N-Glucosyltransferase UGT76C2 is Involved in Cytokinin Homeostasis and Cytokinin Response in Arabidopsis thaliana

Abstract

Cytokinins are a class of phytohormones that play a crucial role in plant growth and development. The gene UGT76C2 encoding cytokinin N-glucosyltransferase of Arabidopsis thaliana has been previously identified. To determine the in planta role of UGT76C2 in cytokinin metabolism and response, we analyzed the phenotypes of its loss-of-function mutant (ugt76c2) and its overexpressors. The accumulation level of the cytokinin N-glucosides was significantly decreased in ugt76c2, but substantially increased in UGT76C2 overexpressors compared with the wild type. When treated with exogenously applied cytokinin, ugt76c2 showed more sensitivity and UGT76C2 overexpressors showed less sensitivity to cytokinin in primary root elongation, lateral root formation, Chl retention and anthocyanin accumulation. Under normal growth conditions ugt76c2 had smaller seeds than the wild type, with accompanying lowered levels of active and N-glucosylated cytokinin forms. The expression levels of cytokinin-related genes such as AHK2, AHK3, ARR1, IPT5 and CKX3 were changed in ugt76c2, suggesting homeostatic control of cytokinin activity. Studies of spatiotemporal expression patterns showed that UGT76C2 was expressed at a relatively higher level in the seedling and developing seed. In their entirety, our data, based mainly on this comparison and opposite phenotypes of knockout and overexpressors, strongly suggest that UGT76C2 is involved in cytokinin homeostasis and cytokinin response in planta through cytokinin N-glucosylation.

Introduction

Cytokinins are a class of plant hormones involved in many aspects of plant growth, development and environmental responses—including cell division, shoot initiation and continuity, root and lateral root growth inhibition, chloroplast development, seed development, leaf senescence, stress tolerance, nutritional signaling and plant–pathogen interactions (Mok and Mok 1994, Sakakibara 2006, Hirose et al. 2008, Argueso et al. 2009).

Naturally occurring cytokinins are adenine derivatives and can be classified by their N⁶ side chain as isoprenoid or aromatic cytokinins. Cytokinins with an unsaturated isoprenoid side chain, such as N⁶-(Δ²-isopentenyl)adenine (iP) and trans-zeatin (tZ), are the most prevalent in planta (Mok and Mok 2001). There have been considerable advances in identifying genes encoding enzymes involved in cytokinin biosynthesis, metabolism and degradation. Genes encoding a key enzyme of cytokinin biosynthesis, adenosine-phosphate isopentenyltransferase (IPT), have been isolated from bacteria and plants (Akiyoshi et al. 1984, Kakimoto 2001, Takei et al. 2001). Cytokinin oxidase/dehydrogenase (CKX) catalyzes the irreversible degradation of cytokinins and is responsible for the majority of metabolic cytokinin deactivation in many plant species (Mok and Mok 2001). CKX genes have been isolated from several plants including maize (Houba-herin et al. 1999, Morris et al. 1999), Arabidopsis thaliana (Bilyeu et al. 2001, Werner et al. 2001), rice (Oryza sativa) (Ashikari et al. 2005) and orchids (Yang et al. 2003). In Arabidopsis, the CKX gene family was found to have seven members (i.e. AtCKX1–AtCKX7) (Schmülling et al. 2003). The overexpression of AtCKX genes in transgenic tobacco and Arabidopsis plants caused cytokinin deficiency traits (Werner et al. 2001, Werner et al. 2003). Recently, LOG, which encodes the cytokinin riboside 5′-monophosphate phosphoribohydrolase, was identified as a cytokinin-activating enzyme through the analysis of lonely guy (log) mutants that are deficient in the maintenance of shoot meristems of rice (Kurakawa et al. 2007). Nine LOGs were also identified from Arabidopsis, and seven of them had enzymatic activities equivalent to that of rice LOG. Results from mutants and overexpression lines of LOG genes suggested that the direct activation pathway via LOGs plays a pivotal role in regulating cytokinin activity during normal growth and development in Arabidopsis (Kuroha et al. 2009).

Cytokinin homeostasis and cytokinin responses in planta depend mainly on its biosynthesis, degradation, transport, signaling and modification. Glycosylation, one of the modifications of cytokinins, has been found in many plant species. However, the physiological role of cytokinin glycosylation in planta is not well characterized. Glycosylation in which a sugar molecule, generally glucose, is attached to cytokinin can occur at N³, N⁷ and N⁹ of the purine ring, or at the hydroxyl group of the zeatin and dihydrozeatin side chain, forming N-glucoside or O-glucoside, respectively. O-Glucosyl conjugates of the side chain of cytokinins can be converted into active cytokinins by β-glucosidases, suggesting that O-glucosides of cytokinins represent inactive storage forms of the hormones. Genes encoding O-glycosyltransferases of cytokinins from several plant species have been identified by Mok and co-workers, including ZOG1 from Phaseolus lunatus, ZOX1 from P. vulgaris, and cis-ZOG1 and cis-ZOG2 from maize (Martin et al. 1999a, Martin et al. 1999b, Martin et al. 2001, Veach et al. 2003). When inducing the expression of ZOG1 in transgenic tobacco leaf disks, a 10-fold higher zeatin level was required for the formation of shoots and callus compared with the controls (Martin et al. 2001, Bowles et al. 2006), whereas overexpression of ZOG1 in maize led to growth retardation and tasselseed formation (Rodo et al. 2008), both suggesting that the O-glucosylation modification could affect the activity of cytokinins. Glucosyl conjugates at the N³, N⁷ and N⁹ positions of cytokinins are usually inactive in bioassays. Because the N-glucosides are resistant to glucosidases, which is in contrast to O-glucosides, it is thus assumed that the N-modifications represent the irreversibly inactive cytokinins. When plants are treated with a high concentration of cytokinin, the major conjugated form is the N⁷-glucoside (Cowley et al. 1978). It is assumed that the N- modifications irreversibly inactivate cytokinins, but the precise in vivo function of these N-glucosyl conjugates is unknown (Mok and Mok 2001).

Recently, two genes (UGT76C1 and UGT76C2) encoding cytokinin N-glucosyltransferases from Arabidopsis were isolated and their catalytic activities were studied in vitro (Hou et al. 2004). It was found that UGT76C1 and UGT76C2 glucosylate all classical cytokinins: N⁶-benzyladenine (6-BA), tZ, cis-zeatin (cZ), dihydrozeatin, iP and kinetin, but not glucosylate purine bases such as adenine and guanine. The glucosylation occurred at either the N⁷ or the N⁹ position, but not at the N³ position of cytokinins, with a higher specific enzyme activity towards the N⁷ compared with the N⁹ position for most of the tested cytokinins. Although the enzymatic activities that glucosylate cytokinins have been biochemically identified, whether these enzymes play any physiological role in the control of cytokinin homeostasis and in the cytokinin responses of plants is still not clear.

We are interested in investigating the physiological role of the cytokinin N-glucosyltransferases. In the present study, we report our results on one of the cytokinin N-GTs, UGT76C2, by exploiting its loss-of-function mutant and overexpressors. Our data demonstrated that UGT76C2 was involved in cytokinin homeostasis and the cytokinin response through catalyzing the N-glucosylation of cytokinins in planta.

Results

Identification of T-DNA insertion mutant and overexpression lines of UGT76C2

We obtained ugt76c2 T-DNA insertion mutant seeds (SALK_135793C) from NASC (http://nasc.life.nott.ac.uk/) (Fig. 1A). The ugt76c2-1 line (abbreviated as ugt76c2 hereafter) was confirmed as T-DNA insertion homozygotes by PCR (data not shown) and reverse transcription–PCR (RT–PCR) (Fig. 1B). The RT–PCR results showed that UGT76C2 was expressed in the wild type but not in the mutant ugt76c2. In addition, UGT76C2-overexpressing homozygous lines were also obtained, and the expression level of UGT76C2 was examined by RT–PCR and quantitative real-time PCR (RT-qPCR) in independent transgenic lines (Fig. 1C, D). According to the expression level, we selected UGT76C2-overexpressing lines 4 and 19 (UGT76C2OE4-2 and UGT76C2OE19-2, respectively) for the following analyses. Despite the change in gene expression level, we noted that the mutant and overexpresssing lines identified above did not show apparently different phenotypes during vegetative growth under normal growth conditions.

Fig. 1

Schematic of the ugt76c2 T-DNA insertion mutant and expression assays of UGT76C2 in mutant and overexpression lines. (A) Position of T-DNA insertion in gene UGT76C2. P1 and P2 indicate relative positions and orientation of primers used for RT–PCR. (B) Expression assay of UGT76C2 in the wild type and mutant. (C) Expression assays of UGT76C2 in the wild type and UGT76C2-overexpressing lines. (D) Quantitative real-time PCR analysis of UGT76C2 expression in wild type and overexpressor lines. Total RNA prepared from 2-week-old seedlings grown on 1/2 MS medium were subjected to RT–PCR analysis. The TUBULIN gene was used as a control.

Contents of cytokinins and their glucosides in the loss-of-function mutant and overexpressors

To determine the in vivo activity of UGT76C2, the levels of endogenous cytokinins and their respective glucosides were first determined from the loss-of-function mutant, overexpressors and wild type grown under normal conditions. The most prevalent naturally occurring cytokinins are N⁶-substituted adenine derivatives with an unsaturated isoprenoid side chain (e.g. zeatin and iP) (Mok and Mok, 2001). Therefore, the iP and zeatin-type cytokinins—including their free forms, N-glucosides, O-glucosides, ribosides and nucleotides—were chosen for determination. For the ugt76c2 mutant, the concentrations of the most active free forms of cytokinins (i.e. iP and tZ) were comparable with those of the wild type, although other forms of cytokinins including ribosides, nucleotides and free cZ were somewhat changed in concentration between the mutant and the wild type (Fig. 2, see Supplementary Fig. S1). The change in content of cytokinin glucosides was particularly noteworthy. The ugt76c2 mutant showed a substantial decrease in cytokinin N-glucosides including tZ7G, tZ9G, iP7G and iP9G (Fig. 2). For cytokinin O-glucoside (tZOG and cZOG), however, the ugt76c2 mutant showed an increased concentration compared with the wild type. For the UGT76C2 overexpressors, the data showed that tZ, iP and other nucleobase cytokinins maintained a similar level to the wild type. However, N-glucosides including tZ7G, tZ9G, iP7G and iP9G showed a dramatic increase in the UGT76C2 overexpressors (Fig. 3). These data indicated that UGT76C2 was required for the N-glucosylation of cytokinins in planta. Meanwhile, the data implied a metabolic homeostasis to maintain the normal level of free nucleobase forms of cytokinins in the transgenic lines.

Fig. 2

Contents of cytokinins and their glucosides of the ugt76c2 mutant and the wild type. Seedlings were grown on vertical MS plates for 2 weeks and then used for hormone determination: ∼100 mg of seedlings per sample was pooled and three independent biological samples were analyzed for each genotype. Error bars represent the SD (n = 3). Asterisks over bars indicate significant differences or very significant differences between the wild type and mutants. Set at *P < 0.05 and **P < 0.01 (t-test). tZ, trans-zeatin; tZR, tZ riboside; cZ, cis-zeatin; cZR, cZ riboside; DZRPs, dihydrozeatin riboside phosphates; iP, N⁶-(Δ²-isopentenyl)adenine; iPR, iP riboside; tZRPs, tZR phosphates; cZRPs, cZR phosphates; iPRPs, iPR phosphates; tZROG, tZR-O-glucoside; tZRPsOG, tZRPs-O-glucoside; cZRPsOG, cZRPs-O-glucoside; iP9G, iP-9-N-glucoside; tZ7G, tZ-7-N-glucoside; tZ9G, tZ-9-N-glucoside; tZOG, tZ-O-glucoside; cZOG, cZ-O-glucoside; cZROG, cZR-O-glucoside; iP7G, iP-7-N-glucoside; and DZ9G, dihydrozeatin-9-N-glucoside.

Fig. 3

Contents of cytokinins and their glucosides of UGT76C2 overexpressors and the wild type. Seedlings were cultured on vertical MS plates for 2 weeks and then used for hormone determination: ∼100 mg of seedlings per sample was pooled and three independent biological samples were analyzed for each genotype. Error bars represent the SD (n = 3). Asterisks over bars indicate significant differences or very significant differences between the wild type and overexpression lines. Set at *P < 0.05 and **P < 0.01 (t-test). tZ, trans-zeatin; tZR, tZ riboside; cZ, cis-zeatin; cZR, cZ riboside; DZRPs, dihydrozeatin riboside phosphates; iP, N⁶-(Δ²-isopentenyl)adenine; iPR, iP riboside; tZRPs, tZR phosphates; cZRPs, cZR phosphates; iPRPs, iPR phosphates; tZROG, tZR-O-glucoside; tZRPsOG, tZRPs-O-glucoside; cZRPsOG, cZRPs-O-glucoside; iP9G, iP-9-N-glucoside; tZ7G, tZ-7-N-glucoside; tZ9G, tZ-9-N-glucoside; tZOG, tZ-O-glucoside; cZOG, cZ-O-glucoside; cZROG, cZR-O-glucoside; iP7G, iP-7-N-glucoside; and DZ9G, dihydrozeatin-9-N-glucoside.

Effects of exogenously applied cytokinin on main root elongation and lateral root growth

Since cytokinins have strong inhibitory effects on root growth (Cary et al. 1995, Beemster and Baskin 2000, Werner et al. 2003, Ioio et al. 2007, Bishopp et al. 2009), root elongation can be used as a criterion to analyze the responses of a specific plant genotype to cytokinins. We analyzed the effect of exogenously applied 6-BA on root development of the loss-of-function mutant and overexpressors. It should be noted that UGT76C2 can utilize 6-BA as a substrate (Hou et al. 2004). The main root length of the ugt76c2 mutant was significantly shorter than that of the wild type on media containing 0.1–1.0 μM 6-BA (Fig. 4A). In contrast, the main roots of UGT76C2 overexpressors were significantly longer than those of the wild type on media containing 0.1–5.0 μM 6-BA (Fig. 4B). To determine whether root growth was specifically affected by cytokinins, IAA was added to the medium in a separate experiment. No difference in root length among the loss-of-function mutant, overexpressors and wild type was observed (data not shown).

Fig. 4

Statistical analysis of root elongation of the ugt76c2 mutant (A) and overexpression lines (B). Seedlings with equal root length were transferred onto vertical 1/2 MS plates supplemented with 0, 0.1, 0.5, 1 or 5 μM 6-BA and grown for 1 week. Then root elongation was measured. Three independent biological replicates for each genotype were performed. Error bars represent the SD (n = 30). Asterisks over bars indicate significant differences or very significant differences between the wild type and mutants or overexpression lines under the same treatment, set at *P < 0.05 or **P < 0.01 (t-test).

Several reports described the inhibitory effect of cytokinins on lateral root formation (Li et al. 2006, Laplaze et al. 2007). To detect the response of lateral root formation to cytokinins in the loss-of-function mutant and overexpression lines, we further investigated the lateral root density of the 17-day-old mutant, overexpression lines and wild type on 1/2 Murashige and Skoog (MS) medium with or without 0.01–5.0 μM 6-BA. The lateral root density of the ugt76c2 mutant was lower than that of wild type on media containing 0.05–0.1 μM 6-BA, while the lateral root density of UGT76C2 overexpressors was higher than that of the wild type on media containing 0.05–5.0 μM 6-BA (Fig. 5).

Fig. 5

Lateral root density of the ugt76c2 mutant (A) and overexpression lines (B). Seedlings with equal root length were transferred onto vertical 1/2 MS plates supplemented with 0, 0.01, 0.05, 0.1, 0.5, 1 or 5 μM 6-BA and grown for 2 weeks. Then lateral root number was counted. Three independent biological replicates for each genotype were performed. Error bars represent the SD (n = 30). Asterisks over bars indicate significant differences or very significant differences between the wild type and mutants or overexpression lines under the same treatment, set at *P < 0.05 or **P < 0.01 (t-test).

The combined results from root and lateral root experiments suggest that the inhibition of the cytokinin N-glucosyltransferase UGT76C2 greatly enhanced the responses of main root and lateral root to exogenously applied cytokinins, resulting in more sensitive phenotypes. In contrast, overexpression of UGT76C2 decreased the cytokinin responses of plants, resulting in less sensitive phenotypes. Therefore, the tentative conclusion drawn from the experiments is that UGT76C2 could affect the activity of cytokinins via N-glucosylation and thus change the plant response to cytokinins in vivo.

Changes in Chl retention of detached leaves

The influence of cytokinins on the Chl content of leaves and their ability to retard leaf senescence were described soon after their discovery (Richmond and Lang 1957, Mothes and Baudisch 1958). We investigated the participation of UGT76C2 in mediating Chl retention of detached leaves when applying exogenous cytokinin (0–5 μM 6-BA) and dark conditions. The experimental results showed that, before treatment, the loss-of-function mutant and overexpressors had the same level of Chl as the wild type (Fig. 6A, C). After 1 week of incubation in darkness in solution containing 6-BA, the Chl content of the ugt76c2 mutant was significantly higher than that of the wild type with the application of 0, 0.5 and 1.0 μM 6-BA; whereas the Chl contents of UGT76C2 overexpressors were lower than those of the wild type with the application of 0.5–5.0 μM 6-BA (Fig. 6B, D). These data suggested that UGT76C2 contributed to regulating the effects of exogenously applied cytokinins on Chl retention in leaves.

Fig. 6

Relative Chl contents of the ugt76c2 mutant (A and B) and overexpression lines (C and D). Seventh leaves of 4-week-old plants for each genotype were collected, weighed and Chl contents determined at the zero time point (without treatment) (A and C) or after 1 week of treatment with different concentrations of 6-BA in darkness (B and D). Three independent biological replicates for each genotype were performed. Error bars represent the SD (n = 30). Chl contents are presented as a proportion of that of the wild type without treatment or with a control treatment, set to a value of unity. Asterisks over bars indicate significant differences or very significant differences between the wild type and mutants or overexpression lines under the same treatment, set at *P < 0.05 or **P < 0.01 (t-test).

Changes in anthocyanin accumulation

Anthocyanins are pigmented flavonoids responsible for most red, pink, purple and blue colors found in plants. Deikman and Hammer (1995) reported that an accumulation of anthocyanin was induced by cytokinin and thus it was a useful marker for cytokinin responsiveness. We observed that the ugt76c2 mutant exhibited a deeper purple color and that UGT76C2 overexpressors had a lighter purple color than the wild type when cultured on media containing the cytokinin 6-BA for 2 weeks (data not shown). We measured the anthocyanin contents and found that there was a much higher anthocyanin content in the ugt762 mutant than in the wild type for most media supplemented with different concentrations of 6-BA. In contrast, much lower anthocyanin contents were found in UGT76C2 overexpressors (Fig. 7). These results suggested that the activity level of N-glucosyltransferase in the mutant or overexpressors regulated the effects of exogenously applied cytokinins on anthocyanin accumulation in leaves.

Fig. 7

Relative anthocyanin contents of the ugt76c2 mutant (A) and overexpression lines (B). Seedlings were grown for 2 weeks on vertical plates supplemented with different concentrations of 6-BA and then anthocyanin contents were measured. Three independent biological replicates for each genotype were performed. Error bars represent the SD (n = 15). Anthocyanin contents were presented as a proportion of that of the wild type, with a control treatment set to a value of unity. Asterisks over bars indicate significant differences or very significant differences between the wild type and mutants or overexpression lines, set at *P < 0.05 or **P < 0.01 (t-test).

Seed size and endogenous cytokinins in developing siliques of the ugt76c2 mutant

Compared with the wild type, the ugt76c2 mutant and UGT76C2 overexpressors did not show obvious phenotypic changes throughout the growth period under normal growth conditions. However, the seeds of the ugt76c2 mutant were smaller and seed weights significantly reduced compared with the wild type (Fig. 8A, B). Seed weights of UGT76C2 overexpressors were not significantly different from those of the wild type (data not shown). To determine the possible reason for the seed size change in the ugt76c2 mutant, cytokinins and their glucosides in developing siliques of the ugt76c2 mutant were measured—because collecting and measuring only the developing seeds is technically difficult. Our data showed that the developing siliques of the ugt76c2 mutant had significantly reduced iP7G and iP9G in comparison with the wild type (Fig. 9). In addition, other cytokinins (e.g. tZ, tZR, cZ, cZR, iP and iPR) and cytokinin glucosides (e.g. tZ7G, tZ9G, tZOG, tZROG and cZROG) were also slightly reduced in developing ugt76c2 siliques. These results suggest that a defect in UGT76C2 expression could affect seed development through changing the metabolite homeostasis of cytokinins and their glucosides.

Fig. 8

Seed size and thousand seed weight of the ugt76c2 mutant. (A) Observation of wild-type and ugt76c2 seeds under a dissecting microscope with the same magnification. Bar = 1 mm. (B) Thousand seed weights of the wild type and ugt76c2. Three independent biological replicates for each genotype were performed using seeds of at least two independent batches. Error bars represent the SD (n ≥ 500). An asterisk over bars indicates significant difference between the wild type and mutant, set at *P < 0.05 (t-test).

Fig. 9

Contents of cytokinins and their glucosides in developing young siliques of the ugt76c2 mutant. Young siliques (∼100 mg) were collected from 35-day-old plants and then used for hormone determination. Three independent biological samples were analyzed for each genotype. Error bars represent the SD (n = 3). Asterisks over bars indicate significant differences between the wild type and mutants, set at *P < 0.05 (t-test). tZ, trans-zeatin; tZR, tZ riboside; cZ, cis-zeatin; cZR, cZ riboside; iP, N⁶-(Δ²-isopentenyl)adenine; iPR, iP riboside; tZRPs, tZR phosphates; cZRPs, cZR phosphates; iPRPs, iPR phosphates; tZROG, tZR-O-glucoside; tZRPsOG, tZRPs-O-glucoside; cZRPsOG, cZRPs-O-glucoside; iP9G, iP-9-N–glucoside; tZ7G, tZ-7-N-glucoside; tZ9G, tZ-9-N-glucoside; tZOG, tZ-O-glucoside; cZROG, cZR-O-glucoside; iP7G, iP-7-N-glucoside; and DZ9G, dihydrozeatin-9-N-glucoside.

Expression levels of cytokinin-related genes in the ugt76c2 mutant

As mentioned above, there were no obvious phenotypic variations from the loss-of-function mutant or overexpressors, except for the changed seed size of the ugt76c2 mutant grown under normal conditions. The most active free forms of cytokinins (i.e. iP and tZ) also stayed at a similar level in the mutant or overexpressors compared with the wild type. To determine whether, in a situation of significantly changed cytokinin N-glucosylation, the expression levels of other cytokinin-related genes were changed so as to maintain the homeostasis of free-form cytokinins and normal growth, the ugt76c2 mutant was used to perform RT-qPCR. The tested cytokinin genes were involved in synthesis (IPT genes), activation (LOG genes), degradation (CKX genes), O-glucosylation (UGT85A1) (Hou et al. 2004), perception (AHK genes) and signaling (ARR genes). We also checked the expression of a cyclin (CYC3B) and a putative disease resistance response protein (DRRP), which were previously shown to be target genes of cytokinin signaling (Pischke et al. 2006, Taniguchi et al. 2007, Argyros et al. 2008). The RT-qPCR showed that the expression levels of the receptor genes AHK2 and AHK3, cytokinin type-B response regulator gene ARR1 and cytokinin synthesis gene IPT5 were all down-regulated in the mutant (Fig. 10). However, the cytokinin oxidase gene CKX3 was up-regulated. Other cytokinin-related genes remained unchanged. Interestingly, a putative disease resistance response gene DRRP was also down-regulated in the ugt76c2 mutant. These results suggested that the up-regulation or down-regulation of cytokinin-related genes in the mutant enabled plants to maintain homeostasis of free-form cytokinins and normal growth under normal conditions.

Fig. 10

Relative expression levels of cytokinin metabolism- and signaling-related genes in the wild type and the mutant. Total RNA prepared from 2-week-old seedlings grown on 1/2 MS medium was subjected to RT-qPCR. The relative expression of each sample was calculated by dividing the expression level of the analyzed gene by that of TUBULIN. Gene-to-TUBULIN ratios were then averaged and presented as a proportion of the wild type, set to a value of unity. Values are the means of three biological replicates ± SD (n = 3).

Expression pattern of UGT76C2

To investigate the expression patterns of UGT76C2, we generated a β-glucuronidase (GUS) reporter construct fused to the UGT76C2 promoter and obtained transgenic plants. The expression regions of UGT76C2 were examined by histochemical staining for GUS activity in the transgenic lines. We found that there was strong expression of UGT76C2 in roots, hypocotyls, cotyledons, young leaves, young lateral roots and immature seeds, but weak expression in inflorescences, especially inflorescence stems, and other tissues (Fig. 11). In addition, the expression levels of UGT76C2 in different tissues were also analyzed by RT-qPCR. A good consistency in the GUS activity assay in terms of expression levels was obtained (Fig. 12). These results indicated that UGT76C2 has a spatio-temporal expression pattern in Arabidopsis growth and development.

Fig. 11

GUS expression analysis of the UGT76C2 promoter. Between five and 10 independent transformants were analyzed and the representative staining patterns were photographed. (A) One-day-old seedling; (B) 4-day-old seedling; (C) 8-day-old seedling; (D) young inflorescence; (E) immature seeds; (F) roots of 35-day-old plants. Bars = 0.5 mm for (A), and 1 mm for (B–F).

Fig. 12

Expression levels of UGT76C2 in the wild type. Total RNA prepared from different tissues was subjected to RT-qPCR. The relative expression of UGT76C2 was calculated by dividing its expression level by that of TUBULIN. UGT-to-TUBULIN ratios were then averaged and presented. Values are the means of three biological replicates ± SD (n = 3).

Discussion

Glycosyltransferases (GTs) are a multigene superfamily in plants that can transfer single or multiple activated sugars to a wide range of metabolites, resulting in glycosylated compounds. Glycosylation is thought to be one of the most important modification reactions and plays a key role in maintaining cell homeostasis, thus probably participating in the regulation of plant growth and development (Jones and Vogt 2001, Lim and Bowles 2004, Wang and Hou 2009).

In vitro biochemical studies have shown that UGT76C2 can glucosylate active cytokinins including tZ, dihydrozeatin, iP, 6-BA, kinetin and cZ to form their N-glucosides (Hou et al. 2004). However, in vitro experimental results may not necessarily reflect the true activities of those enzymes in planta due to an absence of factors influencing GT activity and its kinetic properties in the cell, such as cofactors, protein–protein interactions, metabolite channeling and the effects of crowding. In the present study, the quantification of endogenous cytokinins and their glucose conjugates provided strong evidence for UGT76C2 activities in planta. Our data showed that the ugt76c2 mutant had much lower concentrations of cytokinin N-glucosides than the wild type, while UGT76C2 overexpressors had much higher concentrations of cytokinin N-glucosides. These results confirmed that UGT76C2 was responsible for cytokinin N-glucosylation and the formation of cytokinin N-glucosides in planta, which was in accordance with its biochemical activity (Hou et al. 2004). Furthermore, it is noteworthy that the knockout mutant and overexpressors show opposite phenotypes of the cytokinin sensitivity of main root elongation, lateral root formation, Chl retention and anthocyanin accumulation. These findings indicate that UGT76C2 plays an important role in influencing the cytokinin response of plants possibly via glucosylating and deactivating cytokinins.

In previous study, cytokinins were used in millimolar concentrations to serve as substrates for the UGT76C2 enzyme in vitro (Hou et al. 2004). This is far above the endogenous and effective cytokinin concentrations examined here and in other studies, which are in the low nanomolar range (Catterou et al. 2002, Werner et al. 2003). However, it should be noted that almost nothing is known about the local distribution of cytokinin metabolites and their local concentrations, which might be higher than those measured in total tissue extracts. It might also be that cofactors which were missing in the in vitro assay are required for optimal enzyme activity

Under normal growth conditions, the mutant and overexpressors of UGT76C2 did not show obvious phenotypic changes. Additionally, the biologically most active cytokinin forms (tZ and iP) did not clearly change in the mutant or overexpressors when there was a dramatic change in cytokinin N-glucoside levels. The expression levels of cytokinin-related genes may provide clues for us to consider the regulation mechanisms occurring in planta. From our RT-qPCR results, it can be seen that, in the ugt76c2 mutant, the cytokinin synthesis-related gene IPT5 was down-regulated, while the cytokinin degradation-related gene CKX3 was up-regulated. It is likely that the changes in expression of these two genes reduced the disturbance in cytokinin levels caused by the lack of N-glucosylation, which would have caused an increased level of the biologically most active free forms (tZ and iP). Furthermore, the down-regulation of genes involved in cytokinin perception and signaling (AHK2, AHK3 and ARR1) may further maintain a homeostasis in cytokinin responses and normal growth and development. A putative disease resistance response protein (DRRP) was also down-regulated in the ugt76c2 mutant. DRRP is a cytokinin primary response gene previously determined to be a direct target of the type B response regulator ARR1 and be induced by the cytokinin 6-BA (Taniguchi et al. 2007, Argyros et al. 2008). In this study, the down-regulation of DRRP expression was consistent with the ARR1 expression change. On the other hand, the DRRP expression change in the ugt76c2 mutant may reflect that UGT76C2 also takes part in disease resistance responses.

Recently, the gene encoding O-glycosyltransferases of cytokinins, ZOG1, has been identified by Mok and co-workers from Phaseolus lunatus and heterologously transformed into tobacco and maize (Martin et al. 2001, Rodo et al. 2008). ZOG1 overexpression lines of tobacco and maize had greatly increased levels of zeatin-O-glucoside and caused a higher to lower sensitivity shift in response to exogenous zeatin. These findings are similar to our data on the gene encoding the N-GT of cytokinins, UGT76C2, whose overexpression lines had greatly increased levels of cytokinin-N-glucoside and decreased responses to exogenous 6-BA. These results indicate that ZOG1 and UGT76C2 are both important players in regulating cytokinin homeostasis and cytokinin responses. On the other hand, ZOG1 overexpression causes marked phenotypes in transgenic tobacco plants and transgenic maize. In the present study, however, UGT76C2 overexpression did not produce obvious phenotypic changes in Arabidopsis. We speculated that possibly the UGT76C2 overexpression level was not high enough to cause a phenotype. Otherwise, it is possible that plants have evolved a stronger regulatory capacity or plasticity in cell homeostasis to cope with the expression change of its own gene than a ‘foreign’ gene.

Seed size is an agronomically important trait; however, little is known about the factors regulating it. Several reports have indicated a relationship between Arabidopsis seed size and cytokinin signaling or cytokinin catabolism. For example, increased seed size was observed in transgenic plants overexpressing CKX genes (Werner et al. 2003), the ahk2,3,4 triple mutant (Riefler et al. 2006), the ahp1,2,3,4,5 quintuple mutant (Hutchison et al. 2006) and the cki1 mutant (Deng et al. 2010). In the present study, ugt76c2 mutant seeds were smaller than wild-type seeds, a phenotype opposite to that observed in the other studies mentioned above. This result suggests that UGT76C2 may play a role as a cytokinin-modifying enzyme in seed development. In the case of plants overexpressing CKX genes, increased seed size was accompanied by a decreased concentration of cytokinins and their N-glucosides in seedlings (Werner et al. 2003). However, in the ahk2,3,4 triple mutant, the increased seed size was accompanied by increased concentrations of cytokinins and their N-glucosides (Riefler et al. 2006). Data in the present study showed that, in ugt76c2 mutant seedlings and developing siliques, the concentrations of the most active cytokinin forms (i.e. tZ and iP) were similar to those of the wild type, while concentrations of their N-glucosides were much lower. Thus, the experimental results indicate a potential contribution of cytokinin N-glucosylation to seed size. However, as the cytokinin measurement was from developing siliques, whether the cytokinin contents found here reflect the situation of developing seeds is a concern that needs to be addressed.

Materials and Methods

Plant materials and growth condition

Arabidopsis thaliana ecotype Columbia-0 (Col-0) was used as the wild-type plant. The T-DNA insertion mutant seeds of ugt76c2 (SALK_135793C) were ordered from NASC (http://nasc.life.nott.ac.uk/) and confirmed by PCR and RT–PCR (the primers are listed in Supplementary Table S1). The mutant was derived from the Col-0 ecotype. The overexpression lines of UGT76C2 (At5g05860) were obtained as follows: full-length cDNA of UGT76C2 was amplified by PCR using high fidelity Taq enzyme (DR010, TAKARA) and cloned into the pBluescript II SK (+) vector. After the correct sequences were confirmed, the full-length cDNA was cloned into the pBI121 vector to replace the GUS and generate the Cauliflower mosaic virus 35S promoter fusion construct 35S_pro:UGT76C2. In all experiments DH5α was used for subcloning, and the Agrobacterium strain GV3101 was used for plant transformation via the floral dip method (Clough and Bent 1998). Homozygous lines were then established from primary transformants. At least two independent homozygous lines of 35S_pro:UGT76C2 were selected for further experiments. Arabidopsis plants were grown in the greenhouse on Nutrition Soil (Shangdao Biotech Co. Ltd.) with vermiculite (Nutrition Soil : vermiculite = 2 : 1) at 22°C under a 16/8 h light/dark cycle with a light intensity of ∼100 μmol m⁻² s⁻¹. For seedling assays in vitro, seeds were surface-sterilized and cold treated at 4°C for 3 d in darkness and then exposed to white light (∼75 μmol m⁻² s⁻¹). Seedlings were grown at 22°C on horizontal or vertical plates containing MS or 1/2 MS medium supplemented with 3% sucrose and 0.7% agar unless otherwise specified.

Root elongation assay

Surface-sterilized wild-type, mutant and overexpressor seeds were planted in a line on vertical 1/2 MS plates containing 3% sucrose and 0.7% agar, incubated in darkness for 3 d at 4°C, then transferred into a growth chamber for vertical culture at 22°C in a 16/8 h light/dark cycle. When roots had grown to a length of 1–1.5 cm after ∼3 d, 10 healthy seedlings of each of the wild type, mutant and overexpressors with approximately equal root lengths were transferred to fresh vertical 1/2 MS plates containing 3% sucrose and 0.7% agar with different concentration of phytohormones. The positions of the root tips were indicated using a fine-tipped marker pen. After 7 d, images were captured and the advancement of the root tips was measured by NIH-Image, and the root elongation data quantitatively analyzed in Microsoft Excel. There were at least three replicates per experiment.

Lateral root density assay

Surface-sterilized seeds were planted in a line on 1/2 MS plates containing 3% sucrose and 0.7% agar and incubated in darkness for 3 d at 4°C, then transferred into a growth chamber for vertical culture at 22°C in a 16/8 h light/dark cycle. When roots had grown to a length of 1–1.5 cm after ∼3 d, 10 healthy seedlings of the wild type, mutant and overexpressors with approximately equal root lengths were transferred to fresh vertical 1/2 MS plates containing 3% sucrose and 0.7% agar with different concentrations of 6-BA. After 14 d, images were captured, the lengths of roots were measured by NIH-Image and the number of lateral roots for each main root counted with the aid of a magnifier. The lateral root density was calculated by the formula: lateral root density = lateral root number/main root length. All data were quantitatively analyzed in Microsoft Excel. There were at least three replicates per experiment.

Chl content measurement

The Chl measurement was performed essentially as described by Chory et al. (1991). Briefly, the seventh leaves of 4-week-old plants for each genotype were collected, weighed and Chl contents determined at the zero time point or after 1 week of treatment with different concentrations of 6-BA in darkness. The leaves were frozen in liquid nitrogen, ground to a fine powder, total Chl was extracted into 80% acetone, and the Chl a and b contents were calculated using MacKinney's specific absorption coefficients in which Chl a = 12.7 (A₆₆₃) − 2.69 (A₆₄₅) and Chl b = 22.9 (A₆₄₅) − 4.68 (A₆₆₃). The total specific Chl content was expressed as mg of Chl g⁻¹ rosette leaves. There were at least three replicates per experiment.

Anthocyanin content measurement

Surface-sterilized seeds were planted in a line on 1/2 MS plates containing 3% sucrose and 0.7% agar and incubated in darkness for 3 d at 4°C, then transferred into a growth chamber for vertical culture at 22°C in a 16/8 h light/dark cycle. After 2 d, five healthy seedlings each of the wild type, mutant and overexpressors with approximately equal root lengths were transferred to new vertical plates with different concentrations of 6-BA. After 14 d, seedlings were harvested for anthocyanin content measurement according to the procedure described by Neff and Chory (1998). Briefly, the samples were incubated overnight in 450 μl of methanol acidified with 1% HCl in darkness at 22°C. After extraction, 300 μl of water and 300 μl of chloroform were added to the extract and mixed. After centrifugation at 12,000 r.p.m. for 5 min, the absorbance (A) of the supernatant was measured at 530 and 657 nm. By subtracting the A₆₅₇ from the A₅₃₀, the relative amount of anthocyanin per seedling was calculated. There were at least three replicates per experiment.

Seed size measurement

The wild type, mutants and overexpressors were grown in soil under strictly similar conditions until the plants reached senescence. Siliques on the middle position of plants were harvested from the wild type, mutants and overexpressors, respectively. For each genotype, at least 500 seeds were counted and weighed. Three independent biological replicates for each genotype were performed using seeds of at least two independent batches.

Quantification of cytokinins and their glucosides

For the quantification of cytokinins in seedlings, seeds were surface-sterilized and planted on MS plates containing 3% sucrose and 0.7% agar, incubated in darkness at 4°C for 3 d, then transferred to a growth chamber at 22°C under a 16/8 h light/dark cycle for vertical culture. After 2 weeks, seedlings were harvested and weighed; ∼100 mg of seedlings per sample was pooled and three independent biological samples analyzed for each genotype. For the quantification of cytokinins in developing siliques, plants of different genotypes were grown on soil under strictly similar conditions. After 35 d of growth, ∼100 mg of young siliques of similar size were collected from each sample and used for hormone determination. Three independent biological samples were analyzed for each genotype. The procedure used for cytokinin analysis was previously described by Kojima et al. (2009).

RT–PCR and RT-qPCR assay

Seeds were surface-sterilized and planted on 1/2 MS plates containing 3% sucrose and 0.7% agar, incubated in darkness at 4°C for 3 d, and transferred to a growth chamber at 22°C under a 16/8 h light/dark cycle for vertical culture. After 2 weeks, seedlings were harvested. Total RNA was extracted from seedlings with the TRIzol method and treated with RNase-free DNase I (TAKARA). RNA (0.5 μg) was used as a template for first-strand cDNA synthesis using the PrimeScript RT reagent kit (DRR037, TAKARA) and the PCR comprised 24–37 cycles of 30 s at 94°C, 30 s at 55°C and 50 s at 72°C. RT-qPCR was conducted by using the SYBR Premix Ex Taq II kit (DRR081, TAKARA) according to the manufacturer's instructions and was run on an SLA_3296 real-time PCR system (Bio-Rad).The procedures for real-time PCR were: 30 s at 94°C, 40 cycles of 5 s at 94°C, 1 min at 60°C, plate read, 2 s at 82°C, plate read, 5 min at 72°C, and a melting curve from 65 to 95°C with every 0.5°C held for 1 s. All primers used in RT–PCR and RT-qPCR are listed in Supplementary Table S1.

GUS fusion construction and plant transformation

A 2 kb DNA fragment upstream of the start codon of UGT76C2 (AT5G05860), which was used as the UGT76C2 promoter, was amplified by PCR using high fidelity Taq enzyme (DR010, TAKARA) and cloned into the pBluescriptII SK (+) and pEASY-Blunt vector, respectively. This DNA fragment might overlap with a part (∼1.1 kb) of the upstream open reading frame before the 0.9 kb intergenic region of AT5G05860 and AT5G05850. After the correct sequences were confirmed, the 2 kb DNA fragment was cloned into the pBI121 vector to replace the Cauliflower mosaic virus 35S promoter and generate the UGT76C2 promoter–GUS fusion construct UGT76C2_pro:GUS. Homozygous lines were selected from transformants. At least two independent transgenic lines of UGT76C2_pro:GUS were selected and used for GUS staining.

GUS staining assay

Seedlings within 1–14 d after germination grown on MS plates, and roots, leaves, flowers, siliques and immature seeds from soil-grown plants were collected for GUS staining assay. There were 5–10 independent transformants for each transgenic line used for GUS staining. Samples were placed in 90% acetone on ice for 15 min, then washed twice with staining buffer [50 mM sodium phosphate, pH 7.2, 0.2 mM K₃Fe(CN)₆, 0.2 mM K₄Fe(CN)₆ and 0.2% Triton X-100] without X-Gluc (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) before adding staining buffer with X-Gluc (to a final concentration of 2 mM), and incubating overnight at 37°C. The samples were washed in 70% ethanol for 30 min, before observation under a dissecting microscope (Olympus). Only representative staining patterns were photographed.

Funding

This work was supported by the National Natural Science Foundation of China [grant of the General Program (No. 30770214) to B.K.H.]; the National Natural Science Foundation of China [grant of the Major Research Plan (No. 90917006) to B.K.H.].

Acknowledgments

We thank the Nottingham Arabidopsis Stock Centre for providing T-DNA insertion mutant seeds.

Abbreviations

Abbreviations
 
• 6-BA

  N⁶-benzylaminopurine

 
• CKX

  cytokinin oxidase/dehydrogenase

 
• cZ

  cis-zeatin

 
• cZR

  cZ riboside

 
• cZOG

  cZ-O-glucoside

 
• cZROG

  cZR-O-glucoside

 
• GT

  glucosyltransferase

 
• GUS

  β-glucuronidase

 
• iP

  N⁶-(Δ2-isopentenyl)adenine

 
• iP7G

  iP-7-N-glucoside

 
• iP9G

  iP-9-N-glucoside

 
• iPR

  iP riboside

 
• IPT

  adenosine-phosphate isopentenyltransferase

 
• MS

  Murashige and Skoog

 
• RT–PCR

  reverse transcription–PCR

 
• RT-qPCR

  quantitative real-time PCR

 
• tZ

  trans-zeatin

 
• tZR

  tZ riboside

 
• tZOG

  tZ-O-glucoside

 
• tZROG

  tZR-O-glucoside

 
• UGT

  UDP glucose glucosyltransferase.