The interaction between glucose and cytokinin signal transduction pathway in Arabidopsis thaliana

Authors


  • Research Area: Physiology and Development
  • Financial sources: This work was financially supported by the National Institute of Plant Genome Research (NIPGR) core grant, Department of Biotechnology, Government of India (Grant No. BT/PR14398/BRB/843/2010) and Department of Biotechnology, Government of India research fellowship to S.K.

Abstract

Cytokinins (CKs) and glucose (GLC) control a number of common responses in plants. We hypothesize that there may be an extensive overlap between CK- and GLC-signalling pathways. Microarray along with physiological analysis has been performed to find out the interdependence/overlap between CK and GLC signal transduction pathways in Arabidopsis seedlings. GLC could transcriptionally affect 76% of CK-regulated genes at whole genome level, 89% of which are agonistically regulated. GLC may also affect CK-regulated gene expression via non-transcriptional pathways. GLC can regulate several genes involved in CK metabolism and signalling. A number of gene families involved in development and stress are commonly regulated by CK and GLC. Physiologically, both GLC and CK could regulate hypocotyl length in dark. GLC and CK signalling may integrate at the level of type A Arabidopsis response regulators (ARRs) in controlling hypocotyl length. Both GLC and CK signalling cannot alter hypocotyl length in dark in auxin-signalling mutants AUXIN RESPONSE2/INDOLE-3-ACETIC ACID7 (AXR2/IAA7) and AXR3/IAA17 suggesting that they may involve auxin-signalling component as a nodal point. Here, we demonstrate that there is an extensive overlap between CK- and GLC-regulated gene expression and physiological responses.

Introduction

Glucose (GLC) as a signalling molecule influences every aspect of plant growth and development including, cell proliferation and death, cell expansion and elongation, seed germination, seedling growth and development, primary root length, root gravitropism, lateral roots, root hairs, shoot meristem maintenance, reproduction, senescence, photosynthetic gene expression, crop yield and product quality, carbon and nitrogen metabolism and stress responses (Rolland, Moore & Sheen 2002; Rolland & Sheen 2005; Chen, Gao & Jones 2006; Rolland, Baena-Gonzalez & Sheen 2006; Ramon, Rolland & Sheen 2008; Smeekens et al. 2010; Eveland & Jackson 2012). There are three distinct GLC-signalling pathways in plants viz., (1) AtHXK1 (HEXOKINASE1) signalling function-dependent, (2) AtHXK1 catalytic function-dependent and (3) AtHXK1-independent pathway (Xiao, Sheen & Jang 2000). In AtHXK1 signalling function-dependent pathway, HXK1 acts as a GLC sensor. To determine the role of HXK1 in plant physiology, Moore et al. (2003) isolated Arabidopsis HXK loss-of-function mutant glucose insensitive2-1 (gin2-1) that retained the catalytic activities but not the signalling function. Catalytically inactive HXK1 (G104D and S177A) mutants displayed similar GLC-signalling function as the wild-type (WT; Moore et al. 2003). Transgenic plants that expressed the catalytically inactive HXK1 alleles in the gin2-1 null mutant background restored GLC-dependent phenotypes (Moore et al. 2003). This suggests that the role of HXK1 in sugar signalling is independent of its conventional role in metabolism. The AtHXK1-independent pathway involves G-protein signalling (Assmann 2002, 2004; Jones 2002; Jones, Ecker & Chen 2003; Jones & Assmann 2004; Perfus-Barbeoch, Jones & Assmann 2004; Joo et al. 2005; Chen et al. 2006; Temple & Jones 2007). The G-protein complex is represented by a single Gα [G-PROTEIN ALPHA SUBUNIT1 (GPA1)], Gβ [G-PROTEIN BETA SUBUNIT1 (AGB1)] and Gγ [G-PROTEIN GAMMA SUBUNIT1 (AGG1) and AGG2] subunits. The AtRGS1 (regulator of G-protein signalling1) contained a seven transmembrane domain and RGS box; and accelerates intrinsic GTPase activity of G-protein. The THF1 (THYLAKOID FORMATION1) protein located on the outer membrane and stroma of plastids interacts with GPA1 in root hair cells when plasma membrane and plastid membrane are in close contact. The heterotrimeric G-protein complex via GPA1 interacts with the AtRGS1 and THF1 (Huang et al. 2006; Grigston et al. 2008).

Cytokinins (CKs) are adenine-derived essential phytohormones. CKs are involved in a myriad of processes such as cell division and expansion, embryogenesis, vascular morphogenesis, de novo organogenesis, root and shoot development, lateral root formation, leaf morphology and senescence, chloroplast and anthocyanin biogenesis, stress responses, immunity, nodule organogenesis and photomorphogenesis (Kieber & Schaller 2010; Müller 2011; Brenner et al. 2012; Gupta & Rashotte 2012; Hwang, Sheen & Müller 2012; Shi & Rashotte 2012; Spíchal 2012). In Arabidopsis, CK signalling involves multistep phosphorelay in the following order of phosphorylation, (1) hybrid His protein kinases [ARABIDOPSIS HISTIDINE KINASE2 (AHK2), AHK3, AHK4] that serve as CK receptors, (2) His phosphotransfer proteins [ARABIDOPSIS HISTIDINE PHOSPHOTRANSFER PROTEIN1 (AHP1), AHP2-AHP6] and (3) type A ARABIDOPSIS RESPONSE REGULATORS (ARR3-ARR9, ARR15-ARR17) and type B response regulators (ARR1, ARR2, ARR10-ARR14, ARR18-ARR21; Hwang & Sheen 2001). After phosphorylation, AHPs are translocated into the nucleus, where they phosphorylate type B and type A ARRs. Phosphorylated type B ARRs act as transcription factors and induce the transcription of type A ARRs, CYTOKININ RESPONSE FACTORS (CRFs) and other CK early-responsive genes. Type B ARRs act as positive regulators, whereas type A ARRs functions as negative regulators of CK-signalling pathway (Hwang & Sheen 2001).

There are several reports indicating that sugars and CK can control similar responses. Our group has recently reported a novel CK-induced root directional response, which is enhanced in the presence of GLC (Kushwah, Jones & Laxmi 2011). CK signalling modulates sugar-induced anthocyanin biosynthesis through a two-component signalling cascade via regulating sugar-inducible structural and regulatory genes (Das et al. 2012). Sucrose is able to induce expression of cyclin D3 (CycD3) alone and in combination with CK. However, CK in the absence of sucrose does not induce CycD3 expression (Riou-Khamlichi et al. 1999, 2000). This suggests that sucrose acts upstream of CK or synergistically with CK in regulating CycD3 expression (Riou-Khamlichi et al. 2000; Hartig & Beck 2006).

Sugars and CK can also work antagonistically to each other. The delayed leaf senescence phenotype found in glucose receptor gin2 mutant links sugar and CK signalling in antagonistic manner (Moore et al. 2003). In tissue culture, gin2 mutants are hypersensitive to CKs for shoot regeneration in comparison to the WT (Moore et al. 2003). Plants constitutively active in CK signalling [CYTOKININ-INDEPENDENT1 (CKI1) and ARR2] displayed decreased sugar sensitivity, further confirming antagonistic interaction between CK and GLC signalling (Moore et al. 2003). CK-insensitive mutants CYTOKININ RESISTANT1 (cre1), ahk3, cre1ahk3 are hypersensitive, whereas plants ectopically expressing AHK3 are more resistant to high sucrose concentrations suggesting antagonistic interaction between CK and sucrose (Franco-Zorrilla et al. 2005).

The hypersenescence1 (hys1) mutant hypersensitivity for the inhibitory effects of GLC upon seedlings greening is HXK1 dependent, which cannot be prevented by the antagonistic activities of ethylene and CK. This suggests that GLC and CK can also work independent of each other (Aki et al. 2007).

These reports suggest that for many responses, GLC and CK act agonistically, whereas for others, they act antagonistically and for some responses, they act independently of each other. To date, there have been no systematic studies that explore the global effect of GLC and CK signalling on gene expression and in turn on plant growth and development. In this study, whole genome transcript profiling along with physiological analysis has been performed to find out the interdependence/overlap between GLC and CK signal transduction pathways in Arabidopsis.

Materials and Methods

Plant materials

Following seed stocks were procured from the Arabidopsis Biological Resource Center at Ohio State University: ahk4 (At2g01830, CS6563); arr1,10,11 (At3g16857/At4g31920/At1g67710, CS6993); arr3,4,5,6,8,9 (At1g59940/At1g10470/At3g48100/At5g62920/At2g41310/At3g57040, CS25279); tir1-1 (At3g62980, CS3798); axr1-3 (At1g05180, CS3075); axr3-1 (At1g04250, CS57504), eir1-1 (At5g57090, CS8058); gin2-1 (At4g29130, CS6383); pARR5::GFP (At3g48100, CS23893); pARR6::GUS (At5g62920, CS25262). The axr2-1 (At3g23050, CS3077) and HS::AXR3NT-GUS (At1g04250, N9572) seed stocks were obtained from the European Arabidopsis Stock Centre. The following lines were obtained from the original published source as: pin3-4 (At1g70940) (Friml et al. 2002a); pin4-3 (At2g01420) (Friml et al. 2002b); pin7-2 (At1g23080) (Friml et al. 2003); mdr1-101 (At3g28860) (Lin & Wang 2005); pgp1-100 (At2g36910) (Lin & Wang 2005); rgs1-1 and rgs1-2 (At3g26090) (Chen et al. 2003); gpa1-1, gpa1-2 and gpa1-3 (At2g26300) (Ullah et al. 2001); thf1-1 (At2g20890) (Huang et al. 2006). The slr-1 (At4g14550) (Fukaki et al. 2002) mutant line was generously provided by Dr Hidehiro Fukaki at Nara Institute of Science and Technology, Nara, Japan. All mutant lines were in Col background except the following: ahk4; arr1,10,11; gpa1-1; gpa1-2 were derived from Ws background. The gin2-1 mutant was in the Ler background. All chemicals were purchased from Sigma (St Louis, MO, USA) except agar, which is purchased from Himedia (Mumbai, India). The 6-benzylaminopurine (BAP) was prepared as 10−2 m stock solution in dimethyl sulfoxide (DMSO). 5-Bromo-4-chloro-3-indolyl-β-D-glucuronic acid (X-Gluc) was prepared as 100 mgL−1 stock solution in N, N-dimethylformamide (DMFO).

Seedling growth

Seeds were surface sterilized and imbibed at 4 °C for 48 h. For dark-grown seedlings, imbibed seeds were directly sown on square (120 × 120 mm) Petri plates containing half-strength Murashige and Skoog (MS) medium (MS, 2.2 gL−1; MES, 0.5 gL−1) supplemented with different concentrations of GLC (0, 1, 3, 5%; w/v), BAP (0 m, 10−7 m, 5 × 10−7 m, 10−6 m) and 0.8% agar (w/v; pH 5.7). Hormone and solvent were added after autoclaving. Thereafter, seeds on plates were first exposed to 16 h light to stimulate germination; subsequently, the plates were wrapped with aluminium foil and placed vertically in the growth chamber (22 °C ± 2 °C temperature).

For light-grown seedlings, seeds were directly sown on square (120 × 120 mm) Petri plates containing half-strength MS medium supplemented with 1% sucrose and 0.8% agar (w/v; pH 5.7). Seed germination was carried out in climate-controlled growth room under long day conditions (16 h light and 8 h darkness, 80 μmol m−2 s−1 light intensity) at 22 °C ± 2 °C temperature. Afterwards, 5-day-old seedlings were transferred on square (120 × 120 mm) Petri plates containing half-strength MS medium supplemented with different concentrations of GLC (0, 1, 3, 5%; w/v), BAP (0 m, 10−7 m, 5 × 10−7 m, 10−6 m; pH 5.7) and 0.8% agar (w/v; pH 5.7). The root tips were marked at the moment of seedling transfer. Thereafter, seedlings were grown vertically for next 3 d in growth chambers. In all experiments, plates were sealed with gas-permeable tape to avoid ethylene accumulation.

Measurement of hypocotyl and root length

All end point analyses were taken on the eighth day. For hypocotyl and root length, digital images were captured using Nikon Coolpix (Nikon Corporation, Tokyo, Japan) digital camera. Hypocotyl and root length was quantified using ImageJ (http://rsb.info.nih.gov/ij/). The length represents the average and error bars represent the standard deviation. The experiment was repeated at least thrice. For all experiments, Student's t-test with paired two-tailed distribution was used for statistical analysis. The panel represents hypocotyl growth of one of the three biological replicates yielding similar results.

Chlorophyll and anthocyanin estimation

For chlorophyll and anthocyanin estimation seeds were directly sown on square (120 × 120 mm) Petri plates containing half-strength MS medium supplemented with different concentrations of GLC (0, 1, 3, 5%; w/v), BAP (0 m, 10−7 m, 5 × 10−7 m, 10−6 m; pH 5.7) and 0.8% agar (w/v; pH 5.7). Plates were kept vertically in climate-controlled growth room under long day conditions (16 h light and 8 h darkness, 80 μmol m−2 s−1 light intensity) at 22 °C ± 2 °C temperature for 8 d. The amount of chlorophyll and anthocyanin was estimated as mentioned by Laxmi et al. (2006).

The seedlings were harvested and their fresh weight was determined. Chlorophyll was extracted by keeping the seedlings in dark in 2 mL 80% (v/v) acetone for overnight at 4 °C. The following day, the cell debris was pelleted by centrifugation at 8000 × g at 4 °C for 10 min. The amount of chlorophyll was determined by measuring the absorbance of the supernatant at 663 and 646 nm by using SmartSpec™ Plus spectrophotometer (Bio-Rad, Hercules, CA, USA). The amount of chlorophyll a (12.21xλ663-2.81xλ646) and chlorophyll b (20.13xλ646-5.03xλ663) was quantified. The total amount of chlorophyll (chlorophyll a + chlorophyll b) was normalized to per gram fresh weight. Anthocyanin was extracted by keeping the seedlings in 3 mL 1% (v/v) acidic methanol for overnight at 4 °C. The phase partitioning was done in the second day by adding 3 mL chloroform and 2 mL RO water. Absorbance of aqueous phase was determined at 530 and 657 nm by using SmartSpec™ Plus spectrophotometer (Bio-Rad). The total amount of anthocyanin (λ530–λ657) estimated was normalized to per gram fresh weight. The experiment was repeated at least thrice, yielding similar results. Data shown is the average of two biological replicates each with three technical replicates having at least 20 seedlings and error bars represent standard deviation. For all experiments, Student's t-test with paired two-tailed distribution was used for statistical analysis.

GUS histochemical staining and fluorometry

Seeds of pARR6::GUS lines were directly sown on half-strength MS medium supplemented with 0% GLC, 0% GLC + 1 μm BAP, 3% GLC and 3% GLC + 1 μm BAP, 0.8% agar and grown vertically in culture room conditions for 7 d. The 7-day-old HS::AXR3NT-GUS seedlings grown in half-strength MS medium containing 1% sucrose and 0.8% agar were heat-shocked at 37 °C for 2 h followed by recovery for 30 min to induce HS::AXR3NT-GUS reporter. Seedlings were subsequently incubated at culture room temperature in the 0% GLC-, 0% GLC + 5 μm BAP-, 0% GLC + 10 μm BAP-, 3% GLC-, 3% GLC + 5 μm BAP- and 3% GLC + 10 μm BAP-containing liquid half-strength MS medium for 1 h. GUS activities were then determined by incubating the seedlings at 37 °C in a GUS staining solution [sodium phosphate buffer pH 7.0, 0.1 m; K3Fe(CN)6, 0.5 mm; K4Fe(CN)6, 0.5 mm; EDTA, 50 mm; X-Gluc, 1 mg mL−1] for 3 to 4 h. The seedlings were then kept in 70% ethanol for the removal of chlorophyll. The seedlings were then observed under Nikon SMZ1500 Stereo-Zoom microscope and photographs were taken by Nikon Coolpix digital camera connected with Nikon SMZ1500 Stereo-Zoom microscope. The experiment was repeated twice with each replicate having 10 seedlings, yielding similar results.

GUS enzyme activity in HS::AXR3NT-GUS seedlings was determined according to the method of Jefferson (1987). Seedlings were homogenized in extraction buffer (sodium phosphate buffer pH 7.0, 50 mm; β-Mercaptoethanol, 10 mm; Na2EDTA, 1 mm; Sodium lauryl sarcosine, 0.1%; Triton X-100, 0.1%). Protein was quantified using Bradford assay and data was normalized against total protein level. Samples were assayed using 1 mm 4-methylumbelliferyl-β-D-glucoronide (MUG) in extraction buffer followed by 1 h incubation. The reaction was stopped by adding 0.2 m Na2CO3 and the fluorescence was measured with a fluorescence spectrophotometer (Varian Cary Eclipse Varian Optical Spectroscopy Instruments, Mulgrave,Victoria, Australia).

Laser confocal scanning microscopy (LCSM)

The 5-day-old pARR5::GFP seedlings were treated with 0% GLC-, 0% GLC + 1 μm BAP-, 3% GLC- and 3% GLC + 1 μm BAP-containing liquid half-strength MS medium for 3 h. GFP fluorescence in root sections was observed using the TCS SP2 (AOBS) Laser Confocal Scanning Microscope (Leica Microsystems, Heidelberg, Germany) and images were captured. For imaging GFP, the 488 nm line of the argon laser was used for excitation and emission was detected at 520 nm. The laser and pinhole settings of the confocal microscope were kept identical among different treatments. The experiment was repeated at least thrice, yielding similar results. The data represents one of the biological replicates with 10 seedlings in each replicate.

Gene expression analysis: microarray and quantitative real-time PCR

The growth conditions used here for microarray and real-time PCR are already described by Price et al. (2004) and Mishra et al. (2009) involving certain modifications. For gene expression studies, seeds were sown on half-strength MS medium supplemented with 1% sucrose, 0.8% agar (w/v) and grown vertically in culture room conditions. The 5-day-old homogenous seedlings were washed five times with sterile water and lastly with liquid half-strength MS medium without sugar to remove residual exogenous sugar. In order to deplete internal sugars, seedlings were grown in sugar-free liquid half-strength MS medium for 24 h in dark. Thereafter, seedlings were incubated with treatment medium as mentioned separately for 3 h in dark. For microarray and validation of microarray data, seedlings were treated with 0% GLC-, 0% GLC + 1 μm BAP-, 3% GLC- and 3% GLC + 1 μm BAP-containing liquid half-strength MS medium. To determine the basal level of CK- and GLC-signalling genes in GLC- and CK-signalling mutants, respectively, seedlings were treated with liquid half-strength MS medium without sugar. All the above mentioned steps were carried out in dark and the cultures were shaken at 140 r.p.m., 22 °C ± 2 °C. Afterwards, seedlings were flash frozen in liquid nitrogen and the mRNA was prepared from frozen tissue using the RNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA) following the manufacturer's protocol.

Three biological replicates were used for performing microarray analysis. Labelling of RNA probe and hybridization to Arabidopsis Gene Chip were conducted using standard Affymetrix protocols by the University of California, Irvine DNA MicroArray Facility. Isolated total RNA samples were processed as recommended by Affymetrix, Inc. (Affymetrix GeneChip Expression Analysis Technical Manual, Affymetric, Inc., Santa Clara, CA, USA). Eluted total RNAs were quantified with a portion of the recovered total RNA adjusted to a final concentration of 1 μg μL−1. All starting total RNA samples were quality assessed prior to beginning target preparation/processing steps by running out a small amount of each sample (typically 25–250 ng well−1) onto a RNA Lab-On-A-Chip (Caliper Technologies Corp., Mountain View, CA, USA) that was evaluated on an Agilent Bio-analyzer 2100 (Agilent Technologies, Palo Alto, CA, USA). Single-stranded, then double-stranded cDNA was synthesized from the poly(A) + mRNA present in the isolated total RNA (5.0 μg total RNA starting material each sample reaction) using the SuperScript Double-Stranded cDNA Synthesis Kit (Invitrogen Corp., Carlsbad, CA, USA) and poly (T)-nucleotide primers that contained a sequence recognized by T7 RNA polymerase. A portion of the resulting ds cDNA was used as a template to generate biotin-tagged cRNA from an in vitro transcription reaction (IVT), using the BioArray High-Yield RNA Transcript Labeling Kit (T7; Enzo Diagnostics, Inc., Farmingdale, NY, USA). The 15 μg of the resulting biotin-tagged cRNA was fragmented to strands of 35–200 bases in length following prescribed protocols (Affymetrix GeneChip Expression Analysis Technical Manual). Subsequently, 10 μg of this fragmented target cRNA was hybridized at 45 °C with rotation for 16 h (Affymetrix GeneChip Hybridization Oven 640) to probe sets present on an Affymetrix (Arabidopsis ATH1 genome) array. The GeneChip arrays were washed and then stained (SAPE, streptavidin-phycoerythrin) on an Affymetrix Fluidics Station 450, followed by scanning on an Affymetrix GeneChip 3000 Scanner 7G. The results were quantified and analysed using Expression Console ver.1.1 software (Affymetrix, Inc.) using the PLIER algorithm default values (quantification scale: linear; quantification type: signal and detection P-value; background: PM-GCBG; normalization method: sketch-quantile). Cyber-T comparison were performed at PPDE(p) ≥ 0.995 and greater than or equal to twofold (+/−). Additional microarray data presentation and manipulation were assessed using Microsoft Excel. All data is MIAME (Minimum Information About a Microarray Experiment) compliant and the raw data has been deposited in Array Express database through MIAMExpress (accession number E-MEXP-3827).

For real-time PCR, first-strand cDNA was synthesized by reverse transcription using 4 μg of total RNA in 40 μL of reaction volume using high-capacity cDNA Reverse Transcription kit (Applied Biosystems, Carlsbad, CA, USA). Diluted cDNA samples (1:20 dilution) were used for quantitative real-time PCR analysis, 5 μm of each primer mixed with SYBR Green PCR master mix as per the manufacturer's instructions. Primers for all the candidate genes were designed preferentially from the 3’ end of the gene using PRIMER EXPRESS version 3.0 (PE Applied Biosystems) with default parameters. The reaction was carried out in 96-/384-well optical reaction plates (Applied Biosystems) using ABI 7900HT Fast Real-time PCR System (Applied Biosystems). To normalize the variance among samples, 18S rRNA and ubiquitin10 (UBQ10) were used as the endogenous control. The mRNA levels for each candidate gene in different samples were determined using the ΔΔCT method (Livak & Schmittgen 2001). The values represent the average of the two biological replicates (each with three technical replicates) and error bars represent SE. For all experiments, Student's t-test with paired two-tailed distribution was used for statistical analysis. The primer sequences for all the genes tested have been included in Supporting Information Table S1.

Results

The GLC and CK interaction at whole genome level: microarray analysis

Effect of GLC on BAP-regulated genes

Since a number of phenotypes are simultaneously controlled by GLC and CK, whole genome transcript profiling was performed to investigate the extent of overlap at the gene expression level. BAP could regulate 941 genes (Supporting Information Fig. S1), which include 625 (66%) up-regulated (Supporting Information Fig. S2) and only a small number of down-regulated 316 (34%) genes (Supporting Information Fig. S3; Fig. 1a). Out of 941 BAP-regulated genes, 713 (76%) genes were affected by GLC treatment alone (Fig. 1b; Supporting Information Fig. S4). Out of these 713 genes, 633 (89%) genes were agonistically (Supporting Information Fig. S5) and 80 (11%) genes were antagonistically regulated by BAP and GLC treatment alone (Supporting Information Fig. S6; Fig. 1b). Thus, at the whole genome level, GLC and CK largely act agonistically.

Figure 1.

Effect of GLC alone on BAP-regulated genes. (a) BAP could regulate 941 genes, which include 625 (66%) up-regulated and 316 (34%) down-regulated genes. (b) Out of 941 BAP-regulated genes, 713 (76%) genes were affected by GLC treatment alone. Out of these 713 genes, 633 (89%) genes were agonistically and 80 (11%) genes were antagonistically regulated by BAP and GLC treatment alone.

Effect of GLC on extent of regulation of BAP-affected genes

The effect of GLC on the extent of regulation of BAP-affected genes was examined. Out of 941 BAP-regulated genes (0% GLC versus 0% GLC + BAP), the extent of 699 (74%) genes (0% GLC versus 0% GLC + BAP as compared to 0% GLC versus 3% GLC + BAP) was significantly affected (twofold or more/less or lost) in presence of GLC (Fig. 2, Supporting Information Fig. S7). Out of these 699 genes, the extent of 490 (70%) genes was agonistically (BAP up-regulation and down-regulation increased) and only of 209 (30%) genes was antagonistically (BAP up-regulation and down-regulation decreased or lost) affected in presence of GLC (Fig. 2b, Supporting Information Figs S8 & S9). Among the agonistically (490) and antagonistically (209) affected genes, 480 (98%) and only 87 (42%) genes, respectively, were transcriptionally affected by GLC alone (Fig. 2b). Thus, majority of agonistically affected genes were transcriptionally, while a large number of antagonistically affected genes were non-transcriptionally regulated by GLC alone. To explore if GLC and CK can affect protein stability, HS::AXR3NT-GUS transgenic line was used. In this line, the promoter of auxin inducible gene AUXIN RESPONSE3 (AXR3)/INDOLE-3-ACETIC ACID17 (IAA17) has been replaced with heat shock promoter and the entire construct has been fused with GUS. This line has been used previously to find out the effect of various signals/treatments on AXR3 protein stability after its induction by heat shock (Gray et al. 2001; Mishra et al. 2009; Quint et al. 2009; Savatin et al. 2011). The AXR3 protein was found to be degraded on giving BAP treatment but the presence of GLC could stabilize it (Fig. 3a,b) which is also consistent with previous study by Mishra et al. (2009). This suggests GLC and CK can also interact via non-transcriptional pathways.

Figure 2.

Effect of GLC on extent of regulation of BAP affected genes. (a) Out of 941 BAP up- and down-regulated genes the extent of 473 (76%) and 226 (72%) genes, respectively, was significantly affected in presence of GLC. (b) Out of 941 BAP-regulated genes, the extent of 699 (74%) genes was significantly affected in presence of GLC. Out of these 699 genes, the extent of 490 (70%) genes was agonistically and only of 209 (30%) genes was antagonistically affected in presence of GLC.

Figure 3.

CK- and GLC-regulation of AXR3 protein stability. (a) The AXR3 protein was degraded faster in the presence of CK while presence of GLC stabilized it as suggested by histochemical GUS assay. (b) The stabilizing effect of GLC was also supported by quantitative GUS assay.

Quantitative real-time PCR validation of microarray results

Validation of microarray data was done by conducting quantitative real-time PCR of some well-known CK-regulated genes and an additional set of selected genes. The expression of CK–related genes AHK1, AHK2, AHK4, ARR2, ARR3, ARR4, ARR5, ARR6, ARR16, ATP/ADP-ISOPENTENYLTRANSFERASE 3 (IPT3), CYTOKININ OXIDASE 4 (CKX4) was checked (Fig. 4a–c). The expression of some auxin- and brassinosteroid (BR)-regulated genes, few expansin and xyloglucan:xyloglucosyl transferase gene family members and one expressed protein with unknown function was also studied (Supporting Information Fig. S10a). In addition to this, some GLC-regulated genes were also tested (Supporting Information Fig. S10b). We have also repeated real-time PCR of some candidate genes with another reference gene UBQ10 (Supporting Information Fig. S10c). The microarray data was further supported by pARR6::GUS (Fig. 4d) and pARR5::GFP (Fig. 4e) expression pattern which was up-regulated by CK and agonistically regulated by CK and GLC both. The expression pattern of representative genes from microarray data, which were picked for validation, was shown in Supporting Information Fig. S10d.

Figure 4.

Validation of microarray results. (a–c) The relative expression of a few well-known CK-regulated genes as revealed by real-time PCR. (d) The expression of ARR6 gene as revealed by pARR6::GUS line. (e) The expression of ARR5 gene as revealed by pARR5::GFP line.

Effect of GLC on genes involved into CK metabolism and signalling

A number of genes involved in CK metabolism and signalling were also observed to be regulated by GLC (Table 1). GLC also affects those genes that were themselves transcriptionally not regulated by CK itself (Table 1). CK biosynthesis gene IPT3 was up-regulated by GLC. AtIPT3 is located in the plastids, expressed in vegetative organs and in the vasculature throughout the plant and up-regulated by nitrate (Hwang & Sakakibara 2006; Hirose et al. 2008). CK catabolism-related gene CKX4 was up-regulated, whereas CKX5 was down-regulated by GLC. AtCKX4 and AtCKX5 are specifically expressed in root cap and root procambium, respectively (Werner et al. 2006). CK perception gene, AHK4, was up-regulated while another CK receptor AHK2 was down-regulated by GLC. AHK4 is primarily involved in embryogenesis, vascular morphogenesis in root, lateral root organogenesis, while AHK2 is involved in vascular morphogenesis in shoot, inhibition of lateral root formation; whereas both are involved into shoot meristem development and abiotic stress responses (Hwang et al. 2012). Among the CK-signalling components, type B ARRs ARR1, ARR2, ARR11 were down-regulated, whereas ARR10 was up-regulated by GLC. CK early responsive genes ARR6 and ARR8 (type A ARRs) as well as CRF1, CRF2, CRF3 were up-regulated by GLC. ARR6 is involved in controlling abiotic stress response (Jeon et al. 2010), while ARR8 is expressed strongly throughout the root and its disruption affects lateral root numbers (To et al. 2004). CRF gene family members function redundantly to regulate the development of embryos, cotyledons and leaves (Rashotte et al. 2006).

Table 1. The effect of GLC on genes involved in CK metabolism, perception and signalingThumbnail image of

CK- and GLC-regulated genes are involved in stress and developmental pathways

The Arabidopsis Information Resource (TAIR, http://www.arabidopsis.org) gene ontology (GO) (http://www.geneontology.org) search program was used to categorize genes identified by microarray (Berardini et al. 2004). Subcategories of molecular functions, cellular components and biological processes were not mutually exclusive, so GO assigned some genes to more than one category. Categorizations based on GO revealed that, there was a significant difference between CK and GLC agonistically and antagonistically regulated genes in terms of molecular functions (Supporting Information Table S2), cellular components (Supporting Information Table S3) and biological processes (Supporting Information Table S4). Genes agonistically regulated by CK and GLC largely encode hydrolases and were involved in developmental and stress processes, whereas antagonistically affected genes largely encode transferases and were predominantly involved in stress responses (Supporting Information Tables S2 & S4).

The BAP-regulated genes include members belonging to different gene families. The percentage of each BAP-regulated gene family was calculated by comparing it to the total numbers of genes belonging to that particular gene family. The information about the gene family members at the whole-genome level was obtained from PLAZA 2.5 (http://bioinformatics.psb.ugent.be/plaza) (Bel et al. 2012). The important BAP-regulated gene families are xyloglucan xyloglucosyl transferase (58.3%), DEAD box RNA helicase (13.3%), auxin-responsive family (12.7%), peroxidase (11.7%), cytochrome P450 (9%), transducin/WD-40 repeat-containing family (7%), disease resistance family (6.1%), pentatricopeptide (PPR) repeat-containing gene family (4.4%) and zinc finger family (1.8%; Supporting Information Table S5). These BAP-regulated gene families were mostly agonistically regulated by GLC treatment alone (Supporting Information Table S5). The presence of GLC also significantly (mostly agonistically) affects the extent of BAP-regulation of these gene family members (Supporting Information Table S5).

A significant number of BAP-regulated genes encode hormone response genes including CK (23), ethylene (6), auxin (22), abscisic acid/ABA (16), BR (6), gibberellin/GA (5), jasmonic acid/JA (3) and salicylic acid/SA (10) [Arabidopsis Hormone Database 2.0 (AHD2.0), http://ahd.cbi.pku.edu.cn, Jiang et al. (2011)] (Supporting Information Fig. S11).

By using the TAIR GO search program, a large number of BAP-regulated genes were found to encode genes belonging to abiotic and biotic stresses category (89) and most of them (57) were also regulated by GLC alone (Supporting Information Fig. S11). There are various reports suggesting the role of CK in several stress responses (Hwang et al. 2012). CK contents and transport was reported to get reduced by drought and/or salinity in various plant species. The decrease in CK content and transport may be due to the repression of IPT1,3,5 genes and/or activation of CKX1,3,6 genes and reduced transport of root-borne CKs in the xylem (Argueso, Ferreira & Kieber 2009). CK receptor genes AHK2 and AHK4 were up-regulated by dehydration, salinity or cold stresses in one study but down-regulated in another study although AHK3 was up-regulated in both studies (Argueso et al. 2009). Expression of type A and type B ARR genes were also regulated by various stresses, but the patterns varied significantly among different ARR members (Argueso et al. 2009; Jeon et al. 2010).

Cis-regulatory element analysis of promoters of genes regulated by CK and GLC

ATHENA (http://www.bioinformatics2.wsu.edu/Athena) promoter sequence analysis tool was used to identify statistically over-represented (P-value < 10−3) transcription factor (TF) binding sites occurring in a selected set of promoters (1000 bp upstream of the translational start site) (O'Connor, Dyreson & Wyrick 2005). The BAP up-regulated genes have TATA-box and TELO-box promoter motif, whereas BAP down-regulated genes contain TF binding enriched sites ATHB5ATCORE, DREB1A/CBF3, RY-repeat, CARGCW8GAT, MYB1AT, T-box, W-box and TATA-box motif (Supporting Information Table S6). Genes agonistically regulated by BAP and GLC alone have TATA-box and TELO-box promoter motif, whereas genes antagonistically regulated by BAP and GLC alone contain AtMYC2 binding site (BS) in RD22, I-box, RY-repeat, CARGCW8GAT, MYCATERD1 and TATA-box motif (Supporting Information Table S6).

BAP up-regulated genes whose extent significantly affected in presence of GLC have TELO-box promoter motif, whereas BAP down-regulated genes whose extent significantly affected in presence of GLC contain ATHB5ATCORE, RY-repeat, W-box, CARGCW8GAT and TATA-box motif (Supporting Information Table S6). BAP-regulated genes whose extent was agonistically affected in presence of GLC have TELO-box promoter motif, whereas BAP-regulated genes whose extent antagonistically affected in presence of GLC contain AtMYC2 BS in RD22, MYCATERD1 and TATA-box motif (Supporting Information Table S6).

The TELO promoter motif is the best classifier of GLC-induced genes and is significantly enriched in the promoters of nucleotide and protein synthesis genes as well as in genes encoding components of the translational machinery (Li et al. 2006). ATHB5ATCORE binding site motifs are found in ABA up-regulated genes (Huang et al. 2007). MYB1AT is found in the promoters of the dehydration-responsive gene rd22 and many other genes in Arabidopsis (Yamamoto et al. 2007). MYB1AT, MYCATERD1 and AtMYC2 BS in RD22 are found in stress inducible genes in Arabidopsis (Yamamoto et al. 2007). W-box elements are found in genes involved in the regulation of various physiological pathways such as disease defence pathways (Yang et al. 2010). The DREB1A/CBF3 cis-element motif is related to abiotic stresses as drought, salinity, cold, dehydration and also found in ABA up-regulated genes (Huang et al. 2007). The RY-repeat promoter motif is involved in seed-specific gene expression and interacts synergistically with ABA-responsive elements (Nakabayashi et al. 2005). CARGCW8GAT motif is particularly over-represented within the promoters of nectary-enriched genes (Kram, Xu & Carter 2009). Altogether, a number of GLC- and CK-regulated cis-regulatory elements were either ABA- or biotic/abiotic stress-responsive elements further suggesting that GLC and CK may together control stress responses in plants.

TF analysis of genes regulated by CK and GLC

According to TAIR, there are 27 416 protein coding genes in the Arabidopsis genome (TAIR10 genome release). The AGRIS (Arabidopsis Gene Regulatory Information Server, http://Arabidopsis.med.ohio-state.edu.) AtTFDB (Arabidopsis thaliana transcription factor database, http://arabidopsis.med.ohio-state.edu/AtTFDB/) identified 1770 TFs (Davuluri et al. 2003; Palaniswamy et al. 2006). According to various TF databases, Arabidopsis genome encodes ∼5–7% TFs (Riechmann & Ratcliffe 2000). However, this study may be conservative since small changes (i.e. less than twofold) in expression of TFs may significantly affect the biological processes.

Among the BAP-regulated genes, 6.5% genes were found to encode for TF. BAP down-regulated genes encode slightly more numbers of TFs (7.2%) as compared to the BAP up-regulated genes (6.1%). The genes antagonistically regulated by BAP and GLC alone encode slightly more numbers of TFs (7.5%) as compared to agonistically regulated genes, which only encodes 4.7% TFs. BAP-regulated genes whose extent was agonistically affected in presence of GLC encodes only 5.1%, whereas BAP-regulated genes whose extent was antagonistically affected in presence of GLC encodes 7.7% TFs (Supporting Information Table S7). The BAP-regulated TFs whose extent antagonistically affected in presence of GLC (7.7%), were mainly non-transcriptionally (63%) regulated by GLC alone. There are reports which suggest that several TFs are more vulnerable for regulation via post-transcriptional pathways such as phosphorylation, proteasomal pathways, etc. (Yanagisawa, Yoo & Sheen 2003). The TFs encoded by BAP-regulated genes mainly belong to the bHLH (9), C2H2 (8), AP2-EREBP (6), C3H (5) and WRKY (5) TF families (Supporting Information Table S8).

Altogether, GO-, gene family data-, Arabidopsis hormone database results-, cis-element- and TF-analysis suggest that BAP-regulated genes mainly belong to two major categories involved in (1) stress responses and (2) developmental pathways. In this study, we are primarily focusing on developmental pathways as a tool to dissect out GLC–CK interaction in the model plant system A. thaliana.

Physiological significance of GLC and CK interaction

The role of GLC and CK was studied in regulating some of the important physiological parameters such as hypocotyl length, root length, chlorophyll level and anthocyanin content. These physiological responses were quantified in response to GLC and BAP treatment alone and in combination. GLC at low concentration increased and at higher concentrations decreased the hypocotyl length, whereas BAP decreased the hypocotyl length at all the concentrations tested (Fig. 5a). Therefore, GLC and CK act antagonistically at low GLC concentration and agonistically at higher GLC concentrations (Fig. 5a). The sensitivity of hypocotyl towards BAP was reduced at higher GLC concentrations (Supporting Information Fig. S12a). GLC at low concentrations increased and at higher concentration decreased the root length, whereas BAP decreased the root length at all the concentrations tested (Fig. 5b). The high concentrations of GLC reduce the inhibitory effect of CK on root growth (Fig. 5b). The sensitivity of root towards BAP was reduced at higher GLC concentration (Supporting Information Fig. S12b). Increasing CK concentrations act antagonistically to GLC for chlorophyll production (Fig. 5c). The effect of CK on reducing the chlorophyll content is stronger in seedlings grown on high GLC concentrations (Fig. 5c). The sensitivity of seedling towards BAP was enhanced at higher GLC concentration for chlorophyll production (Supporting Information Fig. S12c). CK acts agonistically to GLC to induce anthocyanin accumulation (Fig. 5d, Supporting Information Fig. S12d). CK enhances anthocyanin production in response to high concentrations of GLC (Fig. 5d, Supporting Information Fig. S12d).

Figure 5.

Physiological parameters regulated by GLC and BAP treatment alone and in combination. (a) For controlling hypocotyl length, GLC and CK work antagonistically at lower GLC concentration but act agonistically at higher GLC concentrations. Data shown is the average of at least 45 seedlings and error bars represent standard deviation. (b) For controlling root length, GLC and CK work antagonistically at lower GLC concentrations but act agonistically at higher GLC concentration. Data shown is the average of at least 11 seedlings and error bars represent standard deviation. (c and d) For chlorophyll and anthocyanin accumulation, GLC and CK work antagonistically and agonistically, respectively.

The involvement of the GLC-signalling (HXK1-dependent and -independent pathway) components in controlling hypocotyl growth

Since a large numbers of genes were regulated by BAP and GLC together at the whole genome level, CK sensitivity of GLC-signalling mutants was examined to find out if CK–GLC interaction occurs via HXK-dependent or HXK-independent pathway. HXK-dependent pathway mutant gin2-1 was found less sensitive to CK for hypocotyl length control as compared to WT (Fig. 6a, Supporting Information Fig. S13, S14). HXK-independent pathway mutant rgs1-1, rgs1-2, gpa1-1, gpa1-2, gpa1-3, thf1-1 were found to have similar sensitivity towards CK for hypocotyl length control as compared to WT (Supporting Information Fig. S15–S19). These results suggest that CK interacts with GLC via HXK1 dependent pathway for hypocotyl length control.

Figure 6.

Hexokinase (HXK) dependence of CK signaling. (a) HXK-dependent pathway mutant gin2-1 was found less sensitive to CK for hypocotyl length control as compared to WT. Data shown is the average of at least 45 seedlings and error bars represent standard deviation. (b and c) The relative expression of genes involved in CK signaling in WT and GLC signaling mutants.

Real-time expression analysis of the genes involved in CK signalling was figured out in HXK-dependent pathway mutant gin2-1 and HXK-independent pathway mutants rgs1-1, gpa1-3. The basal level of almost all the tested CK-signalling genes was low in gin2.1 mutant as compared to WT (Fig. 6b). In rgs1-1 and gpa1-3 mutants, the basal level of almost all the tested CK-signalling genes is either almost equivalent or high in comparison to Col (Fig. 6c).

The involvement of the CK-signalling components in controlling hypocotyl growth

GLC sensitivity of CK-signalling mutants was checked to find out the involvement of CK-signalling components. The CK-receptor mutant ahk4 and type B ARR triple mutant arr1,10,11 were found to have similar sensitivity towards GLC for hypocotyl length control. Type A ARR sextuple mutant arr3,4,5,6,8,9 was found to be less sensitive towards GLC for hypocotyl length control at lower GLC concentration (i.e. 1%) while more sensitive at higher GLC concentrations (i.e. 3 and 5%) as compared to WT (Fig. 7a, Supporting Information Figs S20 & S21).

Figure 7.

CK signaling dependence of GLC response/gene expression. (a) CK-receptor mutant ahk4 and type B ARR triple mutant arr1,10,11 were found to have similar sensitivity towards GLC for hypocotyl length control. Type A ARR sextuple mutant arr3,4,5,6,8,9 was found to be less sensitive towards GLC at lower GLC concentration while more sensitive at higher GLC concentrations for hypocotyl length control as compared to WT. Data shown is the average of at least 45 seedlings and error bars represent standard deviation. (b) The relative expression of genes involved in GLC signaling in WT and CK signaling mutants.

In CK receptor mutant ahk4, type B ARR triple mutant arr1,10,11 and AHPs quintuple mutant ahp1,2,3,4,5, the basal level of GLC-signalling genes is almost equivalent to Ws (Fig. 7b). In type A ARR sextuple mutant arr3,4,5,6,8,9 and ahp6 mutant, the basal level of GLC-signalling genes is high in comparison to Col (Fig. 7b).

The involvement of the auxin-signalling components in controlling hypocotyl growth

CK is known to interact with other hormones, especially ethylene (Rashotte et al. 2005) and auxin (Su, Liu & Zhang 2011; Hwang et al. 2012) in controlling plant growth and development. Auxin maxima antagonized CK signalling in hypophysis-derived basal cell lineage by directly activating the transcription of ARR7 and ARR15 during establishment of root stem cell niche (Müller & Sheen 2008). CK signalling plays an important role during root vascular morphogenesis, a high auxin level at the xylem axis activates AHP6 expression, which acts as a negative regulator of CK signalling (Hwang et al. 2012). The root meristem size is established by a fine balance between the antagonistic interaction of CK and auxin, which converges on SHY2 (Dello Ioio et al. 2008). CK mediates lateral root inhibition via perturbation of PIN-FORMED (PIN) genes transcription (Laplaze et al. 2007) and by direct regulation of PIN1 protein levels (Marhavý et al. 2011) in lateral root founder cells. Shoot apical meristem size and activity is positively controlled by CK, whereas negatively controlled by auxin (Zhao et al. 2010). The auxin and CK signalling converges on type A ARRs ARR7 and ARR15 during SAM development (Zhao et al. 2010). In contrast to the above processes, CK and auxin seems to act synergistically to promote nodule organogenesis (Hirsch et al. 1989). There are also various reports suggesting sugar and auxin interact for controlling plant growth and development. The gin2 mutant is resistant for exogenous auxin application (Moore et al. 2003). Increasing GLC concentrations can induce differential root length, lateral roots, gravitropism and root hair elongation in auxin perception mutant TRANSPORT INHIBITOR RESPONSE1 (TIR1) and signalling mutants SOLITARY ROOT1 (SLR1)/IAA14, AXR2/IAA7 and AXR3/IAA17 (Mishra et al. 2009). GLC has also been shown to act via G-protein signalling to attenuate auxin-mediated bimodality in controlling lateral root formation (Booker et al. 2010). The NDL1 (N-MYC DOWNREGULATED-LIKE1) protein interacts with Gβγ (AGB1/AGG1 and AGB1/AGG2) dimers in Arabidopsis and establish local auxin maxima by perturbing auxin transport thus regulating lateral root initiation and emergence (Mudgil et al. 2009). These results suggest that GLC affects above mentioned root growth parameters via affecting auxin signalling and/or transport. Therefore, GLC and CK sensitivity of auxin signalling and transport mutant was checked to find out if both these pathways converge at auxin signalling/transport or not.

The auxin-signalling mutants carrying null allele in TIR1 and weak allele in AUXIN RESISTANT1-3 (AXR1-3) were found to have similar sensitivity towards GLC and CK for hypocotyl length control in comparison to WT (Fig. 8a, Supporting Information Figs S22a & S23). The Aux/IAA gain of function mutant SLR1/IAA14 was also found to have similar sensitivity, whereas AXR2/IAA7 and AXR3/IAA17 mutants were found highly resistant for different concentrations of GLC and CK for hypocotyl length control as compared to WT (Fig. 8b, Supporting Information Figs S22b & S24). These AUX/IAAs are auxin-inducible genes, which encode for transcription regulators that repress auxin-signalling pathway (Gray et al. 2001). Auxin decreases the AUX/IAA protein levels by forcing them for degradation via proteasomal pathway. The AUX/IAA mutants used here are gain-of-function mutants, and the mutation in them makes these repressors stable such that they are not recognized by proteasomal machinery and become constitutively active.

Figure 8.

Involvement of auxin signaling downstream to both GLC and CK in controlling hypocotyl growth. (a and b) The auxin signaling mutants tir1, axr1-3 and slr1 were found to have similar sensitivity, whereas, axr2 and axr3 mutants were found highly resistant for different concentrations of GLC and CK for hypocotyl length control as compared to WT. Data shown is the average of at least 32 (a) and 25 (b) seedlings and error bars represent standard deviation.

The transport mutants ETHYLENE INSENSITIVE ROOT1 (eir1-1), pin3-4, pin4-3, pin7-2, MULTIPLE DRUG RESISTANCE1 (mdr1-101) and P-GLYCOPROTEIN1 (pgp1-100) were found to have similar sensitivity towards GLC and CK for hypocotyl length control as compared to WT (Supporting Information Figs S25-S27). This suggest the involvement of intact auxin-signalling pathway rather than auxin transport for normal GLC and CK sensitivity of Arabidopsis seedlings for hypocotyl length control.

Discussion

Sugar and CK are fundamental to plants and regulate a number of similar processes. GLC and CK can act agonistically (Riou-Khamlichi et al. 1999, 2000; Hartig & Beck 2006; Kushwah et al. 2011; Das et al. 2012), antagonistically (Moore et al. 2003; Franco-Zorrilla et al. 2005) and independent of each other (Aki et al. 2007). To dissect out GLC–CK interaction, whole genome transcript profiling was coupled with characterizing some physiological parameters regulated by GLC and CK both.

GLC transcriptionally affects a large number of BAP-regulated genes agonistically

At whole genome level under experimental conditions, CK could regulate 941 genes out of which a significant number of genes (76%) were also affected by GLC alone. Earlier, Mishra et al. (2009) has shown that GLC can regulate 62% of auxin-regulated genes. This suggests larger overlap between GLC and CK in comparison to GLC and auxin. GLC and CK could regulate 89% genes agonistically as compared to 68% genes agonistically regulated by GLC and auxin, suggesting GLC and CK profoundly act agonistic to each other at whole genome level.

Agonistically affected genes were mainly transcriptionally and antagonistically affected genes were mainly non-transcriptionally regulated by GLC

GLC could significantly affect the extent of regulation of 74% BAP regulated genes. Out of these genes, 70% genes were agonistically and only 30% genes were antagonistically regulated in presence of GLC. Among the agonistically affected genes, majority of BAP-regulated genes (98%) were transcriptionally affected by GLC alone. On the contrary, among the antagonistically affected genes, majority of BAP-regulated genes (58%) were non-transcriptionally affected by GLC alone. Mishra et al. (2009) reported that GLC could regulate the extent of only 63% auxin-regulated genes (Mishra et al. 2009). Among the agonistically (40%) and antagonistically (60%) affected auxin-regulated genes, 93% were transcriptionally and 53% were non-transcriptionally regulated by GLC alone, respectively (Mishra et al. 2009). This indicates that GLC may employ a common mechanism to antagonistically regulate CK- and auxin-regulated gene expression. The antagonistic regulation of genes by GLC may either require a CK-dependent factor for sensitivity of these genes towards GLC or may employ post-transcriptional regulation of affected genes. AXR3, a member of AUX/IAA family of proteins degraded in the presence of CK. However, the AXR3 protein level was stabilized by GLC, which is consistent with a previous study by Mishra et al. (2009). Type A ARRs in CK signalling are similar to AUX/IAA proteins which also involve proteasomal pathway as a common means of regulation (Ren et al. 2009). Similar to AUX/IAA proteins, GLC may also stabilize type A ARR proteins from proteasomal degradation either via inhibiting the activity of proteasome complex or by modulating the affinity of type A ARR proteins and proteasome complex itself or its intermediate component. GLC enhanced degradation of ETHYLENE-INSENSITIVE3 (EIN3) TF via proteasomal pathway (Yanagisawa et al. 2003). This suggests that GLC may affect different signal transduction pathways via differentially regulating the proteasomal degradation of some of the key signalling elements.

GLC can transcriptionally affect a number of genes involved in CK metabolism and signalling

Briefly, GLC was found to affect almost all the important genes involved in CK metabolism, perception and signalling. GLC up-regulation of AtIPT3 suggests that GLC may up-regulate CK-biosynthesis. GLC may act agonistically to CK by up-regulating AHK4 involved in regulating both root and shoot development. Some type A ARRs and CRFs were also up-regulated by GLC. Type A ARRs are also known to be an integrating node to converge CK signalling with light, circadian rhythm, pathogen and other hormone-signalling pathways (Sweere et al. 2001; Mira-Rodado et al. 2007; To & Kieber 2008). Regulation of Type A ARRs and CRFs by GLC provide an integration node where GLC and CK signalling can interact with each other. Based on all the observations, it appears that GLC might extensively affect/modulate CK signalling at various levels from biosynthesis to signalling.

GLC and CK mainly control stress responses and developmental pathways

GO analysis revealed that majority of the BAP-regulated genes were involved in stress responses and developmental processes. The BAP-regulated genes include a large number of important gene families. Among these BAP-regulated gene families, the highly enriched gene family is xyloglucan xyloglucosyl transferase, which encodes cell wall-modifying enzymes involved in cell growth differentiation and plant response to varying environmental conditions (Nishitani & Tominaga 1992; Xu et al. 1996). The other important gene families include auxin-responsive, zinc finger and PPR repeat-containing proteins. The members of these gene families are involved in controlling developmental and stress responses (Ciftci-Yilmaza & Mittler 2008; Schmitz-Linneweber & Small 2008; Ma et al. 2009). The enriched gene families, cytochrome P450 (Werck-Reichhart, Bak & Paquette 2002), disease resistance (R) proteins (Martin, Bogdanove & Sessa 2003) and peroxidases (Tognolli et al. 2002) are specifically related to stress processes, whereas WD-40 repeat proteins and RNA helicases of the DEAD-box protein family (Cordin et al. 2006) are involved in other processes. GO and Arabidopsis hormone database analysis also support that a large number of stress- and hormone-related genes are affected by both CK and GLC.

Our microarray data showed that BAP significantly changed transcript abundance of members of various TF families. CK- and GLC-regulated TF families bHLH (Toledo-Ortiz, Huq & Quail 2003), WRKY (Zhang & Wang 2005), AP2/EREBP (Feng et al. 2005), zinc finger family members are involved in plant growth and development as well as in response towards various stresses. The TF-enriched sites of promoters (1 Kb) of BAP down-regulated genes and genes antagonistically regulated by BAP and GLC alone differs from BAP up-regulated genes and genes agonistically regulated by BAP and GLC alone, respectively. Thus, a number of GLC- and CK-regulated cis-regulatory elements are either ABA responsive or biotic/abiotic stress responsive further suggesting that GLC and CK may together control stress response in plants.

Altogether, GO analysis, gene family data, Arabidopsis hormone database results, TF analysis, cis-regulatory elements analysis of BAP-regulated genes, along with reports in literature suggest that CK- and GLC-regulated genes mainly belong to either plant growth and development or to the stress responses (Argueso et al. 2009; Jeon et al. 2010; Brenner et al. 2012; Gupta & Rashotte 2012; Hwang et al. 2012; Shi & Rashotte 2012; Spíchal 2012).

Regulation of physiological responses by GLC and CK

GLC and CK act antagonistically at low GLC concentration and agonistically at higher GLC concentrations for hypocotyl length regulation. GLC receptor mutant gin2-1 was found less sensitive towards CK in terms of hypocotyl length inhibition as compared to WT. HXK-independent pathway mutants have similar sensitivity towards CK for hypocotyl length control as compared to WT. Therefore, for hypocotyl length control, CK interacts with GLC via AtHXK1-dependent pathway. The basal level of almost all the tested CK-signalling genes either the positive or the negative regulator was low in gin2-1 mutant as compared to WT. In AtHXK1-independent pathway mutants rgs1-1 and gpa1-3, the basal level of almost all the tested CK-signalling genes is either almost equivalent or high in comparison to Col. This suggests that the interaction between GLC and CK involves both AtHXK1-dependent and -independent pathway. There are several reports in literature that point towards agonistic interaction between sugar and CK for regulation of CycD3 expression (Riou-Khamlichi et al. 1999, 2000; Hartig & Beck 2006), anthocyanin accumulation (Wade, Sohal & Jenkins 2003; Guo, Hu & Duan 2005; Das et al. 2012) and early seedling development (Németh et al. 1998; Salchert et al. 1998; Kushwah et al. 2011). The cytokinin resistant1 (cnr1) mutant shows CK resistance and sucrose hypersensitivity (Laxmi et al. 2006). Moore et al. (2003) reported that gin2 mutant display delayed leaf senescence and hypersensitivity towards CKs for shoot regeneration as compared to the WT. Similarly, constitutive CK signalling (CKI1 and ARR2) plants could also overcome the GLC repression, further confirming antagonistic interaction between GLC and CK (Moore et al. 2003). In addition, there are some reports in Arabidopsis (Higuchi et al. 2004; Franco-Zorrilla et al. 2005; Wu, Dabi & Weigel 2005; Skylar et al. 2010) as well as in crop plants (Ikeda et al. 1999; Swartzberg et al. 2006, 2011), which suggest antagonistic interaction between sugar and CK.

CK-signalling mutant analysis suggested that type A ARRs are involved in GLC control of hypocotyl length. Type A ARRs are also known to be an integrating node to converge CK signalling with light, circadian rhythm, pathogen and other hormone-signalling pathways (To & Kieber 2008; Hwang et al. 2012). The interaction of phosphorylated ARR4 with phyB provides a mechanism for the integration of light and CK signalling (Sweere et al. 2001; Mira-Rodado et al. 2007). During embryogenesis for establishment of root stem cell niche and shoot meristem development, CK and auxin signalling converges on type A ARRs (ARR7 and ARR15; Müller & Sheen 2008; Zhao et al. 2010). CK receptor and signalling mutants (positive regulators) in which endogenous CK signalling is low, the basal level of GLC signalling genes was almost equivalent to WT. CK-signalling mutants (negative regulators) in which endogenous CK signalling is high, the basal level of GLC-signalling genes was also high in comparison to WT.

There are previous reports about the functional importance of auxin signalling and transport interaction with GLC and CK, in controlling several developmental processes. Resistance of the AUX/IAA gain-of-function mutants axr2 and axr3 towards GLC and CK control of hypocotyl length suggest that GLC and CK may converge at AXR2 and AXR3 proteins for controlling hypocotyl length. Comparable sensitivity of the auxin transport mutants and WT towards CK and GLC suggest that GLC and CK work via auxin-signalling elements rather than involving auxin transport for controlling hypocotyl growth.

Our studies suggest that GLC can affect almost all aspects of CK homeostasis from CK biosynthesis to degradation, perception and signalling. GLC affects CK-regulated gene expression transcriptionally and non-transcriptionally. The agonistic responses are profoundly controlled transcriptionally, whereas antagonistic responses could also involve post-transcriptional mechanisms. An elaborate analysis of tissue-specific, developmental and temporal control of GLC–CK interaction would shed further light on how these two very important signals integrate to affect plant growth and development in broader context.

Acknowledgments

We are grateful to the National Institute of Plant Genome Research Central Instrument Facility (Real-Time PCR Division and Confocal Imaging Facility) for their assistance and help. We are thankful to UCI DNA & Protein MicroArray Facility, University of California, Irvine, for conducting microarray experiment and initial data analysis. We are also thankful to the Department of Biotechnology, Government of India, for research fellowship to S.K.

Ancillary