Starch metabolism and antiflorigenic signals modulate the juvenile-to-adult phase transition in Arabidopsis

Authors


Correspondence: I. G. Matsoukas, Faculty of Advanced Engineering & Sciences, The University of Bolton, Deane Road, Bolton BL3 5AB, UK. Fax: +44 (0) 120 490 3088; e-mail: I.Matsoukas@bolton.ac.uk

Abstract

The physiology and genetics underlying juvenility is poorly understood. Here, we exploit Arabidopsis as a system to understand the mechanisms that regulate floral incompetence during juvenility. Using an experimental assay that allows the length of juvenility to be estimated and mutants impaired in different pathways, we show that multiple inputs influence juvenility. Juvenile phase lengths of wild type (WT) accessions Col-0, Ler-0 and Ws-4 are shown to differ, with Col-0 having the shortest and Ws-4 the longest length. Plants defective in sugar signalling [gin1-1, gin2-1, gin6 (abi4)] and floral repressor mutants [hst1, tfl1, tfl2 (lhp1)] showed shortened juvenile phase lengths compared to their respective WTs. Mutants defective in starch anabolism (adg1-1, pgm1) and catabolism (sex1, sex4, bam3) showed prolonged juvenile phase lengths compared to Col-0. Examination of diurnal metabolite changes in adg1-1 and sex1 mutants indicates that their altered juvenile phase length may be due to lack of starch turnover, which influences carbohydrate availability. In this article, we propose a model in which a variety of signals including floral activators and repressors modulate the juvenile-to-adult phase transition. The role of carbohydrates may be in their capacity as nutrients, osmotic regulators, signalling molecules and/ or through their interaction with phytohormonal networks.

Introduction

Plants undergo a series of qualitative transitions during their life cycle in response to both environmental and endogenous cues. One of the most distinguishable is the transition from a vegetative-to-reproductive phase of development. This stage is preceded by the juvenile-to-adult phase transition within the vegetative phase. During the juvenile phase, plants are incapable of initiating reproductive development and are insensitive to environmental stimuli such as photoperiod and vernalization, which induce flowering in adult plants. The juvenile-to-adult phase transition has long attracted interest as an important developmental trait, especially in those species where juvenility is prolonged. Knowledge gained about regulation of the juvenile-to-adult phase transition could help with crop scheduling, decrease time to flowering and reduce waste with resulting benefits for the environment through lower inputs and energy required per unit of marketable product.

The genetics and physiology underlying the juvenile-to-adult phase transition is poorly understood. This transition may be associated with physiological, morphological and biochemical markers (Thomas & Vince-Prue 1984; Poethig 1990). However, these changes are often less distinct in herbaceous plants than in woody species, and in many cases no clear association exists. Floral competence is the most reliable determinant that can be used to distinguish between plants that are juvenile or adult. Within the context of this work, juvenility is defined and measured by insensitivity to long day (LD) photoperiods, which would induce flowering in adult plants.

The greatest advances in our understanding of the genetic regulation of plant developmental transitions have derived from studying the vegetative-to-reproductive phase transition in several dicot and monocot plant species. This has led to the elucidation of multiple environmental and endogenous pathways that promote, enable and repress floral induction (Massiah 2007; Jackson 2009; Matsoukas, Massiah & Thomas 2012). The photoperiodic pathway (also known as the LD pathway) is known for its promotive effect by relaying light and photoperiodic timing signals to floral induction (reviewed in Matsoukas et al. 2012). This pathway involves genes such as PHYTOCHROMES (PHYs) and CRYPTOCHROMES (CRYs), which are involved in the regulation of light signal inputs. Genes such as GIGANTEA (GI), CIRCADIAN CLOCK ASSOCIATED1 (CCA1) and LATE ELONGATED HYPOCOTYL (LHY) are components of the circadian clock, whereas CONSTANS (CO), FLOWERING LOCUS T (FT), TWIN SISTER OF FT (TSF) and FLOWERING LOCUS D (FD) encode proteins that specifically regulate floral induction. The action of the photoperiodic pathway ultimately converges to control the expression of so-called floral pathway integrators (FPIs), which include FT (Kardailsky et al. 1999; Kobayashi et al. 1999), TSF (Yamaguchi et al. 2005), SUPPRESSOR OF CONSTANS1 (SOC1; Yoo et al. 2005) and AGAMOUS-LIKE24 (AGL24; Lee et al. 2008; Liu et al. 2008). These act on floral meristem identity (FMI) genes LEAFY (LFY; Lee et al. 2008), FRUITFUL (FUL; Melzer et al. 2008) and APETALA1 (AP1; Wigge et al. 2005; Yamaguchi et al. 2005), which result in initiation of flowering. Under short day (SD) conditions transcription of FT, the major output of the photoperiodic pathway is repressed. However, as plant growth and development proceeds, FT expression levels show a clear increase (Yanovsky & Kay 2002).

On the other hand, pathways that enable floral induction regulate the expression of floral repressors or translocatable florigen antagonists, known as antiflorigens (Matsoukas et al. 2012). The pathways that regulate the floral repressor FLOWERING LOCUS C (FLC) are the best-characterized (Michaels 2009). In addition, genetic analysis suggests that genes such as TERMINAL FLOWER1 (TFL1; Bradley et al. 1997), LIKE HETEROCHROMATIN PROTEIN1 (LHP1, TFL2; Gaudin et al. 2001; Kotake et al. 2003), TEMPRANILLO (TEM1, TEM2; Castillejo & Pelaz 2008; Osnato et al. 2012) and HASTY1 (HST1; Telfer & Poethig 1998) extend the vegetative phase by repressing the FPIs. Functional analysis of the hst1 Arabidopsis mutant reveals that the juvenile-to-adult phase transition is accompanied by a decrease in miR156 abundance and a concomitant increase in abundance of miR172, as well as the SQUAMOSA PROMOTER BINDING PROTEIN-LIKE (SPL) transcription factors (TFs). Expression of miR172 activates FT transcription in leaves through repression of AP2-like transcripts SCHLAFMÜTZE (SMZ), SCHNARCHZAPFEN (SNZ) and TARGET OF EAT 1–3 (TOE1-3; Jung et al. 2007; Mathieu et al. 2009), whereas the increase in SPLs at the shoot apical meristem (SAM), leads to the transcription of FMI genes (Wang, Czech & Weigel 2009; Yamaguchi et al. 2009). The FMI genes trigger the expression of floral organ identity genes (Causier, Schwarz-Sommer & Davies 2010), which function in a combinatorial fashion to specify the distinct floral organ identities.

Overexpression of CORNGRASS1, a tandem miR156 locus, prolongs juvenility and delays time to flowering in response to starch catabolism (Chuck et al. 2007; Gandikota et al. 2007). Genetic and physiological approaches have demonstrated an involvement of starch with/without an interaction with other plant signal transduction pathways in control of floral induction (Corbesier, Lejeune & Bernier 1998; Dijken, Schluepmann & Smeekens 2004; Chuck et al. 2011; Wahl et al. 2013). In Arabidopsis, mutation in loci such as PHOSPHOGLUCOMUTASE1 (PGM1; Caspar, Huber & Somerville 1985; Caspar et al. 1991), ADP GLUCOSE PYROPHOSPHORYLASE1 (ADG1; Lin et al. 1988; Wang et al. 1998), STARCH-EXCESS1 (SEX1; Yu et al. 2001), SEX4 (Zeeman et al. 1998), CHLOROPLASTIC β-AMYLASE3 (BAM3; Lao et al. 1999) and GI (Eimert et al. 1995) alter the rate of starch synthesis, accumulation or mobilization conferring late flowering phenotypes under non-inductive SD conditions. The late flowering phenotype of starch-deficient mutants in SDs can be rescued by exogenous sucrose application (Corbesier et al. 1998; Yu et al. 2000; Xiong et al. 2009). In LDs or under constant light conditions, the extended daily periods partially suffice to supply enough sucrose to the SAM and the starch-impaired mutants flower similar to wild type (WT).

Arabidopsis mutants allow the investigation of functional interaction between genes involved in different genetic pathways, revealing the complex genetic and physiological regulatory networks that orchestrate developmental transitions in plants. The objective of this study was to investigate the physiological and genetic mechanisms that regulate floral incompetence during juvenility in Arabidopsis. Using an experimental assay that allows the length of the juvenile phase to be estimated based on attainment of floral competence and examination of mutants impaired in different genetic pathways, we demonstrated that multiple inputs influence the timing of the juvenile-to-adult phase transition.

Materials and Methods

Plant material and growth conditions

Mutant and WT Arabidopsis plants used were in Columbia-0 (Col-0), Landsberg erecta-0 (Ler-0) and Wassilewskija-4 (Ws-4) backgrounds. The background and stock number of each genotype used in this study is listed in Supporting Information Table S1. Mutant and WT seeds were sown into Plantpak P24 module trays containing Levingtons F2 compost. Plants were grown in controlled environment cabinets (Saxcil®, Chester, UK) under 100 μmol m2 s−1 photosynthetically active radiation (PAR) at 22 ± 0.5 °C and 70 ± 2% relative humidity. When 50% of seedlings emerged, the trays with the seedlings were transferred into growth cabinets (Saxcil®) and the day-length treatments were initiated. Seven to nine replicate plants were transferred every day from SD to LD conditions with the exception of plants grown in continuous SD and LD conditions where 16 replicate plants were used.

Light sources and spectral measurements

To ensure Arabidopsis plants under LD conditions received similar irradiance to those grown under SD conditions, photoperiod was artificially increased without modifying the total quantity of light available for photosynthesis, by extending the SD treatment with very low intensity wavelengths that are less efficient for photosynthesis and more efficient for a photoperiodic response. SD conditions (8 h d−1; 100 μmol m2 s−1 PAR) were achieved using a combination of fluorescent (General Electric Co., 60W, Budapest, Hungary) and incandescent (Philips 32W, Amsterdam, The Netherlands) light tubes. LD conditions (16 h d−1 light) consisted of a combination of fluorescent (General Electric Co., 60W) and incandescent (Philips 32W) light for the first 8 h d−1 (94 μmol m2 s−1 PAR) and low intensity (6 μmol m2 s−1 PAR) incandescent (Philips 32W) light for the 8 h d−1 extension. Light quality and quantity were measured with an EPP 2000 Fiber Optic Spectrometer (StellarNet Inc., Tampa, FL, USA).

Estimation of juvenile phase length

An analytical approach, which is based on floral competence, was used to estimate the length of the juvenile phase in Arabidopsis plants grown under different experimental conditions. The approach determines the phases of photoperiod sensitivity by conducting transfer experiments in which plants are transferred from SD to LD conditions at regular intervals, from seedling emergence to flowering. The approach enables the analysis of the photoperiod-insensitive juvenile vegetative phase and photoperiod-sensitive floral inductive phase of plant development. The length of these developmental phases were calculated based on the number of rosette leaves and number of days from 50% of seedling emergence at the appearance of the floral bolt at 1 cm height. Flowering time data obtained from the transfer experiments were analysed by fitting the logistic curve, estimating the maximum slope and then fitting the lag time and stationary phase lines using the non-linear regression analysis directive of Sigma Plot 12® (Systat Software, Chicago, IL, USA). The lag time and stationary phase lines were calculated by the upper and lower asymptote of the logistic curve. Analysis of variance (ANOVA) for a randomized complete block design was carried out for all data obtained using Sigma Plot 12® (Systat Software).

Enzymatic assay of sucrose, reducing sugars and starch

For analyses and quantification of glucose, fructose and sucrose, plant material was sampled and immediately frozen in liquid nitrogen. The freeze-dried materials were ground and 50–100 mg used for analysis. Sugars were determined enzymatically in EtOH extracts at 340 nm by a UV/VIS V530 JASCO spectrophotometer, after digestions with β-fructosidase, hexokinase (HXK), glucose-6-phosphatdehydrogenase and phosphoglucose isomerase, using the EZS 864+ kit (Diffchamb; Lyon, France), following the manufacturer's guidelines. Starch was determined from the pellets of the soluble sugar extractions after extensive washing with water. Two millilitres of water was added per pellet, resuspended and centrifuged at 3000 × g for 5 min. The supernatant was removed and the pellet was re-extracted twice using the same procedure. Starch from the air-dried pellets was quantitatively dissolved in dimethyl sulfoxide (DMSO). Pellets were resuspended in 85% v/v DMSO and heated for 30 min at 90 °C. After cooling, 8 M HCl was added and the solution was incubated for a further 30 min at 60 °C. The sample was then centrifuged at 4000 × g for 15 min. After adjusting pH to 4.5 with 5 M NAOH, the starch was precipitated with EtOH (96%; v/v) part of the suspension was digested with amyloglucosidase and HXK/glucose-6-phosphatdehydrogenase. Starch was determined by a UV/VIS V530 JASCO (Easton, MD, USA) spectrophotometer at 340 nm, using the EnzyPlus™ determination kit (Diffchamb), following the manufacturer's guidelines.

Results

Defining the juvenile phase length in Arabidopsis genotypes

Transfer of plants between non-inductive SD and inductive LD photoperiods and measurement of flowering times allowed the length of juvenile phase to be measured. Three Arabidopsis WT accessions and several mutants impaired in different genetic pathways were exploited (Supporting Information Table S1). Among the WT accessions under SD conditions, Ws-4 showed the earliest flowering phenotype, while Col-0 exhibited the latest flowering (Fig. 1). However, under LD conditions, the three accessions flowered similarly. Under SDs starch metabolism mutants pgm1, adg1-1, bam3, sex1 and sex4 flowered significantly later than Col-0 WT (Fig. 1a), while, under LDs only adg1-1 and bam3 maintained this late flowering phenotype. In contrast, genotypes impaired in sugar sensing and signalling glucose insensitive1 (gin1; aba2), gin2 (hxk1) and gin6 (abi4) flowered with their respective WT under LDs, but earlier under SD conditions (Fig. 1b). The floral-repression pathway mutants hst1, tfl1 and lhp1 (tfl2) were early flowering under SD conditions, while under LDs they flowered with their respective WT (Fig. 1b,c).

Figure 1.

Flowering time profiles of Arabidopsis wild types (WTs) and mutant genotypes grown under short day (SD) and long day (LD) conditions.

(a) Number of rosette leaves at flowering of Col-0 WT and mutants in the Col-0 background. (b) Number of rosette leaves at flowering of Ler-0 WT and mutants in the Ler-0 background. (c) Number of rosette leaves at flowering of Ws-4 WT and lhp1-1 mutant. Dark bars denote number of rosette leaves in LDs and light bars number of rosette leaves in SDs. SD conditions (8 h d−1) were achieved using a combination of fluorescent and incandescent light. LD conditions (16 h d−1) consisted of a combination of fluorescent and incandescent light for the first 8 h d−1 and low intensity incandescent light for the 8 h d−1 extension. Data are represented by analysis of 14 plants for each photoperiod treatment. Flowering time is expressed as the number of rosette leaves from 50% of seedling emergence to the appearance of the floral bolt at 1 cm height. Error bars indicate ± standard error of mean (SEM). Values followed by the same letter are not significantly different according to the Student's t-test at a 0.05 level of significance.

The three Arabidopsis WT accessions displayed differences in juvenile phase length (Fig. 2; Table 1). Plants transferred from SDs to LDs while juvenile flowered similarly to plants receiving constant LDs. A linear increase in leaf number and days to flower with successive transfer date can be seen for all the genotypes transferred after the end of juvenile phase (Fig. 2). This illustrates the delay in inflorescence initiation caused by extended time spent in non-inductive SD conditions. Seedlings of Ws-4 and Ler-0 WT were insensitive to photoperiod for longer periods after their emergence than Col-0, which signifies a prolonged juvenile phase. The different type of mutants tested showed different durations in the length of juvenile phase. The starch-deficient mutants adg1 and pgm1 had longer juvenile phase lengths than Col-0 WT (Supporting Information Fig. S1a; Table 1). The starch-excess mutants sex1, sex4 (data not shown) and bam3 exhibited a longer juvenile phase than Col-0 WT (Supporting Information Fig. S1b; Table 1). The glucose insensitive mutants gin1 (aba2), gin2 (hxk1) and gin6 (abi4) had shorter juvenile phases, compared to their respective WT (Supporting Information Fig. S2; Table 1). The floral repressors tfl1, hst1 and lhp1 (tfl2) had a shortened juvenile phase length compared to their respective WT (Supporting Information Fig. S3; Table 1).

Figure 2.

Estimation of the juvenile phase length of Col-0 (●), Ler-0 (○) and Ws-4 (▾) accessions of Arabidopsis based on attainment of floral competence.

Plants were transferred from short days to long days at time intervals shown in x-axes. Points and vertical error bars denote the mean and standard error of the mean (SEM) of leaves at flowering of replicate plants transferred on each occasion. The dashed line delimits the length of juvenile phase. Horizontal error bars denote the SEM of the estimated juvenile phase length.

Table 1. Estimation of the juvenile phase lengths of the Arabidopsis genotypes based on attainment of floral competence
GenotypeFunctionJuvenile phase length (days)
  1. The juvenile phase length is expressed as number of days from 50% of seedling emergence.
  2. The length of juvenile phase is estimated based on attainment of floral competence. Standard error of mean of the estimated juvenile phase length indicated in parenthesis.
Ler-0Wild type5.4 (± 0.3)
tfl1Floral repressor0.3 (± 0.2)
hst1Floral repressor1.2 (± 0.2)
gin1-1 (aba2)Sugar sensor3.0 (± 0.2)
gin2-1 (hxk1)Sugar sensor2.5 (± 0.3)
Col-0Wild type2.0 (± 0.2)
gin6 (abi4)Sugar sensor/ABA1.7 (± 0.2)
pgm1Starch deficient5.0 (± 0.2)
adg1-1Starch deficient5.9 (± 0.4)
sex1Starch excess3.5 (± 0.2)
sex4Starch excess5.0 (± 0.3)
bam3Starch excess4.0 (± 0.2)
Ws-4Wild type7.6 (± 0.4)
lhp1 (tfl2)Floral repressor1.0 (± 0.2)

Juvenile to adult phase transition and carbohydrate relationships in mutants impaired in starch anabolism and catabolism

Having established that mutants involved in starch metabolism-related events have longer juvenile phase lengths compared to WT, led to the conclusion that a starch catabolism-derived signal might be involved in the juvenile-to-adult phase transition. To test this hypothesis, the diurnal metabolite changes in sex1 and adg1 mutant seedlings were determined. Col-0 WT and mutant seedlings were collected under SD and LD conditions on day 9 from emergence when all were in adult phase of development.

Starch progressively accumulated in both Col-0 WT and sex1 mutant as the seedlings age (Supporting Information Fig. S4). In Col-0 WT, photosynthate assimilates were generated in excess of sink demand causing elevated starch accumulation at the end of the light period (Fig. 3a,b). By the end of 16 h−1 light period, Col-0 WT seedlings grown under SD conditions accumulated greater amounts of starch, compared to plants grown in LDs. This is due to low intensity incandescent light provided in the 8 h−1 extension in the LD treatment that is less efficient for photosynthesis, and more efficient for a photoperiodic response. During the dark period, reduced sucrose content triggered a degradation of starch, which was almost fully remobilized by the end of the dark period (Fig. 3a,b). During the scotoperiod, seedlings growing in SD conditions had a slightly faster rate of starch degradation than plants growing in LD conditions. However, at the end of the dark period, similar starch contents were determined in Col-0 WT grown under both photoperiods. In contrast, sex1 mutant showed high starch content throughout the day/night cycle and less diurnal variation under both SD and LD conditions, compared to Col-0 WT.

Figure 3.

Diurnal metabolite changes in Col-0 WT and sex1 mutant genotypes of Arabidopsis.

Diurnal changes of starch (a, b), sucrose (c), glucose (d) and fructose (e) in Col-0 and sex1 mutant genotypes. Col-0 (●) and sex1 (○) seedlings were collected under short days (A) and long days (b, c, d, e) on day 9 from emergence, at time intervals shown in x-axes. All results are the averages of three biological replicates ± standard error of the mean (SEM). White and black bars on the top are subjective day and night, while the grey bar in the LD treatment indicates the photoperiod extension with low intensity incandescent light.

Determination of soluble carbohydrates extracted from sex1 mutant seedlings showed that glucose, sucrose and fructose accumulated in large amounts during the day, relative to Col-0 WT (Fig. 3c,d,e). With the start of 8 h−1 light extension with low intensity incandescent light, soluble carbohydrates were depleted in WT and sex1. However, sucrose accumulation in sex1 mutant seedlings were slightly reduced compared to Col-0 WT sucrose levels, at the end of dark period (Fig. 3c).

No starch accumulated in adg1 mutant seedlings, irrespective of day length (Fig. 4a,b) or developmental phase (data not presented). Compared to Col-0 WT, adg1 mutant seedlings accumulated considerable amounts of sucrose (Fig. 4c,d), glucose (Fig. 4e,f) and fructose (Fig. 4g,h) during the day, rather than being used for biosynthesis, growth and development. During the dark period, soluble carbohydrates were depleted in adg1, in a pattern similar to that of sex1 (Fig. 3c,d,e). Noticeably, at the end of dark period lesser amounts of sucrose remained in adg1 mutant seedling grown under both photoperiods, compared to sucrose levels of Col-0 WT (Fig. 4c,d).

Figure 4.

Diurnal metabolite changes in Col-0 WT and adg1-1 mutant genotypes of Arabidopsis.

Diurnal changes of starch (a, b), sucrose (c, d), glucose (e, f) and fructose (g, e) in Col-0 and adg1-1 genotypes. Col-0 (●), and adg1-1 (○) seedlings were collected under short days (a, c, e, g) and long days (LDs) (b, d, f, h) at time intervals shown in x-axes. All results are the averages of three biological replicates ± standard error of the mean (SEM). White and black bars on the top are subjective day and night, whereas the grey bar in the LD treatment indicates the photoperiod extension with low intensity incandescent light.

Discussion

Estimation of the juvenile phase length of a number of mutants acting in different genetic pathways has led to a model describing a simplified integrated network of pathways that quantitatively control the timing of the juvenile-to-adult phase transition. This model divides the genetic pathways into those that enable the juvenile-to-adult phase transition and those that promote it (Fig. 5).

Figure 5.

A model describing a simplified-integrated network of pathways that quantitatively control the timing of the juvenile-to-adult phase transition in Arabidopsis.

Mutations in different genetic pathways are grouped into those that promote (↓) and those that repress (┴) the juvenile-to-adult phase transition. The enabling pathways regulate the ability of the leaf and meristem to respond to floral promotive signals from different environmental and endogenous cues.

Defining the juvenile phase length in Arabidopsis WT accessions

In this study, estimates on the length of juvenile phase in Col-0, Ler-0 and Ws-4 suggested that Arabidopsis WT accessions differ in the length of the juvenile vegetative phase. Despite the hastened juvenile phase of Col-0, its late flowering phenotype in SD conditions might be attributed to a prolonged photoperiod-sensitive phase. Conversely, in Ws-4, despite its prolonged juvenile phase compared to the other two WT accessions, the early flowering phenotype under LD conditions might be due to a shortened photoperiod-sensitive phase. These indicate that the juvenile and adult phases of plant development can vary independently.

Defining the juvenile phase length in starch-deficient mutants

The developmental differences between WT and starch-deficient mutants of the same chronological age reveal the importance of transitory starch for normal growth and development. The adg1 has no detectable alpha-D-glucose-1-phosphate adenyl transferase (AGP) activity, as it is deficient in the small subunit protein ADP glucose pyrophosphorylase (Lin et al. 1988; Wang et al. 1998). The pgm1 mutant is unable to synthesize starch due to inactivation of the chloroplastic isozyme of PGM (Caspar et al. 1985, 1991). With very low starch levels the rate of growth and net photosynthesis of both mutants, and Col-0, WT are indistinguishable when the genotypes are grown under continuous fluorescent light conditions (data not presented). However, under SD and LD conditions the growth of adg1 and pgm1 is impaired and flowering is significantly delayed compared to Col-0 WT. It has been demonstrated that vernalization completely suppresses the late flowering phenotype of pgm1 (Bernier et al. 1993; Eimert et al. 1995) suggesting that the late-flowering phenotype observed in starch-deficient mutants is not due to the defect in starch accumulation and slow growth rates, but more to their inability to mobilize the stored carbohydrates during the scotoperiod. It has also been demonstrated that maltose and glucose are the two major forms of carbon exported from chloroplasts during the scotoperiod as a result of starch catabolism (Servaites & Geiger 2002; Weise, Weber & Sharkey 2004). Maltose is exported by MALTOSE EXPORTER1 (MEX1; Niittyla et al. 2004), whereas HXK operates as a glucose sensor (Moore et al. 2003). The long juvenile phase length of starch-deficient mutants compared to Col-0 WT provides evidence for the involvement of starch catabolism-related events in the juvenile-to-adult phase transition in Arabidopsis.

Defining the juvenile phase length in starch-excess mutants

Mutants that are unable to catabolize starch provide a valuable tool to study the scotoperiodic effects of carbon exported from chloroplasts on plant developmental transitions. In WT, starch is degraded by phosphorylating enzymes to maltodextrin, which is then, converted to maltose and glucose by BAM and DPE1 in the chloroplast for scotoperiodic export (Lao et al. 1999; Critchley et al. 2001; Scheidig et al. 2002). Reduced activity of glucan water dikinase and SEX4 lead to a reduced rate of starch breakdown and in the accumulation of high levels of starch in sex1 and sex4 mutants, respectively (Kotting et al. 2005, 2009). The bam3 mutant leads also to a starch-excess phenotype (Lao et al. 1999). Mutants with starch-excess phenotypes were late flowering compared to WT, flowering later in SD than they do in LD conditions. Furthermore, sex1, sex4 (data not shown) and bam3 mutants displayed prolonged periods of photoperiod insensitivity after seedling emergence, signifying longer juvenile phase lengths than Col-0. However, the longer juvenile phase length of starch-excess mutants, compared to Col-0 provides a further piece of evidence for the involvement of starch catabolism-related events in the transition within the vegetative phase in Arabidopsis. It is plausible that plants in the juvenile phase may require starch accumulation to reach a threshold level, in order to sustain a steady supply of maltose and/or sucrose during the scotoperiod to undergo the juvenile-to-adult phase transition. This is supported by observations in development of mex1 and dpe-1 mex-1 double mutant. Both mutants are very small and pale and under normal growth conditions, often fail to reach a mature developmental state (Niittyla et al. 2004). However, this severe phenotype can only partially be rescued by supplying both mutants with sucrose (Stettler et al. 2009).

Defining the juvenile phase length in mutants involved in carbohydrate-hormone interactions

Arabidopsis mutants showing sugar insensitive phenotypes represent a valuable tool in unravelling sugar-response pathways affecting developmental transitions in plants. The exploitation of gin1 (aba2) mutant in photoperiod transfer experiments revealed an early flowering phenotype and hastened juvenile phase length, compared to Ler-0. GLUCOSE INSENSITIVE1 (ABA2) encodes a unique short-chain dehydrogenase/reductase that is required for ABA synthesis. A phenotype similar to gin1 (aba2) has been determined in the gin2 (hxk1) mutant. GLUCOSE INSENSITIVE2 (HXK1) encodes an HXK that functions as a glucose sensor to integrate nutrient, light intensity and hormone-signalling systems for controlling plant development in response to environmental conditions (Moore et al. 2003). The gin6 (abi4) mutant in the Col-0 background contains a T-DNA insertion in the promoter of the At2g40220 locus, which encodes an APETALA-2 domain TF (Finkelstein et al. 1998; Arenas-Huertero et al. 2000). The short juvenile phase length of gin6 and other glucose insensitive mutants demonstrates the involvement of these loci in the juvenile-to-adult phase transition in Arabidopsis. Noticeably, it has been shown that in addition to GIN6, several mutants insensitive to sucrose are allelic to abi4 (Huijser et al. 2000; Rook et al. 2001). This might disclose the tight interplay between sugar and ABA phytohormone pathway in the regulation of the juvenile-to-adult phase transition through multiple pathways (Choi et al. 2000; Domagalska et al. 2010).

Defining the juvenile phase length in mutants acting as floral repressors

Floral incompetence during the juvenile phase has led to the hypothesis that the underlying mechanism of juvenility may involve activities of strong floral repressors based at the leaf or at SAM. Terminal flower1 (tfl1), lhp1 (tfl2) and hst1 mutants were shown to flower early under non-inductive SD conditions compared to their respective WT. Assessment of juvenility in tfl1, hst1 and lhp1 (tfl2) seedlings showed they had shortened juvenile phase lengths compared to their respective WT accessions, which is significantly longer. This provides evidence for the involvement of these three loci in the juvenile-to-adult phase transition in Arabidopsis. Genetic and molecular approaches have identified the functions of TFL1, HST1 and LHP1 (TFL2). Transcripts of these genes are detected in all plant tissues (Schmid et al. 2005). However, it has been shown that they mainly act at the SAM by regulating the FMI genes.

HASTY1 (HST1) is the Arabidopsis ortholog of the miRNA nuclear export receptor EXPORTIN5 (Bollman et al. 2003). HST1 was isolated in a screen for mutations that accelerated the vegetative phase change with respect to leaf morphological traits. Functional analysis of hst mutations revealed a role for miR156 and miR172 in synchronization and harmonization of the juvenile-to-adult phase transition (Wu & Poethig 2006; Chuck et al. 2007; Jung et al. 2007; Mathieu et al. 2007). It has been shown that the levels of miR156 and miR172 exhibit contrasting expression patterns (Chuck et al. 2007; Wu et al. 2009). As age proceeds, the decline of miR156 levels, and the increase in levels of miR172 and certain SPL genes, leads to the activation of FT in leaves, whereas the increase in SPLs in the meristem leads to the activation of FPIs and FMI genes that promote the transition to flowering (Wang et al. 2009; Yamaguchi et al. 2009).

TFL1 has been demonstrated to function as a signal to coordinate shoot meristem identity by regulating the FMI genes LFY and AP1 (Ratcliffe, Bradley & Coen 1999; Conti & Bradley 2007). Further support for the involvement of TFL1 in the vegetative phase transition is derived from the study of Bradley et al. (1997) on inflorescence commitment in Arabidopsis. By applying LD to SD transfer of seedlings, the juvenile phase length of tfl1 was shortened compared to Ler-0 WT.

LIKE HETEROCHROMATIN PROTEIN1 (LHP1/TFL2), functions as a negative regulator during the vegetative-to-reproductive phase transition by repressing the expression of FT, but with no effect on expression of the other FPIs (Gaudin et al. 2001; Kotake et al. 2003). It has been demonstrated that HP1, with which LHP1 shares homology, maintains genes in a transcriptionally inactive state by remodelling chromatin structure in the heterochromatin region (Nakahigashi et al. 2005). It is plausible that not only the activators but also the repressors are required for the precise synchronization and harmonization of the juvenile-to-adult phase transition in Arabidopsis. Taken together, these results suggest that regulation of juvenility might be through repression of FPIs transcription by antiflorigenic molecules such as TFL1, HST1 and LHP1 (TFL2).

Juvenile-to-adult phase transition and carbohydrate relationships

Starch catabolism-related events might be the cause of the prolonged juvenile phase in starch-deficient and starch-excess mutants. It has been proposed that the inhibition of growth in starch-deficient mutants is primarily caused by a disturbance of metabolism and growth, which is triggered by a transient period of sugar depletion during the scotoperiod (Bernier et al. 1993). Furthermore, it has been demonstrated that the expression of hundreds of genes is altered in the pgm1 starch-deficient mutant at the end of the dark period, compared to WT at the same time (Thimm et al. 2004). This includes many genes that are required for nutrient assimilation, biosynthesis and growth. Taken together, these results suggest that soluble carbohydrate depletion during the dark period leads to marked changes in gene expression stimulating an inhibition of carbohydrate utilization, which could directly affect the juvenile-to-adult phase transition. However, in addition to transitory starch providing a source of carbon for growth during the following night (Thimm et al. 2004) and for the beginning of the next light period, it may also act as an overflow for newly assimilated carbon (Stitt & Quick 1989), when assimilation exceeds the demand for sucrose. This mechanism is inactivated in the adg1 mutant, as demonstrated by the elevated soluble carbohydrate levels, and hardly any starch at the end of the light period, in adg1 mutant seedlings.

The sex1 mutant is known for its impaired ability to catabolize starch (Caspar et al. 1991; Yu et al. 2001). The lack of starch turnover in sex1 has an influence on general carbohydrate availability, reducing the amount of sucrose and maltose contents (Chia et al. 2004; Niittyla et al. 2004) at the end of scotoperiod. The importance of temporal availability of maltose in the regulation of plant growth has already been demonstrated (Niittyla et al. 2004; Stettler et al. 2009). Furthermore, despite the fact that sex1 and adg1 mutants being impaired in different genetic pathways their metabolism and growth inhibition might be triggered by a transient period of soluble carbohydrate depletion during the scotoperiod. It is possible that both mutants function in the same physiological pathway controlling juvenility. As with adg1, it is possible that in sex1 carbohydrate depletion during the dark leads to critical changes in gene expression stimulating an inhibition of carbohydrate utilization with direct effects on the length of the juvenile phase.

A number of physiological, biochemical and molecular approaches have shown that early growth and development in Arabidopsis seedlings can be arrested in the presence of high glucose and sucrose levels. Several developmental characteristics are subject to high-level soluble carbohydrate repression (Jang et al. 1997; Arenas-Huertero et al. 2000; Gibson 2000; Gazzarrini & McCourt 2001). The physiological rationale for soluble carbohydrate repression during the early phases of plant development could be that elevated soluble carbohydrate accumulation levels reflect suboptimal growth conditions (Lopez-Molina, Mongrand & Chua 2001). Therefore, inhibition of developmental programmes such as the juvenile-to-adult phase transition in starch-deficient and starch-excess mutants may result from the activation of soluble carbohydrate repression events. Based on this repression of growth response, a series of gin mutants have been isolated (Zhou et al. 1998; Moore et al. 2003). The finding that the gin mutants have a shortened juvenile phase and an early flowering phenotype further supports this hypothesis.

The data presented shows that a variety of signals act to promote and enable the juvenile-to-adult phase transition that involves both floral activators and repressors. Starch metabolism is involved in the juvenile-to-adult phase transition. Carbohydrates might be involved through their function as nutrients, osmotic regulators and signalling molecules, and/ or by their interaction with phytohormonal networks.

Funding

This work was supported by the Hellenic State Scholarships Foundation (IKY) and UK Department for Environment, Food and Rural Affairs (DEFRA).

Acknowledgments

We would like to thank Professor Samuel Zeeman (ETH, Zurich) for providing the sex4 and bam3 mutants, Professor Jen Sheen (Harvard Medical School) for providing the gin mutants and Professor Valérie Gaudin (INRA, Versailles) for providing the lhp1 mutant.

Ancillary