- American Society of Plant Biologists
Abstract
As a plant shoot matures, it transitions through a series of growth phases in which successive aerial organs undergo distinct developmental changes. This process of phase change is known to be influenced by gibberellins (GAs). We report the identification of a putative transcription factor, GLABROUS INFLORESCENCE STEMS (GIS), which regulates aspects of shoot maturation in Arabidopsis thaliana. GIS loss-of-function mutations affect the epidermal differentiation of inflorescence organs, causing a premature decrease in trichome production on successive leaves, stem internodes, and branches. Overexpression has the opposite effect on trichome initiation and causes other heterochronic phenotypes, affecting flowering and juvenile–adult leaf transition and inducing the formation of rosette leaves on inflorescence stems. Genetic and gene expression analyses suggest that GIS acts in a GA-responsive pathway upstream of the trichome initiation regulator GLABROUS1 (GL1) and downstream of the GA signaling repressor SPINDLY (SPY). GIS mediates the induction of GL1 expression by GA in inflorescence organs and is antagonized in its action by the DELLA repressor GAI. The implication of GIS in the broader regulation of phase change is further suggested by the delay in flowering caused by GIS loss of function in the spy background. The discovery of GIS reveals a novel mechanism in the control of shoot maturation, through which GAs regulate cellular differentiation in plants.
INTRODUCTION
Organ initiation in plants, in contrast with animals, takes place through much of the life cycle. Successive organs may, however, develop distinctive morphological and physiological characteristics that are dependent on the developmental stage of the plant at the time the organ primordia are fated. This maturation process occurs both before and after the plant develops reproductive competence. During embryogenesis, the shoot apex produces embryonic leaves or cotyledons, which contrast with postembryonic leaves in their morphology and role in storage reserve accumulation. After germination, successive leaves go through a juvenile phase before adopting an adult form, a process termed vegetative phase change, which is largely age dependent but independent of growth rate (Telfer et al., 1997). Shoot maturation also affects leaves that are generated as the plant is committed to flowering, and their pattern of differentiation is thought to depend on the developmental state of corresponding primordia in the shoot apex at the time of floral induction (Hempel and Feldman, 1994; Telfer et al., 1997). In Arabidopsis thaliana, the changing distribution of trichomes, which are spiky epidermis-derived structures at the surface of aerial organs, is a robust morphological marker of phase change. During vegetative development, juvenile leaves only develop trichomes on their upper (adaxial) side, whereas adult leaves produce trichomes on their adaxial and lower (abaxial) sides. After flowering, trichomes forming on the adaxial side of inflorescence leaves (cauline leaves) follow a contrasting pattern of initiation and are gradually less abundant on successive leaves (Telfer et al., 1997).
Phase change and trichome distribution are known to be influenced by gibberellins (GAs). In maize (Zea mays), GAs promote the expression of adult traits, in particular trichome production, during the vegetative phase (Evans and Poethig, 1995). In Arabidopsis, loss-of-function mutations impairing GA biosynthesis or sensitivity delay the appearance of adult leaves. By contrast, exogenous GA applications accelerate the transition from juvenile to adult phase (Chien and Sussex, 1996; Telfer et al., 1997). In the embryo, the retention of embryonic characteristics by cotyledons is also thought to implicate GAs (Gazzarrini et al., 2004). In relation to their role in phase change, GAs are known to affect the expression of GLABROUS1 (GL1), which encodes a MYB transcription factor that is a core component of a complex necessary for Arabidopsis trichome initiation (Perazza et al., 1998). The trichome initiation complex also comprises the basic helix-loop-helix transcription factors GL3 and ENHANCER of GL3 (EGL3) and the WD40 protein TRANSPARENT TESTA GLABRA1 (TTG1); the transcriptional effects of GAs on GL1 may be important for modulating the activity of the initiation complex as a whole (Larkin et al., 1994; Walker et al., 1999; Payne et al., 2000; Zhang et al., 2003). Despite the growing body of information relating to the regulation of vegetative phase change (Telfer and Poethig, 1998; Clarke et al., 1999; Groot and Meicenheimer, 2000; Berardini et al., 2001; Prigge and Wagner, 2001; Hunter et al., 2003), the molecular mechanisms through which GAs regulate different aspects of shoot maturation in adult plants, in particular trichome initiation, are unknown.
We report here the identification of a putative C2H2 transcription factor, which modulates the regulation of shoot maturation by GAs, and show that it plays a central role in epidermal differentiation through its influence on GL1 activity.
RESULTS
Overexpression of GLABROUS INFLORESCENCE STEMS in Arabidopsis Stimulates Trichome Initiation and Causes the Heterochronic Expression of Juvenile Traits
In an effort to uncover novel functions of transcription factors in plants, a genome-wide reverse genetics analysis of gain-of-function phenotypes was conducted at Mendel Biotechnology across different Arabidopsis transcription factor families (Riechmann et al., 2000). This experiment identified novel genes that play a role in the control of developmental phase change and trichome production. Among them, GLABROUS INFLORESCENCE STEMS (GIS; corresponding to At3g58070) was first uncovered by screening transgenic Arabidopsis plants overexpressing transcription factor genes. Multiple independent 35S:GIS overexpression lines displayed an abnormally high density of trichomes on inflorescence organs (Figures 1A and 1B ). Two representative lines were selected for detailed studies: line 6 and line 8, which exhibited high and moderate levels, respectively, of GIS overexpression (81- and 18-fold wild-type levels). In both of these lines, the distribution of trichomes on cauline leaves was significantly altered: instead of decreasing in numbers on the adaxial side of successive leaves, trichomes remained unusually abundant, which suggested that progression of the epidermal differentiation program was delayed. In lines 6 and 8, the average adaxial trichome density on the second cauline leaf was 68 and 45% higher, respectively, than in control plants (Figure 2F ). Significantly more trichomes were also noticeable on stems and sepals in the overexpressors. In addition to suppressing the progressive decline in trichome production, GIS overexpression caused the formation of ectopic trichomes on carpels, petals, and even stamens (Figure 1B). By contrast, trichome density was normal on rosette leaves of 35S:GIS lines (Figure 2F).
Phenotype of Loss-of-Function gis Mutants and 35S:GIS Overexpressors.
(A) Main inflorescence stems (2nd internode) of the wild type (left), 35S:GIS transgenic line 8 (center), and gis (right).
(B) Exposed floral organs from a wild-type (left) and 35S:GIS (right) dissected flower. Arrows point to ectopic trichomes.
(C) Rosette formation on the primary inflorescence stem of a strong GIS overexpressor.
(D) Comparison inflorescence stems of a complemented 35S:GIS gis line (right) and gis (left).
(E) Trichomes on an inflorescence stem (2nd internode) of a transgenic plant in which GIS has been silenced by RNAi (line GIS-Ri-1).
(F) and (G) Trichome branching pattern on gis (F) and control stems (G). Note that gis trichomes are smaller and more branched than their wild-type counterparts.
Trichome Initiation on Inflorescence Organs of gis Mutants and 35S:GIS Overexpressing Lines.
(A) to (D) Trichome density in gis and wild-type plants (A), first internodes of successive paraclades (B), main stem internodes (C), and sepals (D). Flower trichome counts represent the total for 20 flowers. Values are averages, and error bars correspond to standard error. Black bars, wild type; white bars, gis mutant; par., paraclade.
(E) Trichome density on inflorescence organs in gis homozygotes and gis/+ heterozygotes. Dark gray bars, wild-type controls; gray bars, heterozygotes (Het.); white bars, homozygotes (Hom.).
(F) Trichome density on inflorescence organs and on rosette leaves in 35S:GIS-overexpressing lines. O-6, 35S:GIS-overexpressing line 6; O-8, 35S:GIS-overexpressing line 8; int., internode.
35S:GIS plants also displayed a number of phenotypic changes that we interpreted as heterochronic shifts in development. For example, 35S:GIS plants flowered significantly later than wild-type plants, after producing more leaves (Table 1 ). GIS overexpression also caused the occasional appearance of aerial rosettes on the inflorescence stems of transgenic plants, in place of cauline leaves (Figure 1C).
Effect of GIS Loss of Function and Overexpression on Plant Growth and Development
In summary, GIS overexpressors displayed a phenotype consistent with a delay in shoot maturation, associated with a strong induction of trichome production on the inflorescence.
GIS Loss-of-Function Affects the Timing of Trichome Initiation on Inflorescence Organs
To test whether the true biological function of GIS is to mediate morphological changes associated with phase change, we obtained, from the GABI-Kat library of insertional mutants (Rosso et al., 2003), a line in which GIS is interrupted by a T-DNA insertion. Presence of the T-DNA at the expected location was verified by genomic PCR, and GIS expression was undetectable based on quantitative PCR analysis (see Supplemental Figure 1A online).
We focused our initial analysis on trichome production in the mutant. The pattern of trichome initiation is well documented for leaves but not for other aerial organs. To better assess the possible role played by GIS in epidermal differentiation, we first characterized in detail trichome distribution on stems, paraclades (secondary inflorescence shoots or branches), and flowers in wild-type plants. In addition to the known decline in adaxial trichome production on successive inflorescence leaves (Telfer et al., 1997), we observed that trichome initiation decreases on successive stem internodes, branches, and flowers (Figures 2A to 2D). Overall, we found a striking correlation between trichome density and distance from the base of the inflorescence (see Supplemental Figure 2 online).
In contrast with the overexpression phenotype, trichome initiation was negatively affected on all inflorescence organs in gis mutants, and the effect was stronger in later than in earlier organs (Figures 2A to 2D). The reduction in trichome density was most noticeable on paraclade stems (branches), which were near-glabrous or glabrous (hence the gene name), and was also detectable on main stem internodes and cauline leaves, mostly above the first branch. The flowers of gis followed a similar trend, with a more steady decrease in trichome density. The number of branches or spikes on trichomes was also affected in the mutant. In contrast with wild-type stems, where most trichomes possessed one or two branches, an average 85% of the trichomes on gis stems had three branches and were more similar in this respect to leaf trichomes (Figure 1F). Such an increase in trichome branching was also observed on rosette leaves, most noticeably on their abaxial side (data not shown). However, in contrast with inflorescence organs, rosette leaves produced a normal number of trichomes in the mutant (see below). In addition, unlike some of the trichome initiation mutants, gis did not exhibit any obvious defect in anthocyanin, root hair, or mucilage production (data not shown).
As GIS overexpression not only affected trichome production, but also induced other heterochronic phenotypes, we examined the process of shoot maturation in the mutant. We found that the mutation did not affect flowering time in long or short days, the rate of leaf initiation, or plant size (Table 1). We also examined leaf shape and venation patterns during vegetative and inflorescence development but did not find any significant difference between gis and wild-type plants. Therefore, in contrast with overexpression, GIS loss of function mainly affected epidermal differentiation after floral induction under normal conditions.
To verify the GIS loss-of-function phenotype, we generated multiple transgenic lines in which the GIS gene was silenced by an RNA interference (RNAi)–based approach (see Supplemental Figure 1B online); most of these lines showed a phenotype comparable to that of the T-DNA mutant (Figure 1E). As a further verification that the phenotype was caused by the GIS loss-of-function mutation and not by another linked mutation, we overexpressed GIS under the control of the 35S promoter in the gis T-DNA line background. Constitutive overexpression rescued the mutant phenotype in most lines, also causing an increasing in trichome density on flowers and on stems (Figure 1D).
The gis Mutation Is Semidominant
To determine the influence of gene dosage on the trichome phenotype, we analyzed an F2 population derived from a gis/+ plant. The genotype of segregating progenies was determined using genomic PCR (data not shown), and their phenotype was analyzed. In line with our expectations, we found that all lines homozygous for the T-DNA insertion were defective in trichome initiation. By contrast, all of the plants homozygous for the wild-type allele were found to have a normal trichome phenotype. Heterozygous plants also showed a marked decrease in trichome density and displayed a phenotype that was intermediate between wild-type and homozygous plants (Figure 2E). This observation indicated that the gis mutation is semidominant and that relatively small variations in GIS expression can have a dramatic effect on the processes that the gene controls.
GIS Acts Upstream of the Trichome Initiation Complex
Since trichome initiation is affected in gis and the overexpressor, we investigated the position of GIS in the trichome initiation pathway and overexpressed the GIS gene in the gl1, gl3, and ttg-1 mutant backgrounds. Trichome initiation was not rescued by GIS overexpression in gl1, gl3 (Figures 3A and 3B ), or ttg-1 (data not shown). In a complementary experiment, we asked whether the trichome defect of gis mutants could be overcome by increased activity of the trichome initiation complex. To test this possibility, we overexpressed, in the gis mutant background, the maize basic helix-loop-helix regulator R, which is known to be functionally equivalent to trichome initiation regulators GL3 and EGL3 in Arabidopsis (Lloyd et al., 1992). Expression of R under the control of the 35S promoter led to a strong increase in leaf and stem trichome initiation in transgenic plants (Figure 3C). Taken together, these results suggested that GIS acts either upstream or at the same step as GL1, TTG1, and GL3.
Genetic Interactions between GIS and Trichome Initiation Regulatory Genes.
(A) and (B) Trichome initiation on main stems (first internode) of 35S:GIS overexpressors in two different mutant backgrounds: gl1 (left) and 35S:GIS gl1 (right) (A); gl3 (left) and 35S:GIS gl3 (right) (B).
(C) Trichome initiation on stems (2nd internode) of gis mutants overexpressing the maize R gene. R overexpression rescues the trichome initiation phenotype of the gis mutant.
(D) Effect of the gai mutation on the GIS overexpression phenotype. Trichome initiation is inhibited in the absence of GA signaling. Left, gai flowers; right, 35S:GIS gai flowers.
To further assess the functional relationship between GIS and trichome initiation regulators, we measured the expression level of GL1, GL3, and EGL3 in developing stems and flowers of the gis mutant. As shown in Figure 4 , the expression of GL1, GL3, and EGL3 was significantly lower in the mutant than in wild-type controls. The difference was not simply the result of a decrease in the production of trichomes, where these genes are also expressed, since no significant difference in GL1 expression could be detected in ttg-1 (which is completely glabrous) and since GL3 is upregulated in this mutant (Zhang et al., 2003) (Figure 4). By contrast, we found that the expression of GL1 and GL3 was significantly increased in 35S:GIS plants and that their induction level correlated with the level of expression of GIS overexpression. It was not the case for TTG1 expression, which was neither affected in the overexpressing lines nor in the mutant (Figure 4). Taken together, these genetic experiments and expression studies argued that GIS acts upstream of the trichome initiation complex.
Quantitative PCR Analysis of Regulatory Genes Involved in Trichome Initiation in gis Mutants and 35S:GIS Overexpressors.
Relative expression of GL1, GL3, EGL3, and TTG1 in gis mutants, 35S:GIS-overexpressing lines, and ttg-1 mutants. Values represent the ratios of gene expression in a particular genotype to corresponding wild-type controls (dotted line indicates a ratio of 1, or identical expression to the wild type). Genotypes are indicated above each expression value. Tissues and genes used in the analysis are indicated under the x axis. O-6, 35S:GIS-overexpressing line 6; O-8, 35S:GIS-overexpressing line 8; DI, developing inflorescence shoots; F, flowers.
GIS Encodes a Transcription Factor of the C2H2 Family That Is Highly Expressed in Stem Epidermis and at Early Stages of Inflorescence and Flower Development
Analysis of the predicted amino acid sequence of GIS indicates that it contains a C2H2 domain found in transcription factors of the TFIIIA class and is most similar to KNUCKLES and to the ZFP group of transcription factors (Figure 5 ; see Supplemental Figure 3 online) (Meissner and Michael, 1997; Payne et al., 2004).
Structure of the GIS Gene and Similarity between GIS and Related Proteins.
(A) Schematic representation of the GIS gene. Conserved residues of the C2H2 domain are showed in an insert above the DNA binding domain. The C2H2 motif is underlined. DBD, DNA binding domain; T-DNA, T-DNA insertion.
(B) Alignment of the conserved regions of GIS and 12 most related proteins. Conserved residues are underlined, and the C2H2 motif is highlighted. Asterisks indicate stop codons.
(C) Phylogenetic tree of protein sequences similar to GIS. Bootstrap values are provided near the nodes.
To determine the pattern of GIS expression in wild-type plants, we first performed a quantitative RT-PCR analysis of transcript levels in different organs. In agreement with the phenotype of gis mutants, we found that GIS expression is low in rosette leaves and undetectable in roots. By contrast, GIS is most highly expressed in developing stems and branches, where expression levels are fairly consistent in successive paraclades (Figure 6A ). To refine our analysis, we performed in situ hybridizations using GIS-specific probes on sections of developing inflorescence stems. This experiment confirmed that the gene is expressed broadly in the inflorescence, particularly in primary and secondary inflorescence meristems. It also showed that GIS is expressed strongly in the stem epidermis and in floral meristems (Figures 6B, 6D, and 6E).
Expression Pattern of the GIS Gene.
(A) Quantitative RT-PCR analysis of GIS expression in different tissues of wild-type plants. Analysis of GIS expression in successive paraclades was performed in an independent experiment. All values are normalized using a common internal standard (UBQ10). Ros. Leaf, rosette leaf; Caul. Leaf, cauline leaf; Fl., flower; Dev. stem, developing main stem (1st internode); Mat. stem, fully elongated first internode of the main stem; Sil., silique; Dev. br., developing branch (paraclade).
(B) to (F) In situ hybridization of GIS probes to developing inflorescence shoot sections. GIS is expressed broadly in stems and strongly in inflorescence meristems (IM), axillary meristems (AM), floral meristems (FM), and in the epidermis (E). Antisense hybridizations ([B], [D], and [E]); control sense hybridizations ([C] and [F]).
In summary, we found that the pattern of GIS expression in the inflorescence, epidermis, and trichomes is consistent with the trichome phenotypes of overexpressors and loss-of-function mutants, although it could also support a wider role in the control of inflorescence maturation that is not apparent in the mutant but is suggested by the overexpression phenotype.
Epidermal Differentiation Is Less Responsive to GAs in gis Mutants
The epidermal phenotype of knockout mutants and overexpressors suggests that GIS acts in part to slow or prevent changes in the pattern of trichome initiation that are normally associated with shoot maturation, at least during reproductive development. Due to the known implication of GA signaling in phase change and trichome production, we investigated how fluctuations in GA levels would affect the gis phenotype. To this end, we submitted gis mutants to different levels of GA3 or paclobutrazole (PAC), which is a GA biosynthesis inhibitor. In addition to their effect on flowering and growth, GAs are known to promote trichome initiation on leaves, whereas PAC has the opposite effect (Chien and Sussex, 1996).
As expected, PAC applications induced late flowering and the shortening of internodes, indicating that the treatment was successful. By contrast, plants treated with GA flowered earlier and had a characteristic tall phenotype (Table 2).
We did not observe any significant difference in the mutant response to variations in GA signaling with respect to flowering time and leaf production (Table 2). However, there was a small but significant difference (t test, P < 0.001; n = 20) in the number of juvenile leaves produced by gis mutants compared with wild-type plants when high levels of PAC or GAs were applied: abaxial trichomes appeared noticeably later in the mutant on PAC- or GA-treated rosette leaves (Table 2 ), which suggested that the mutant was more sensitive to decreases and less to increases in GA levels. In line with these observations, we noticed that trichome initiation on later leaves was also less responsive to GA and more to PAC applications (Figure 7C ). PAC also inhibited trichome initiation on cauline leaves and stems more strongly in the mutant than in wild-type plants, whereas trichome production was less induced if at all by GA applications (Figures 7A, 7B, and 7D). Taken together, these observations indicated that the mutant is altered in its sensitivity to GAs and suggested a role for GIS in the epidermal expression of vegetative phase change that is only revealed when GA signaling is altered.
Developmental Effect of GA3 and PAC Applications on gis Mutants
GAs and the gis Mutant Phenotype.
(A) and (B) Effects of GA applications at different concentrations on trichome initiation in gis mutants and controls.
(A) Trichome initiation on the first internode of the second branch.
(B) Trichome initiation on the abaxial side of the 3rd cauline leaf.
(C) and (D) Effects of GA and PAC applications on leaf (7th rosette leaf; [C]) and main stem trichome (1st internode; [D]) initiation of gis mutants and controls.
Squares, wild-type controls; triangles, gis mutants; black lines, GA treatments; gray lines, PAC treatments. GA and PAC treatments were done in separate experiments with a different set of control plants, and 20 plants were sampled for each condition. All values are averages, and error bars represent the standard error.
GIS Acts Downstream of SPINDLY and Affects the Flowering Phenotype of spindly Mutants
To further probe the role of GAs in modulating its function, we investigated the interactions between GIS and components of the GA signaling pathway. We first characterized genetic interactions between GIS and SPINDLY (SPY), which encodes a putative O-linked β-N-acetylglucosamine transferase playing a repressive role in GA signaling (Jacobsen et al., 1996). spy mutants display an early-flowering and accelerated growth phenotype that is largely reflective of constitutive GA response and bear many similarities with the phenotype of GA-treated plants (Jacobsen and Olszewski, 1993). They also transition between the juvenile and adult phases of vegetative development earlier than wild-type plants (Telfer et al., 1997) and produce more trichomes on rosette and inflorescence leaves (Chien and Sussex, 1996; Perazza et al., 1998). We therefore asked whether GIS was required for the expression of the vegetative and inflorescence trichome phenotypes in this mutant and constructed gis spy double mutants, using for this analysis spy-3, which is in the Col-0 background (Jacobsen and Olszewski, 1993).
We confirmed that spy-3 mutants produce leaf abaxial trichomes earlier than control plants (Telfer et al., 1997) and found that, as in spy-5 mutants, trichome production was elevated on leaves and stems (Perazza et al., 1998). By contrast, and consistent with the phenotype of GA-treated plants, we found that trichome densities on upper cauline leaves and paraclade stems of the gis spy mutant were similar to that of gis mutants (Figures 8A to 8C ). Similarly to GA treatments, SPY loss-of-function caused a small increase in trichome density on lower inflorescence organs in gis mutants but more limited than in the wild-type background. Trichome production on the rosette leaves of double mutants was similar to that of spy-3 and wild-type plants, although abaxial trichomes appeared slightly later on rosette leaves, as in gis seedlings that were established on GA-containing medium (Tables 1 and 2). Interestingly, whereas SPY loss-of-function had a negative effect on trichome branching on inflorescence stems, resulting in most trichomes being unbranched, gis spy double mutants had stem trichomes that were mostly three-branched, as in the gis mutants (Figure 8B). This observation, together with the trichome phenotype of late inflorescence organs in gis and gis spy, suggested that gis is largely epistatic to spy. Furthermore, the elevated trichome production on vegetative leaves and basal parts of the inflorescence in gis spy compared with gis, also seen in GA-treated gis mutants, indicated that GA signaling can promote trichome initiation locally, independently of GIS. Such a residual response to GA is also likely to be responsible for the overall increase in rosette leaf trichome branching that was observed on gis spy (data not shown) compared with gis mutants, as spy is known to promote trichome branching on leaves (Perazza et al., 1998).
Genetic Interactions between GIS and SPY and the Effect of Variations in GA Signaling on GIS and GL1 Expression.
(A) to (C) Trichome initiation phenotype of spy, gis, and gis spy mutants in branches (A), main stems (B), and cauline leaves (C). The gis mutant is largely epistatic to spy. The trichome branching phenotype of the different mutants is illustrated in (B). 1-br, one-branched (unbranched) trichomes; 2-br, two-branched trichomes; 3+-br, trichomes with three or more branches.
(D) to (F) Effect of variations in GA signaling on GIS and GL1 expression.
(D) Effect of GA applications on GIS and GL1 expression in developing inflorescences of ga1-3, gai, gis, and wild-type plants. GA3 (100 μM) was applied, and developing shoots were harvested 4 or 6 h later for gene expression analysis.
(E) GIS and GL1 expression in the gai and ga1-3 mutants. All values correspond to ratios of normalized gene expression values (obtained by quantitative RT-PCR) to appropriate controls. Averages of the ratios to mean control values are presented together with standard errors.
(F) GIS and GL1 expression in developing inflorescence stems of spy, gis, and gis spy mutants and control plants measured by quantitative RT-PCR. All values are normalized using a common internal standard (UBQ10).
The trichome phenotype of gis spy mutants could have been largely predicted from the observation of gis mutants to which GAs had been applied. We noticed, however, an important difference when we recorded the growth and development characteristics of the double mutant: in contrast with GA-treated gis mutants, which flowered at a similar time as GA-treated controls, gis spy mutants flowered significantly later than spy mutants (Table 1). A consistent yet smaller delay in flowering was also found in gis/+ spy heterozygotes. These observations suggested that GIS has a positive influence on flowering, independently of GA signaling, which can only be detected when repression by SPY of flowering is abolished. Interestingly, SPY is known to also act independently of GA signaling to repress flowering in long days (Tseng et al., 2004).
To further investigate how GIS loss-of-function affects the phenotype of spy mutants, we measured GIS expression in spy inflorescence stems. As shown on Figure 8F, we found that GIS expression was significantly higher in spy-3. This was in contrast with the expression of SPY in GIS loss-of-function mutants, which we found was unchanged compared with wild-type controls (data not shown). It therefore appeared that SPY is involved, directly or indirectly, in the repression of GIS and that derepression of GIS in spy mutants is associated with an increase in trichome density. Taken together with the phenotype of gis spy mutants, these observations strongly suggested that GIS acts downstream of SPY.
GIS Is GA Inducible and Its Induction Kinetics Are Similar to Those of GL1
The increase in GIS expression in the spy-3 mutant raised the possibility that the gene is responsive to GA levels in the plants. We therefore set out to test the effect of GA on GIS expression by monitoring transcript levels over time after GA applications in wild-type plants and in the GA-deficient mutant ga1-3 (Koornneef and Van Der Veen, 1980; Koornneef et al., 1985; Wilson et al., 1992; Sun and Kamiya, 1994). Plants were sprayed with 100 μM GA shortly after flowering, developing inflorescence stems were harvested, and GIS transcript levels were measured by quantitative PCR 4 and 6 h after treatment. In wild-type plants, GA applications had a small but significant effect on GIS expression (∼1.5-fold increase) as soon as 4 h after treatment. However, this effect was much more pronounced in GA-treated ga1-3 mutants, where GIS expression increased close to fourfold within the same time frame (Figure 8D). GIS transcript levels subsequently decreased in both backgrounds, while still remaining higher than in untreated controls. As a control, GA-insensitive gai mutants were submitted to the same treatments. In these mutants, GAI, a repressor of GA signaling belonging to the DELLA family, is constitutively active (Peng et al., 1997). In contrast with the effect seen in wild-type and ga1-3 mutants, we did not detect any change in GIS expression in GA-treated gai mutants (Figure 8D).
Since we have shown that GL1 is likely to act downstream of GIS and since GL1 expression levels are known to be regulated by GAs (Perazza et al., 1998), we investigated whether GA-mediated induction of GL1 expression occurs within a similar time frame. Similarly to our analysis of GIS expression, we used quantitative PCR to determine GL1 levels in GA-treated wild-type and ga1-3 plants. Interestingly, we found that GL1 induction also occurred within 4 h and decreased between 4 and 6 h after GA treatment. By contrast, GL1 expression, like GIS expression, was not affected by GA applications in the gai mutant (Figure 8D).
These experiments showed that GIS is GA inducible and that changes in its expression depend on the activity of GAI. They also indicated that the response of GL1 expression to variations in GA levels is similar in time to that of GIS and is also abolished by GAI repression.
GIS Mediates the Induction of GL1 by GAs
The promotion of trichome initiation by GAs is thought to involve the regulation of GL1 expression, as the gene is significantly induced by GA and repressed by PAC applications (Perazza et al., 1998). Since we placed GIS upstream of GL1 in the trichome initiation pathway and found that GIS and GL1 were similarly responsive to GA treatments, we decided to test whether GIS is implicated in the activation of GL1 expression by GAs. To this end, we compared GL1 transcript levels in wild-type and gis mutant plants to which GAs had been applied. GA applications significantly induced GL1 expression in wild-type plants: transcript levels increased between twofold and threefold within 4 h in developing main and secondary inflorescence shoots. By contrast, induction was significantly and reproducibly lower in gis mutants, where changes in GL1 expression were still detectable but more limited (Figure 8D).
To further examine the role played by GIS in the regulation of GL1 expression, we also compared GL1 transcript levels in developing inflorescence shoots of gis, gis spy, and spy mutants. Consistent with the effect of GA applications on GL1 expression, GL1 transcripts were significantly more abundant in spy-3 than in wild-type plants. By contrast, GL1 was more weakly expressed in gis spy mutants, and its expression level was comparable in the double mutant and in gis (Figure 8F).
Taken together, these results suggested that GIS mediates the induction of GL1 expression by GAs in inflorescence shoots.
Repression of Trichome Initiation Involves GAI and Is Independent of GIS Expression Levels
We showed that GIS is sufficient for inducing trichome initiation and GL1 expression in transgenic plants and that it also mediates its induction by GAs in developing inflorescence shoots. We also showed that GIS and GL1 induction by GAs depends on GAI activity and does not occur when the repressor is active.
Trichome initiation is known to require normal levels of GA signaling, since the ga1-3 mutant, which is severely deficient in the production of GAs, is near-glabrous. The gai signaling mutant is also deficient in trichome production, especially on inflorescence organs (Figure 3D), although it still produces leaf trichomes. In addition to its effect on initiation, GA deficiency also leads to a drop in GL1 transcript levels (Perazza et al., 1998). Therefore, whereas increases in levels of GA or GA signaling induce GL1 and trichome production, the opposite causes the repression of GL1 expression and inhibits trichome initiation. Since positive regulation by GA, which is antagonized by GAI, involves the induction of GIS expression, we asked whether negative regulation reciprocally proceeds through the repression of GIS expression and if this repression is mediated by GAI.
To this end, we first measured GIS expression in developing inflorescence shoots of the ga1-3 and gai mutants and found that, in agreement with our hypothesis, GIS is expressed at significantly lower levels in the mutants than in wild-type plants. We also found that GL1 was repressed in both mutants (Figure 8E). To test whether GIS downregulation is necessary for repression of trichome initiation, we overexpressed GIS in the ga1-3 and gai mutants. Surprisingly, we found that GIS overexpression does not have an effect on trichome initiation in either ga1-3 or gai. No increase in trichome density was noticeable on leaves, stems, or flowers of 35S:GIS gai plants, and carpels did not produce ectopic trichomes (Figure 3D). ga1-3 plants overexpressing GIS displayed a similar phenotype (see Supplemental Figure 3 online). To test whether GIS expression levels in transgenic plants were sufficient to induce the overexpression phenotype, we sprayed 35S:GIS ga1-3 plants with 100 μM GA. This treatment restored the overexpression phenotype normally seen in a wild-type background (data not shown). We also verified GIS expression levels in 35S:GIS gai and 35S:GIS ga1-3 transgenic lines and found that GIS was indeed overexpressed to significant levels (27- and 442-fold, respectively, in the lines we tested).
The phenotype of GIS overexpressing gai and ga1-3 plants indicated that GAI-mediated repression of trichome initiation is independent of GIS expression levels. To test whether it was also independent of GL1 expression levels, we measured GL1 transcript levels in 35S:GIS gai, 35S:GIS ga1-3, and control plants. We found that GL1 was still repressed in spite of high levels of GIS expression in transgenic plants and in fact was expressed at similar levels in 35S:GIS gai, 35S:GIS ga1-3, and control plants (see Supplemental Figure 4 online).
Therefore, repression of trichome initiation and GL1 expression in the absence of GA signaling, while it involves the GAI-mediated downregulation of GIS expression, occurs independently of GIS expression levels.
DISCUSSION
We report the identification of a putative C2H2 transcription factor, GIS, which regulates several aspects of shoot maturation in Arabidopsis. The analysis of overexpressors and loss-of-function mutants indicates that GIS plays a central role in the control of trichome initiation during inflorescence development. GIS also plays a role in the regulation of epidermal differentiation during the vegetative phase, although this function only becomes apparent when GA signaling is altered in the plant. The phenotype of gis spy double mutants indicates that GIS also modulates the repressive effect exerted by SPY on flowering. Through the analysis of gene expression profiles, genetic interactions, and effects of modulating GA levels in the plant, we have found that GIS acts downstream of the GA signaling pathway and controls epidermal differentiation by modulating the activity of one or more cognate regulators of trichome initiation.
Role of GIS in Shoot Maturation
The phenotype of loss-of-function mutants and overexpressors indicates that GIS promotes epidermal differentiation during inflorescence development by inducing trichome initiation. Based on the extent of trichome loss in the mutant, it appears that GIS plays a predominant role in inflorescence stems (Figures 2A to 2D), which is consistent with its strong expression in stem epidermal cells (Figure 6).
Our analysis of the genetic interactions between GIS, GL1, GL3, and TTG1 and of gene expression also suggests that GIS acts upstream of the trichome initiation complex and that it modulates the activity of the complex through a direct or indirect transcriptional mechanism.
While the gain-of-function and loss-of-function phenotypes and gene expression data argue that GIS is an activator of inflorescence trichome initiation, the occasional appearance of rosettes on overexpressors could suggest that GIS is more generally repressing inflorescence shoot maturation and controls the timing of trichome initiation rather than initiation itself. This scenario could be supported by the broad expression of GIS during inflorescence development, which would be compatible with a regulatory role that goes beyond epidermal differentiation. Whether GIS plays a direct or indirect role in promoting trichome initiation should become clearer once its immediate targets are identified.
Regardless of this mechanism, the residual production in gis of trichomes on early inflorescence organs where the gene is normally expressed suggests that other regulators act in parallel to control trichome initiation. The regulation of these redundant factors is also likely to be important to the progression of the epidermal differentiation program, and it will be interesting to determine whether they respond to distinct developmental signals. In this respect, it will be sensible to define the function of C2H2 proteins that are most closely related to GIS, as the similarity in sequence between the homologs is suggestive of redundancy.
The role of GIS in other aspects of shoot maturation appears to be more complex. The timings of vegetative phase change and flowering are not affected in the mutant under normal conditions, although the results of exogenous GAs and PAC applications imply that abaxial trichome initiation requires higher GA levels in the mutant than in wild-type plants. This observation is in agreement with the phenotype of gis spy double mutants and suggests that GIS plays a role in promoting abaxial trichome production during the juvenile–adult transition (which would support a role for GIS as an activator of trichome initiation throughout the plant). An interesting finding is that GIS loss of function also affects flowering time in the spy background. This effect appears to be independent of GA levels as GA- and PAC-treated gis mutants do not flower significantly earlier or later than wild-type plants. The two results are not necessarily inconsistent, as SPY has recently been shown to be implicated in the GA-independent control of flowering under long days through its interactions with GIGANTEA (Tseng et al., 2004). GIS could also play a role in this pathway by antagonizing the repressive effect of SPY on flowering. To examine this possibility, we are currently exploring genetic interactions between gis, gigantea, and constans.
A positive role by GIS in the regulation of flowering under long days and in the regulation of vegetative phase change by GA is in apparent contradiction with the phenotype of GIS overexpressors. While in agreement during inflorescence development, the loss-of-function and gain-of-function phenotypes seem inconsistent during the vegetative phase, as both mutant and overexpressors are, under inductive conditions, delayed in phase transitions. One possibility is that overexpression of GIS results in a dominant-negative effect that equates loss of function in vegetative organs but not in inflorescence organs, for example, through the inhibition of a regulatory complex that only forms before flowering. Alternatively, aspects of the overexpression phenotype that are not mirrored in the mutant could be an artifact of ectopic expression. Regardless of what causes the preflowering phenotype of GIS overexpressors, it is clear that GIS is implicated in the control of phase change during vegetative development and acts by integrating GA-dependent and -independent signals. Identification of proteins interacting with GIS and functional characterization of close homologs should bring further definition to its mode of action during the vegetative phase.
GIS and GA Signaling
The lower sensitivity of gis mutants to GAs, indicated by the result of GA and PAC applications, implies that GIS plays a role in modulating GA signaling in the regulation of shoot maturation. Such a role is further suggested by genetic interactions between GIS, GA1, GAI, and SPY. In particular, we found that the inflorescence phenotype of gis spy is most similar to that of gis mutants and that GIS is upregulated in spy, which suggests that GIS acts downstream of SPY. The fact that gis strongly attenuates the induction of GL1 by GAs indicates that the mode of action of GIS is in part to mediate GA signaling in the control of GL1 expression. In this, GIS is antagonized by GAI, which also represses the induction of trichome initiation when GIS is overexpressed in the gai mutant background. The impact of GAI on the competence of the plant to respond to GIS is reminiscent of the effect of the ga1-3 mutation on LEAFY overexpression, although in this case, the overexpressors still respond, albeit weakly, to LEAFY activity (Blazquez et al., 1998).
In summary, GIS is required for the induction of trichome initiation through its role in regulating GL1 expression, but its downregulation in the absence of GA signaling is not necessary for the repression of trichome initiation and GL1 expression. These observations lead us to a model of GA control over trichome initiation, according to which GIS acts downstream of SPY but at the same step as GAI (and possibly other DELLA proteins) in the regulation of GL1 to influence trichome initiation (Figure 9 ).
Proposed Model of the Regulation of Inflorescence Trichome Production and Shoot Maturation by GIS in Arabidopsis.
Arrows in bold correspond to relationships that were investigated in this study, and they reflect either genetic analysis, gene expression data, or both.
In conclusion, the identification of GIS is a first step in the elucidation of mechanisms through which GAs control cellular differentiation and epidermal aspects of phase change. Further definition of these mechanisms and of the interaction between GA-dependent and -independent pathways is likely to come from continued analysis of the biological roles of GIS and other members of its clade.
METHODS
Plant Material and Growth Conditions
Arabidopsis thaliana ecotype Col-0 was used as a control for most experiments in this study. The gl1-1, ttg-1, gai-1, ga1-3, and spy-3 mutants were obtained from the Nottingham Arabidopsis Stock Centre, and the gl3-1 mutant was kindly proved by Alan Lloyd. The gis spy-3 double mutants were selected out of an F2 population by selection on Murashige and Skoog medium containing sulfadiazine and PAC. Double mutants were confirmed by selfing the selected F2s and ensuring all progenies were resistant to both sulfadiazine and PAC. Plants of the Lansdsberg erecta ecotype were grown as a control whenever necessary. In all trichome counting experiments, the plants were grown under a 16-h-light (90 μE·cm−2·s−2; 21°C) and 8-h-dark (18°C) cycle. Short-day experiments were performed with 8-h-light (90 μE·cm−2·s−2; 21°C) and 16-h-dark (18°C) cycles. Sepal trichome numbers were recorded ∼35 d after sowing, when the plant reached ∼11 cm in size, on three apical flowers per plant. Leaf trichomes were counted when the leaves had fully expanded, on a 1-cm−2 area in the middle of the leaf. When the main inflorescence stem reached ∼18 cm in size, the total number of trichomes was counted on the first two to three branches, and on one to three basal internodes, depending on the experiment. At least 20 plants were used for trichome count analysis for each of the treatment × genotype combinations. All the experiments were repeated at least once.
Isolation of a GIS Knockout Mutant
A transgenic line (catalogue number 423G08) carrying a T-DNA insertion in GIS was identified in the GABI-Kat line collection (Rosso et al., 2003). Homozygous mutants were selected by ensuring that all of their progenies were resistant to sulfadiazine (5.2 mg/L). Presence of the T-DNA insertion was confirmed by PCR using gene-specific primers (5′ primer, 5′-GTTCGCGTTTGTGAGCGTTTT-3′; 3′ primer, 5′-TACGAAAATGCCACCCATCCA-3′; and a T-DNA insertion primer, (5′-GGGCTACACTGAATTGGTAGCTC-3′).
GA and PAC Applications for Sensitivity Assays and Gene Expression Analysis
GA3 (Sigma-Aldrich) was used for all GA applications. Wild-type and gis mutant plants were grown on soil until the first three leaves had emerged and then sprayed with GA3 solutions. Spraying was performed twice a week until the plants reached a size of ∼13 cm. Control plants for each treatment were sprayed with a mock solution without GA3.
For experiments aimed at measuring induction of gene expression in response to GA, gai, gal-3, gis, and control plants were grown on soil until young inflorescence shoots or paraclades had reached a size of 2 to 3 cm. The plants were then sprayed with either 100 μM GA3 or a mock solution, and the shoots were harvested 4 or 6 h after treatment for RNA extraction.
Aqueous solutions of PAC (Sigma-Aldrich) were applied to soil-grown plants by soaking the pots for 2 d when the plants had reached the four-leaf stage, unless stem trichome density was to be measured, in which case PAC was applied at the eight-leaf stage. The PAC solution was then removed from trays, and normal watering was resumed. Trichome data were recorded ∼35 d after sowing.
Molecular Biology
RNA Extraction and Real-Time RT-PCR
Plant RNA was extracted using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Pooled samples from at least eight plants were used for extractions. Gene-specific primer sequences went as follows: GIS, 5′-TTCATGAACGTCGAATCCTTCTC-3′ and 5′-ACGAATGGGTTTAGGGTTCTTATCT-3′; UBQ10, 5′-GGTTCGTACCTTTGTCCAAGCA-3′ and 5′-CCTTCGTTAAACCAAGCTCAGTATC-3′; GL1, 5′-CGACTCTCCACCGTCATTGTT-3′ and 5′-TTCTCGTAGATATTTTCTTGTTGATGATG-3′; GL3, 5′-GGTACCACAGAACATATTACGGAAGA-3′ and 5′-CAAGAACGTTGTCGATGTGATAATC-3′; EGL3, 5′-TTGATCCCTTAAGTGACGATAAATACA-3′ and 5′-CAAACCCGCTAGTAGAAGTTGTTG-3′; TTG1, 5′-CCGTCTTTGGGAAATTAACGAA-3′ and 5′-GCTCGTTTTGCTGTTGTTGAGA-3′. The primers were designed to include, when possible, an intron-exon boundary in the amplicon. cDNA was synthesized from 5 μg of total RNA using Superscript reverse transcriptase (Invitrogen) and random hexamer primers in a 40-μL reaction according to the manufacturer's instructions. For real-time PCR, the cDNAs were diluted to 200 μL, and 3.5 μL was added to 12.5 μL of SYBR-green PCR mix (Applied Biosystems) and 4.5 μL of each primer (198 nM final concentration) in triplicate 25-μL reactions. PCR and detection were performed using an ABI Prism 7000 thermal cycler (Applied Biosystems), using the following cycling conditions: 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Optimization experiments were performed to establish the optimal concentration of primers. Melting curve analysis and gel electrophoresis of the PCR products were used to confirm the absence of nonspecific amplification products. UBQ10 transcripts were used as an endogenous control to normalize expression of the other genes. UBQ10 was chosen as the housekeeping gene because its expression appeared to be most stable between different tissues and treatments (Gan et al., 2005). Relative expression levels were calculated by subtracting the threshold cycle (Ct) values for UBQ10 from those of the target gene (to give ΔCt) and then calculating 2−ΔCt. RT-PCR experiments were performed on two independent samples.
Cloning
For all cloning experiments (35S:GIS, GIS-RNAi, GIS promoter:β-glucuronidase fusion [pGIS:GUS], and 35S:R), except for those involving the maize (Zea mays) R gene, all sequences were first inserted into the pENTR-1A vector (Invitrogen) before being recombined into an appropriate destination vector using the Gateway LR reaction (Invitrogen). All destination vectors were obtained from VIB (Flanders Interuniversity Institute for Biotechnology). pH2GW7 was used for preparing the 35S:GIS construct, pK7GWIWG2(II) for the GIS-RNAi construct, and pHGWFS7 for the construction of the pGIS:GUS fusion construct.
For all these cloning experiments, gene-specific fragments were first PCR amplified from cDNA (35S:GIS and GIS-RNAi) or genomic DNA (pGIS:GUS) using primers containing SalI and NotI restriction sites, purified using a gel extraction kit (Clontech) before restriction and cloning. The following primers were used: GIS overexpression, 5′-TTTCTCAGTCGACCGCCCAGTCTTTTTATCTCTC-3′ and 5′-TCATTCAGCGGCCGCACACATCGTGCCGTTTCTT-3′; GIS-RNAi, 5′-CATTGTCGACTTACCGTCATTACCCGTCGT-3′ and 5′-TCGCGGCCGCACACATCGTGCCGTTTCTT-3′. A 1.6-kb genomic fragment upstream of the start codon in GIS was amplified using the primers 5′-ATCTTGGTCGACTGCACACACTTTTATGGCAAA-3′ and 5′-CTAATGGCGGCCGCGAGAGATAAAAAGACTGGGCG-3′ for preparing the pGIS:GUS construct.
The R gene was PCR amplified from maize cDNA using the primers 5′-CTGAGTCGACATCGAGTTGTTGTACTCTTCGCAGA-3′ and 5′-TCGATCCCGCGGCCGCTTCCATGCCCGTCGATGTCCAAA-3′, cloned first into the pMEN065 vector, then subcloned into the pBI101 vector.
All binary vector constructs were introduced into Agrobacterium tumefaciens strain GV3101 by electroporation. Agrobacterium-mediated transformation of all Arabidopsis genotypes was performed using the floral dip method (Clough and Bent, 1998), and transgenic seeds were selected using hygromycin (35S:GIS and pGIS:GUS) or kanamycin (GIS-RNAi and 35S:R).
In Situ Hybridization
Nonradioactive in situ hybridization was performed according to a published protocol (Long and Barton, 1998). For synthesis of the antisense and sense GIS RNA probes, a gene-specific fragment was amplified using the same primers as for generating the GIS-RNAi constructs (see above) and cloned into pGEM-T Easy vector (Promega). The resulting plasmid served as template for in vitro transcription, which was performed using the DIG RNA labeling kit (Roche Molecular Biochemicals).
Phylogenetic Analysis
Phylogenetic trees were generated using alignments of complete predicted protein sequences using the ClustalX program (Figure 5; see Supplemental Figure 3 online) (Thompson et al., 1997). The same program was used to produce alignments of conserved regions of the proteins (Figure 5). Alignment parameters were as follows: gap opening penalty = 10 and gap extension penalty = 0.2. Gonnet weight matrices were selected as a way to determine the similarity of nonidentical amino acids. The trees were generated using the neighbor-joining method (Saitou and Nei, 1987) with a number of bootstrap replicates set at 1000.
Accession Numbers
Sequence data from this article can be found in the GenBank/EMBL data libraries under the following accession numbers: GIS, At3g58070; GL1, At3g27920; GL3, At5g41315; EGL3, At1g63650; TTG1, At5g24520; UBQ10, At4g05320.
Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure 1. GIS Expression in T-DNA and RNAi Lines.
Supplemental Figure 2. Developmental Regulation of Trichome Initiation on Wild-Type Inflorescence Organs.
Supplemental Figure 3. Alignment of the Predicted Amino Acid Sequences of Transcription Factors Related to GIS.
Supplemental Figure 4. Effect of GIS Overexpression on the ga1-3 Phenotype and on GL1 Expression Levels in ga1-3 and gai.
Acknowledgments
We thank Ottoline Leyser for useful comments on the manuscript, the Garfield Weston Foundation for funding Y.G.'s work, and all the scientists at Mendel Biotechnology who have contributed, directly or indirectly, to this work.
Footnotes
-
The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (www.plantcell.org) is: Pierre Broun (pbroun{at}york.ac.uk).
-
Article, publication date, and citation information can be found at www.plantcell.org/cgi/doi/10.1105/tpc.106.041533.
-
↵[W] Online version contains Web-only data.
- Received January 30, 2006.
- Revised March 13, 2006.
- Accepted March 30, 2006.
- Published May 5, 2006.
References
