ARR1, a Transcription Factor for Genes Immediately Responsive to Cytokinins

See allHide authors and affiliations

Science  16 Nov 2001:
Vol. 294, Issue 5546, pp. 1519-1521
DOI: 10.1126/science.1065201


Cytokinins are a class of phytohormones involved in various physiological events of plants. The Arabidopsis sensor histidine kinase CRE1 was recently reported to be a cytokinin receptor. We used a steroid-inducible system to show that the transcription factor–type response regulator ARR1 directs transcriptional activation of the ARR6 gene, which responds to cytokinins without de novo protein synthesis. This fact, together with characteristics ofARR1-overexpressing plants and arr1 mutant plants, indicates that the phosphorelay to ARR1, probably from CRE1, constitutes an intracellular signal transduction occurring immediately after cytokinin perception.

Cytokinins induce a variety of physiological events, including cell division, chloroplast development, and shoot formation (1, 2). These cytokinin responses in Arabidopsis are at least partly triggered through the recognition of cytokinins by the sensor protein CRE1, which is a member of the protein histidine kinase family (3, 4). In bacteria, histidine kinases participate in His-Asp phosphorelays, which respond to environmental stimuli, usually in association with cytoplasmic response regulators, the majority of which are transcription factors (5). If the His-Asp phosphorelay is also the case in plants, an intracellular signal transduction pathway starting from the CRE1 sensor may involve response regulators. The Arabidopsis genome codes for 22 response regulators (ARRs), 12 of which contain a Myb-like DNA binding domain called ARRM (type B) (6–9). The remainder (type A) possess no apparent functional unit other than a signal receiver domain containing two aspartate and one lysine residues (DDK) at invariant positions, and their genes are transcriptionally induced by cytokinins without de novo protein synthesis (7, 9,10). The type B members, ARR1 and ARR2, bind DNA in a sequence-specific manner and work as transcriptional activators (11, 12).

We analyzed the morphological characteristics of transgenic plants carrying 35S::ARR1 and35S::ARR1ΔDDK genes (13), in which the full-length ARR1 and its truncated version missing the NH2-terminal signal receiver domain, respectively, were directed by the cauliflower mosaic virus 35S promoter. On the one hand, transgenic 35S::ARR1 plants showed no phenotypic change under normal growth conditions, except for hypertrophic cotyledons, longer cotyledonary petioles, and shorter roots than wild-type plants (Fig. 1, A and B, and Fig. 2A). On the other hand,35S::ARR1ΔDDK plants did show a phenotypic change under normal growth conditions, forming ectopic shoots on the adaxial surface of their cotyledons (Fig. 1D). In lines expressing the transgene at relatively high levels, plant growth was occasionally inhibited, with concomitant disordered cell proliferation around the shoot apex (Fig. 1C). Because the transactivating function of ARR1 was masked by its signal receiver domain in transient expression experiments (12), the phenotypes generated by35S::ARR1ΔDDK likely resulted from the constitutive transactivating function of ARR1ΔDDK, independent of a signal from an upstream component. To test whether the cytokinin signal regulates ARR1 through its signal receiver domain, we treated35S::ARR1 plants with exogenous 6–benzylaminopurine (BA) at various concentrations and compared their morphologies with those of nontreated35S::ARR1ΔDDK plants. Transgenic35S::ARR1 plants grown in the presence of 1 μM BA showed growth inhibition concomitant with disordered cell proliferation around the shoot apex (Fig. 1, F and G). This phenotype was similar to a severe phenotype seen in nontreated35S::ARR1ΔDDK plants (Fig. 1C). Other kinds of cytokinins (kinetin, trans-zeatin, and isopentenyladenine) had similar effects on 35S::ARR1 plants, whereas a different phytohormone, auxin (3-indoleacetic acid), did not (14). These results suggest that the signal receiver domain of ARR1 suppresses the function of ARR1 in the absence of cytokinins and that a cytokinin signal releases the suppression. Treatment of wild-type plants with BA caused growth inhibition but not the disordered cell proliferation that was observed in35S::ARR1 plants. (Fig. 1E). This difference might imply that ectopic overexpression of ARR1 disturbs homeostasis in cytokinin responses.

Figure 1

Phenotypes of transgenic plants. (A and E) Wild-type plants grown without and with 1 μM BA, respectively. (B and F)35S::ARR1 plants grown without and with 1 μM BA, respectively. (G) An enlargement of (F). (Cand D) 35S::ARR1ΔDDKplants without BA. (H and I)35S::ARR1ΔDDK::GR plants without and with 0.1 μM DEX, respectively. Arrowheads indicate cotyledons. Scale bars, 1 mm.

Figure 2

Root growth of wild-type (circles),35S::ARR1 (triangles), andarr1-1 (squares) plants in the presence of exogenous cytokinin. Seeds were germinated and grown on vertical agar medium containing various concentration of BA. (A) The primary root length of 8-day-old seedlings was measured, and the average and standard deviations from at least 15 individuals for each condition were calculated and plotted against the BA concentration. (B) Root growth is expressed relative to the mean root length of respective plants grown without BA. Dashed lines indicate the BA concentrations causing IC50.

Another finding from morphological analyses is that plants overexpressing ARR1 appear to be more sensitive to cytokinins than are wild-type plants (Fig. 1, E and F). To investigate the relationship between sensitivity to cytokinins and level ofARR1 expression, we examined wild-type,35S::ARR1, and arr1-1 mutant plants by two different procedures: root-elongation and callus-formation assays. The arr1-1 mutant obtained by screening transferred DNA (T-DNA)–tagged lines (15) contains a T-DNA insertion within the first intron of the ARR1 gene and produces no mature ARR1 mRNA (see below). Apparent phenotypic alterations were not detected inarr1-1 plants, except for longer roots than in wild-type plants under normal growth conditions (Fig. 2A). In the root-elongation assay, we grew seedlings on agar media containing various concentrations of BA. In the absence of BA,35S::ARR1 and arr1-1 plants had shorter and longer roots than did wild-type plants, respectively. BA inhibited root growth in all the plants at concentrations of 10 nM and higher (Fig. 2). However, roots of 35S::ARR1 andarr1-1 plants were more and less sensitive, respectively, to BA than were those of wild-type plants (see IC50, the concentration of BA needed to cause 50% inhibition of root growth, shown in Fig. 2B). In the callus-formation assay, we tested the response (green callus formation) of tissue-cultured explants to varying concentrations of the cytokinin kinetin and to the auxin 2,4-dichlorophenoxyacetic acid (2,4-D), using the procedure described by Kubo and Kakimoto (16). Like wild-type explants,35S::ARR1 and arr1-1 explants responded to increasing levels of kinetin with rapid proliferation and greening, and in the case of 35S::ARR1 explants, occasionally with shoot formation (Fig. 3). However, the sensitivity of 35S::ARR1 andarr1-1 explants to kinetin was higher and lower, respectively, than that of wild-type explants. These results show that the level of ARR1 expression correlates with sensitivity to cytokinin, which supports the view that ARR1 is involved in the cytokinin signal transduction pathway.

Figure 3

Callus growth of35S::ARR1, wild-type, andarr1-1 explants exposed to different cytokinin and auxin concentrations. Hypocotyl segments were excised and cultured on MS media containing different levels of kinetin and 2,4-D. After 21 days in culture, the induced calli were arranged and photographed. (A) 35S::ARR1 explants, (B) wild-type explants, and (C) arr1-1explants.

To gain insight into how ARR1 contributes to cytokinin signal transduction at the molecular level, we examined transcription levels of a cytokinin-responsive marker gene in 35S::ARR1and arr1-1 plants, with or without BA treatment, by Northern blot analysis (17). The marker gene we chose wasARR6 (18), a type A response regulator gene, because its expression is rapidly induced by cytokinins without de novo protein synthesis (7, 10). As in wild-type plants, ARR6 responded rapidly to BA in both35S::ARR1 and arr1-1 plants, even in the presence of cycloheximide (CHX), a protein synthesis inhibitor (Fig. 4A). However, the transcription levels induced by the BA treatment of 35S::ARR1and arr1-1 plants were higher and lower, respectively, than in wild-type plants. This induction pattern, together with the appearance of an ARR1 recognition sequence (5′ -AGATT-3′) (12) three times in the 200–base pair region upstream from the ARR6 translation start site, suggests that the transactivation of ARR6 is targeted by ARR1. Consistent with this, 35S::ARR1ΔDDK plants showing severe phenotypes expressed ARR6 at much higher levels than did wild-type plants under normal growth conditions (14). Direct transactivation of ARR6 by ARR1 was shown with transgenic plants carrying a chimeric35S::ARR1ΔDDK::GR gene, whose translation product was expected to become activated by glucocorticoid without de novo protein synthesis (19). Such steroid-inducible systems using the hormone-binding domain of the glucocorticoid receptor (GR) have been used as powerful tools to identify direct target genes for various transcription factors (20–22). Transgenic35S::ARR1ΔDDK::GR plants were indistinguishable from wild-type plants under normal growth conditions (Fig. 1H), but the addition of a synthetic glucocorticoid, dexamethasone (DEX), caused graduated morphological changes, which were dependent on the DEX concentration. A plant treated with 0.1 μM DEX showed growth inhibition concomitant with disordered cell proliferation around the shoot apex (Fig. 1I), which was similar to a severe phenotype in 35S::ARR1ΔDDKplants (Fig. 1C). These results indicate that ARR1 function can be modulated by DEX instead of by cytokinins in this system. Treatment of35S::ARR1ΔDDK::GR plants with DEX for 1 hour enhanced ARR6 transcription, even in the presence of CHX (Fig. 4B) (17). This provides in vivo evidence for direct interaction of ARR1ΔDDK::GR with the ARR6 gene. This in vivo evidence, taken together with the higher and lower ARR6 transcription levels in BA-treated 35S::ARR1 andarr1-1 plants, respectively, versus wild-type plants, lead us to conclude that ARR1 directly activatesARR6 transcription in response to cytokinins.

Figure 4

Northern blot analysis for ARR6transcription. (A) Wild-type,35S::ARR1, and arr1-1 plants were treated with 5 μM BA and/or 30 μM CHX for 30 min and then subjected to Northern blot analysis for ARR6 andARR1, with UBQ5 as a control. (B) Transgenic35S::ARR1ΔDDK::GRplants were treated with 30 μM DEX and/or 30 μM CHX for 1 hour. Numerals below the UBQ5 signals indicate the level of ARR6 mRNA that was quantified by measuring the radioactivity of the signal band with a Fujix BA100 Bio-Image analyzer, divided by the radioactivity of the UBQ5 signal band with the same RNA sample for normalization and presented as the relative value to that for wild-type plants (A) or35S::ARR1ΔDDK::GR plants (B) without chemical treatment.

In Arabidopsis, cytokinin receptors, ethylene receptors, and, possibly, osmosensors are histidine kinases (3, 4, 23–25). However, their cognate response regulators, working as transcription factors, have never been identified. Here we present evidence that ARR1 mediates a cytokinin signal, probably through its NH2-terminal signal receiver domain, and transactivates ARR6, which is immediately responsive to cytokinins. A paralogous response regulator, ARR2, shows almost identical characteristics to ARR1 (12), suggesting a functional overlap. Residual cytokinin responses observed with the arr1-1 mutant may have been provided by ARR2. In addition to ARR6, other type A member genes, including ARR4, ARR5, ARR7,ARR8, and ARR9, were also activated by DEX at various levels in35S::ARR1ΔDDK::GR plants (14), suggesting that all the immediate cytokinin-responsive genes belonging to this group are directly activated by ARR1. Also, other cytokinin-responsive genes whose promoter regions contain the ARR1 recognition sequences are possibly transactivated by ARR1. A screening for ARR1 target genes using transgenic35S::ARR1ΔDDK::GR plants will shed light on the whole view of the early cytokinin signal transduction pathway. We conclude that ARR1 is a principal transcription factor–type response regulator that is involved in an early step of cytokinin signal transduction, possibly as a partner of the sensor histidine kinase CRE1.

  • * These authors contributed equally to this work.

  • Deceased 14 July 2001.

  • To whom correspondence should be addressed. E-mail: oka{at}


View Abstract

Navigate This Article