The purine derivative PI-55 blocks cytokinin action via receptor inhibition Luka ´ s ˇ Spı ´chal 1 , Toma ´ s ˇ Werner 2 , Igor Popa 1 , Michael Riefler 2 , Thomas Schmu ¨ lling 2 and Miroslav Strnad 1 1 Laboratory of Growth Regulators, Institute of Experimental Botany, AS CR & Palacky ´ University, Olomouc, Czech Republic 2 Institute of Biology ⁄ Applied Genetics, Free University of Berlin, Germany Cytokinins, which comprise an important class of plant hormones, are chemically classified as N 6 -substi- tuted derivatives of adenine with either isoprenoid (e.g. zeatin) or aromatic [e.g. 6-benzylaminopurine (BA)] sidechains [1] (Fig. 1). Cytokinins regulate diverse processes during growth and development. The histi- dine kinases cytokinin response 1 (CRE1) ⁄ Arabidopsis histidine kinase (AHK)4, AHK3 and AHK2 have been identified as cytokinin receptors in Arabidopsis [2,3], which feed the signal into a two-component signaling pathway. Cytokinin receptors have also been identified from other species, including maize [4], the legumes Medicago truncatula [5] and Lotus japonicus [6,7], and rice [8]. Genetic analyses of Arabidopsis cytokinin receptor mutants has assigned functions to the recep- tors in the regulation of shoot growth, leaf senescence, seed size and germination, and root elongation and branching. In addition, analysis of loss-of-function and gain-of-function mutations in legume cytokinin recep- tors has revealed their direct involvement in nodule organogenesis [5–7]. Chemical inhibitors of cytokinin action would be very potent tools to study the mechanism of cytokinin perception and signal transduction. Furthermore, they Keywords anticytokinin; cytokinin; cytokinin receptor; cytokinin receptor antagonist; hormone signaling Correspondence L. Spı ´ chal, S ˇ lechtitelu ˚ 11, CZ-78371 Olomouc, Czech Republic Fax: +420 58 5634870 Tel: +420 58 5634855 E-mail: lukas.spichal@upol.cz Website: http://rustreg.upol.cz (Received 26 August 2008, revised 24 October 2008, accepted 3 November 2008) doi:10.1111/j.1742-4658.2008.06777.x One of several potential approaches to study mechanisms of action of bio- logically active compounds is to develop their agonists and antagonists. In the present study, we report the identification of the first known molecule antagonizing the activity of the plant hormone cytokinin at the receptor level. This compound, 6-(2-hydroxy-3-methylbenzylamino)purine, desig- nated PI-55 in the present study, is structurally closely related to cytokinin 6-benzylaminopurine, but substitutions at specific positions of the aromatic side chain strongly diminished its cytokinin activity and conferred antago- nistic properties. PI-55 competitively inhibited the binding of the natural ligand trans-zeatin to the Arabidopsis cytokinin receptors cytokinin response 1 (CRE1) ⁄ Arabidopsis histidine kinase (AHK) 4 and AHK3 and repressed induction of the cytokinin response gene ARR5:GUS. Genetic analysis revealed that CRE1 ⁄ AHK4 is the primary target of PI-55. Cyto- kinin bioassays also demonstrated the anticytokinin effect of PI-55 in several other species. Furthermore, we show that PI-55 accelerated the germination of Arabidopsis seeds and promoted the root growth and formation of lateral roots, thus phenocopying the known consequences of a lowered cytokinin status and demonstrating its potential to inhibit cytokinin perception in planta. PI-55 is the first example for the targeted development of a cyto- kinin antagonist and represents an initial step for the preparation of cyto- kinin antagonists with broad activity and reduced agonistic properties. Abbreviations AHK, Arabidopsis histidine kinase; BA, 6-benzylaminopurine; CDK, cyclin dependent kinase; CRE1, cytokinin response 1; DAG, days after germination; PI-55, 6-(2-hydroxy-3-methylbenzylamino)purine. 244 FEBS Journal 276 (2009) 244–253 ª 2008 The Authors Journal compilation ª 2008 FEBS would be expected to influence plant growth and devel- opment and might thus possess a potential for agricul- tural application. Subsequent to the discovery of cytokinins, several attempts have been made to identify compounds with anticytokinin properties. Diverse adenylate and non-adenylate compounds with agonis- tic and antagonistic effects have been synthesized. Some of them were classified as anticytokinins because of their strong inhibitory effect on cytokinin-induced responses in various bioassays. N 6 -sidechain cytokinin analogues with alterations in the purine ring were pre- pared in the 1970s, based on the rationale that cyto- kinin antagonists are likely to have structural similarities to cytokinins because such features should allow them to compete for the same receptor(s) but render them ineffective as cytokinins [9]. For example, heterocyclic compounds sharing cytokinin structural motifs, such as 7-substituted 3-methyl-pyrazolo[4,3-d]pyrimidines [10,11], 4-substituted pyrrolo[2,3-d]pyrimidines and their 7-glucosides [12,13], and 7-deaza analogues of 2-methylthioadenine cytokinins [14], were prepared and their activities were tested in a callus growth test. Some of the compounds had strong growth-retarding effects and, because the addition of cytokinins reversed the growth arrest, these results were interpreted as evidence indicating that these compounds were specific anticytokinins sharing the same site of action as active cytokinins [9]. Following observations that some phenylureas struc- turally unrelated to cytokinins can exhibit cytokinin- like activity, cytokinin antagonists were sought among urea derivatives. Several compounds were identified based on the reduction of cytokinin-induced effects, such as chlorophyll retention, radish leaf expansion and callus growth [15]. Other anticytokinins with non- adenine structures were sought based on considerations of the biosterism between N 6 -substituted adenine, phenylurea cytokinins and the herbicidal inhibitors of photosystem II, s-triazines and N-arylcarbamates [16]. It was concluded that phenylureas, s-triazines and N-arylcarbamates share the same site of action: a cyto- kinin receptor localized in the chloroplast [9]. How- ever, the effects of cytokinin antagonists have been defined primarily in tobacco callus assays, and hence their antagonistic properties in other plant systems are uncertain [9]. Moreover, direct proof that the molecu- lar target of these compounds is a cytokinin receptor has been hindered by a lack of knowledge of cytokinin receptors and signaling. However, the recent identifica- tion of Arabidopsis cytokinin receptors [2,3] has allowed direct examination of the true mode of action of known anticytokinins. We have recently shown that representatives of the anticytokinins described above are not competitive inhibitors of cytokinin receptors but inhibit cell cycle progression (i.e. the activity used for their classification as anticytokinins) through a dif- ferent mechanism. Indeed, some compounds were shown to inhibit cyclin dependent kinases (CDKs) and also cause a number of cellular changes typical of known CDK inhibitors [17]. The identification of cytokinin receptors enabled us to employ a novel approach. We followed the assump- tion that a compound possessing true antagonistic activity should compete with the natural ligands for the receptor(s) and block cytokinin action in vivo, but should not have intrinsic cytokinin activity in cyto- kinin bioassays, and neither activate the cytokinin signaling pathway, nor be cytotoxic or inhibit CDKs. Based on these criteria, we screened a library of synthetic compounds derived from cytokinins in a receptor-based assay. We have identified a BA deriva- tive, 6-(2-hydroxy-3-methylbenzylamino)purine, desig- nated PI-55 in the present study (Fig. 1), and describe its antagonistic effects in various cytokinin-mediated processes in the model plant Arabidopsis thaliana and other species. Results PI-55 reduces binding of the natural ligand to cytokinin receptors In the search for the cytokinin antagonists that genu- inely act at the receptor level, we focused on structural derivatives of cytokinins having substitutions in specific positions that diminish cytokinin-like activity. We employed transformed Escherichia coli strains expressing the cytokinin receptors CRE1 ⁄ AHK4 and AHK3 and the cytokinin-activated reporter gene cps::lacZ [4,18,19], and used three classical cytokinin Fig. 1. Structures of active cytokinins and the newly identified anticytokinin PI-55. tZ, trans-zeatin. L. Spı ´ chal et al. PI-55 blocks cytokinin action FEBS Journal 276 (2009) 244–253 ª 2008 The Authors Journal compilation ª 2008 FEBS 245 bioassays (tobacco callus, wheat senescence and Amaranthus bioassays) to select compounds from among > 400 synthetic cytokinin derivatives that are not sensed by the receptors and, in addition, have no activity in the cytokinin bioassays. These compounds were further tested in binding assays with isolated E. coli membranes containing CRE1 ⁄ AHK4 or AHK3 [17] for their ability to compete for receptor binding with the radiolabeled natural ligand trans-zeatin, employing unlabeled trans-zeatin and adenine as posi- tive and negative controls, respectively. An example of the strategy for the selection of the candidate cytokinin antagonist is summarized in Table 1. It is well known that a hydroxyl group placed in the ortho- position, in contrast to the meta- position, of the aromatic side chain of the highly active cytokinin BA strongly decreases its receptor sensing and biological activity in cytokinin bioassays (Table 1) [17,20]. Interestingly, the situation is reversed when a methyl group is substi- tuted in these positions (Table 1) [21]. Simultaneous substitution by both hydroxyl and methyl groups (in the ortho- and meta-positions, respectively) minimized the ability of the compound to activate the cytokinin signaling pathway and reduced its cytokinin-like activ- ity in bioassays (Table 1). The disubstituted BA, desig- nated PI-55 in the present study (Fig. 1), did not activate the CRE1 ⁄ AHK4-mediated signaling pathway, even at a concentration 500-fold greater (50 lm) than that required for receptor activation by trans-zeatin (Fig. 2A). Similarly, up to 10 lm, PI-55 no activation of AHK3 was observed, although the highest concen- tration induced a weak response slightly greater than that induced by the negative control substance adenine (Fig. 2B). AHK3 has been shown previously to have broad ligand specificity [19], which might explain this response. Importantly, in our in vitro system, PI-55 reduced the binding of [2- 3 H]zeatin to both receptors in a dose-dependent manner, with IC 50 values that were only four-fold higher than those of the cytokinin BA, whereas adenine was unable to compete even at 1000-fold excess (Fig. 2C; see also Fig. S1A). Thus, a slight modification of the cytokinin structural motif led to a compound that does not activate the cytokinin signaling pathway, but competes for receptor binding with only a fourfold lower efficacy than widely used cytokinin BA and approximately 200-fold lower than the strongest natural cytokinin trans-zeatin. Analysis of double-reciprocal plots showed that PI-55 was com- petitive towards trans-zeatin with K i values of 1.7 ± 0.9 lm and 10.6 ± 0.7 lm in the CRE1 ⁄ AHK4 (Fig. 2D) and AHK3 assays (see Fig. S1B), respec- tively. PI-55 inhibits the activation of a cytokinin primary response gene To support our finding that PI-55 truly competes with cytokinin for binding to cytokinin receptors, we next determined whether PI-55 blocks cytokinin signal transduction in planta. We recorded the response of the ARR5:GUS reporter gene, which is known to be rapidly upregulated by cytokinin [22,23]. PI-55 decreased the blue staining caused by ARR5:GUS expression following treatment by 2.5 lm BA in the root and shoot (Fig. 3A). The quantitative evaluation of the effect of PI-55 on BA-induced ARR5:GUS expression demonstrated its competitive inhibitory effect (Fig. 3B). The competitive mechanism of this inhibition was judged from the previous findings that PI-55 directly competes with BA for binding to the cytokinin receptors (Fig. 2C; see also Fig. S1A). PI-55 alone only weakly induced ARR5:GUS, probably due to weak activation of AHK3 (Fig. 2B) and eventually AHK2. To investigate in more detail to what extent Table 1. Development of a cytokinin receptor antagonist. The ability of different substituted BAs to activate the cytokinin signaling pathway through the receptors AHK3 and AHK4 in a bacterial assay and to block the binding of radiolabeled natural ligand to the receptors, as well as the biological activity in classical cytokinin bioassays, are compared. Adenine is shown as the negative control. The details of the assays are described in the Experimental procedures. For the semi-quantitative evaluation: +++, high; ++, moderate; +, low; –, none. Compound Substituent Activation of signal- ing pathway Competition for receptor binding Activity in cytokinin bioassays R1 R2 AHK3 AHK4 BA H H +++ +++ +++ +++ N H N N N NH R2 R1 2-OH-BA OH H + ) ++ 3-OH-BA H OH +++ +++ +++ +++ 2-Me-BA CH 3 H ++ + + +++ 3-Me-BA H CH 3 + ) +++ 2-OH-3-Me-BA (PI-55) OH CH 3 ))+++ ) Ade ))) ) PI-55 blocks cytokinin action L. Spı ´ chal et al. 246 FEBS Journal 276 (2009) 244–253 ª 2008 The Authors Journal compilation ª 2008 FEBS different AHK proteins are sensitive to inhibition by PI-55, we introduced the ARR5:GUS gene construct in the background of all three AHK double loss-of-func- tion mutants [24] and examined them in competitive assays. In this experiment, the seedlings were grown in the presence of 1 lm BA and 1 or 5 lm PI-55. As shown in Fig. 3C, the strongest reduction of ARR5: GUS activity was observed with 5 lm PI-55 in the ahk2 ahk3 mutant, in which the b-glucuronidase activity was significantly reduced by 40% (calculated from the values of b-glucuronidase specific activ- ity shown in Fig. 3C) compared to statistically not Fig. 2. Effects of PI-55 on the activity of cytokinin receptors and cytokinin binding. (A, B) Comparison of the sensitivity of CRE1 ⁄ AHK4 (A) and AHK3 (B) to the cytoki- nins trans-zeatin (tZ), BA, PI-55 and adenine (Ade) in a bacterial reporter assay. The compounds were tested at the indicated concentrations; 0.5% dimethylsulfoxide (DMSO) was used as control solvent. Error bars show the SD (n = 3). (C) Representa- tive example of competitive binding assay with CRE1 ⁄ AHK4 containing E. coli mem- branes. Binding of 2 n M [2- 3 H]tZ was assayed in the presence of various concen- trations of PI-55, unlabelled BA and tZ (posi- tive control) and adenine (Ade, negative control). (D) Representative example of Lineweaver–Burk plot of 2 n M [2- 3 H]tZ bind- ing to CRE1 ⁄ AHK4-containing E. coli membranes in the presence of different concentrations of PI-55. The inset shows a secondary plot of slopes versus inhibitor concentration. A C B Fig. 3. Effect of PI-55 on expression of a cytokinin response gene. (A) Staining for b-glucuronidase activity in shoots (upper row) and roots (lower row) of 9-day-old ARR5:GUS-expressing transgenic Arabidop- sis seedlings treated with 2.5 l M BA in the presence or absence of 5 l M PI-55. (B) Quantitative fluorometric measurement of PI-55 dose-dependent suppression of cyto- kinin-induced expression of ARR5:GUS in 3-day-old seedlings. (C) Quantitative evalua- tion of b-glucuronidase activity 3 DAG in ahk double receptor mutants harboring ARR5:- GUS after incubation with BA and PI-55. Dimethylsulfoxide (0.1%) was used as con- trol solvent. Error bars show the SD (n = 2). Asterisks indicate a statistically significant decrease of induction of cytokinin-induced ARR5:GUS expression (Student’s t-test, P < 0.05). L. Spı ´ chal et al. PI-55 blocks cytokinin action FEBS Journal 276 (2009) 244–253 ª 2008 The Authors Journal compilation ª 2008 FEBS 247 significant reduction in the ahk2 ahk4 and ahk3 ahk4 seedlings. The results confirmed that PI-55 most strongly competes for cytokinin binding at the CRE1 ⁄ AHK4 receptor, and that its impact on the sig- nal transduction mediated through AHK2 and AHK3 is probably weaker. 1 lm PI-55 alone induced 26%, 38% and 66% of the BA-induced reporter gene activity in the ahk2 ahk3, ahk2 ahk4 and ahk3 ahk4 mutant plants, respectively, which is consistent with the result of the bacterial assay that PI-55 may act also as a weak agonist, in particular with AHK3, and ⁄ or may point to another recognition systems for cyto- kinins leading to expression of cytokinin responsive genes. PI-55 affects root growth and stimulates root branching Cytokinins are known to be negative regulators of root growth [24–27]. To study whether PI-55 is able to repress this cytokinin effect, we grew wild-type Arabid- opsis seedlings on media containing BA and various concentrations of PI-55. PI-55 alone did not affect the elongation of the primary root of wild-type Arabidop- sis seedlings; positive and, importantly, negative effects were not observed even at 1 lm. By contrast, 1 lm BA almost completely inhibited primary root growth (see Fig. S2). However, when 1 lm PI-55 was applied to the growth media of ahk2 ahk3 double mutants, a significant increase (approximately 15%) in the length of the primary roots was observed compared to the untreated control (Fig. 4A). No significant changes in elongation of the primary root were found with other two receptor double mutants (data not shown). This result is consistent with a predominant activity of PI-55 on CRE1 ⁄ AHK4 and indicates that this receptor has a regulatory function in the control of primary root elongation. A clear antagonistic effect of PI-55 on root branch- ing was observed 14 days after germination (DAG). Even at 1 nm, BA caused a strong inhibition of lateral root formation, and its effect was more pronounced at higher concentrations. An equivalent concentration of PI-55 (1 nm) almost completely reversed the suppres- sion of lateral root formation by BA (Fig. 4B). Inter- estingly, PI-55 had a positive effect on root branching when applied alone (Fig. 4B), indicating that it also suppressed the activity of endogenous cytokinins. Fur- ther investigations with receptor double mutant plants showed that, although 10 nm PI-55 had only a slightly positive effect on formation of lateral roots in ahk3 ahk4 seedlings, and no effect in ahk2 ahk4 seedlings, it caused ahk2 ahk3 seedlings to form a significantly higher number of lateral roots than untreated controls (Fig. 4C). This result again points to activity of PI-55 on CRE1 ⁄ AHK4 and indicates its regulatory function in the control of lateral root formation and root response to exogenous cytokinin. Fig. 4. Effect of PI-55 on primary root length and lateral root formation in Arabidopsis seedlings. (A) Increase in the length of the primary roots observed in ahk2 ahk3 double mutant seedlings (6 DAG) grown on medium containing 1 l M PI-55. (B) The number of lateral roots formed by wild-type Arabidopsis seedlings (11 DAG) in the presence or absence of BA and ⁄ or PI-55. (C) The increase in the number of lateral roots observed in wild-type and ahk double mutant seedlings (11 DAG) grown on medium containing 10 n M PI-55. Error bars show the SD (n = 20). Asterisks indicate statistically significant differences from the untreated controls (Student’s t-test, P < 0.05). PI-55 blocks cytokinin action L. Spı ´ chal et al. 248 FEBS Journal 276 (2009) 244–253 ª 2008 The Authors Journal compilation ª 2008 FEBS PI-55 accelerates seed germination Recently, it was demonstrated that cytokinin-deficient Arabidopsis seeds or seeds with reduced cytokinin per- ception germinate faster than wild-type seeds [24]. In the present study, we tested the effect of PI-55 on ger- mination. After sowing and pretreatment in the dark at 4 °C, wild-type Arabidopsis seeds were transferred to the light and ambient temperature. After 30 h, more than 60% of seeds sown on medium containing 1 nm PI-55 were germinating, which was twice as many than on control medium without PI-55. The proportion of germinating seeds was further enhanced at higher PI- 55 concentrations, and 80% of seeds had germinated when 10 nm PI-55 was present in the medium (Fig. 5). PI-55 antagonizes cytokinin effects in bioassays with different plant species Several classical cytokinin bioassays based on biologi- cal responses to exogenous cytokinins were used to evaluate the anti-cytokinin potential of PI-55 in other plant species. The callus bioassay is based on the ability of cytokinins to induce cell division in presence of auxin [28]. Cytokinin-dependent tobacco (Nicotiana tabacum) callus was grown on medium supplemented with various concentrations of PI-55 in the presence or absence of 0.5 lm BA and the callus biomass was scored after 30 days. 0.5 lm BA was found to be optimal for maximal callus growth. PI-55 did not significantly influence the growth of the calli when applied alone. However, the growth stimulated by BA was effectively inhibited by increasing concentrations of PI-55, and concentrations as low as 10 lm caused total inhibition of callus growth (Fig. 6A). It is possible that such inhibitory effects on cell proliferation could be due to the inhibition of proteins involved in cell division, as previously shown for some pyrazolo[4,3- d]pyrimidines and pyrrolo[2,3-d]pyrimidines, which behave as anticytokinins in this bioassay [17], rather than by competing with cytokinins for a common bind- ing site on a receptor protein. However, PI-55 is not cytotoxic and did not inhibit proliferation of human cancer cell lines (G361, HOS and MCF7) and ⁄ or cyto- kinin-dependent tobacco callus, even at concentrations of 50 and 100 lm, respectively (see Table S1). Another standard cytokinin bioassay is based on cytokinin-stimulated delay of senescence in excised leaves of wheat (Triticum aestivum), in which retention of chlorophyll is measured after 4 days of incubation in the dark [20]. As shown in Fig. 6B, 1 lm PI-55 reduced the retention of chlorophyll induced by 1 lm BA, and its effect was more pronounced at 10 lm. PI-55 had a limited effect when applied alone, but an agonistic effect was observed at higher concentrations (data not shown). A similar result was obtained in a bioassay that is based on cytokinin-induced betacyanin formation in hypocotyls of Amaranthus caudatus in the dark [29]. Also in this assay, PI-55 showed an antago- nistic effect at micromolar concentrations (Fig. 6C), whereas higher concentrations of the compound had weak cytokinin-like activity (Fig. 6C). These results revealed that PI-55 has antagonistic activity to exo- genously applied cytokinins in different bioassays in various species. Discussion We identified a structural derivative of the aromatic cytokinin BA, PI-55, as the first known anticytokinin that acts at the receptor level. Specific substitutions in ortho- and meta- positions of the aromatic sidechain of BA resulted in altered perception by the cytokinin receptors CRE1 ⁄ AHK4 and AHK3 in a bacterial assay and isolated membranes of these bacteria. Importantly, these substitutions did not affect the abil- ity of the compound to compete for binding with the natural ligand trans-zeatin (Fig. 2). Hence, only minor changes in the aromatic cytokinin structure converted the agonist BA into an antagonist. We propose that PI-55 binds in a similar fashion to the receptor as does BA, but does not induce the conformational change Fig. 5. Effect of PI-55 on germination of Arabidopsis seeds. Wild- type Arabidopsis seeds were incubated on MS medium containing varying concentrations of PI-55. Error bars show the SD (n = 3). DMSO, dimethylsulfoxide. L. Spı ´ chal et al. PI-55 blocks cytokinin action FEBS Journal 276 (2009) 244–253 ª 2008 The Authors Journal compilation ª 2008 FEBS 249 necessary to activate the receptor and initiate signaling, as suggested previously by a hypothetical concept of cytokinin action [9]. The antagonistic role of PI-55 during the early steps of cytokinin signal transduction was confirmed in planta in experiments showing that PI-55 inhibits the cytokinin-induced activation of a cytokinin-respon- sive marker gene. This contrasts with the inability of previously described putative cytokinin antagonists, which do not reduce the cytokinin-dependent activa- tion of this reporter gene [17]. Consistently, it was shown that these compounds do not interact with cytokinin receptors of Arabidopsis but interfere at a later stage with cytokinin action, namely by block- ing the cell cycle through inhibition of CDK proteins [17]. The results obtained in the present study further support the hypothesis that the receptors have different structural requirements for the stereochemical or phys- icochemical features of the ligands. Whereas CRE1 ⁄ AHK4 was not activated by PI-55 even at a concentration 500-fold higher than that required for receptor activation by trans-zeatin, AHK3 showed par- tial activation at high concentrations of PI-55. This is in accordance with results of a study of ligand specific- ities of these two receptors [19] and of a screen for cytokinin activities of various substituted BAs [21]. It was shown that only a few tested BA derivatives were able to activate CRE1 ⁄ AHK4 significantly, but most of them were recognized by the AHK3 receptor, albeit less efficiently than trans-zeatin [21]. The ligand speci- ficity of AHK2 has not been described yet. However, the results obtained based on PI-55 induction of cyto- kinin reporter expression in ahk3 ahk4 mutant propose that AHK2 also differs in ligand binding properties from other two cytokinin receptors. The activities of PI-55 in wild-type Arabidopsis plants are in accordance with the results obtained in several studies considering the consequences of geneti- cally caused alterations of cytokinin signaling or metabolism. Collectively, these studies, which involved receptor loss-of-function mutations [19], mutations of signaling proteins [30] and cytokinin-synthesizing genes [31], or caused cytokinin-deficiency by enhanced cyto- kinin degradation [25,32], showed that cytokinin acts as a negative regulator of root growth. Consistently, the presence of PI-55 in the growth medium led to enhancement of the growth of root systems of wild-type and receptor double mutant plants (Fig. 4). Furthermore, PI-55 caused a more rapid germination of Arabidopsis seeds (Fig. 5), which is also a character- istic of seeds from plants with a reduced cytokinin sta- tus [24]. The observation that PI-55 primarily blocks the CRE1 ⁄ AHK4 receptor is in good agreement with the finding that CRE1 ⁄ AHK4 is the receptor that plays the main role in the control of germination by cytokinin [24]. PI-55 interfered with cytokinin action in various bioassays in other species, indicating that it might be widely applicable and that a similar cytokinin sensing mechanism to those mediated by CRE1 ⁄ AHK4 from Arabidopsis might be active in these species. Taken together, the results obtained in the present study indicate that PI-55 acts as an inhibitor of cyto- kinin receptors and antagonizes cytokinin action in vivo. Hence, in studies on the functions of cytokinins, the compound can comprise a modulator of endo- genous cytokinin activity as an alternative to genetic approaches. In addition, it may have interesting Fig. 6. Antagonistic effect of PI-55 in standard cytokinin bioassays. BA was used as the active cytokinin and its effect was studied in the presence and absence of indicated concentrations of PI-55. (A) Effect of PI-55 on cytokinin-dependent growth of tobacco callus. (B) Effect of PI-55 on cytokinin-induced retention of chlorophyll in excised wheat leaves. (C) Effect of PI-55 on cytokinin-stimulated betacyanin formation in Amaranthus seedlings in the dark. Values represent the difference between absorption at 537 and 620 nm. Error bars show the SD (n = 5). PI-55 blocks cytokinin action L. Spı ´ chal et al. 250 FEBS Journal 276 (2009) 244–253 ª 2008 The Authors Journal compilation ª 2008 FEBS applications as a growth regulator for modifying traits of crop plants. Synthetic compounds interfering with the biosynthesis, metabolism or translocation of auxin, gibberellin and ethylene are commonly used in basic studies of the functions of these hormones. Some of them are commercially used, such as inhibitors of eth- ylene perception in controlling longevity in ornamental crops [33]. Previously, such compounds have not been available for cytokinin research. PI-55 is the first example of the targeted development of a cytokinin antagonist. Genetic and biochemical analyses revealed that the primary target of PI-55 in Arabidopsis is CRE1 ⁄ AHK4. The compound showed weak agonistic activity resulting from the partial agonistic interference with the other two cytokinin receptors. This corroborates the differences in cytokinin binding sites of individual receptors and indicates that the preparation of a ‘universal’ cytokinin receptor antagonist will not be an easy task. PI-55 represents the first step demonstrating the feasibility of preparing such compounds. However, further steps (e.g. modulation of PI-55 structure) have to be undertaken to broaden the antagonistic interference and reduce the agonistic activity. Experimental procedures Chemicals Synthesis of PI-55 is described in Doc. S1. The identity and purity of the synthesized compound was confirmed by elemental and melting point analyses, analytical TLC, HPLC and MS and NMR analysis (for data, see Support- ing information). trans-zeatin and BA were obtained from Olchemim Ltd (Olomouc, Czech Republic). Radiolabeled trans-zeatin ([2- 3 H]zeatin) was obtained from J. Hanus ˇ (Isotope Laboratory, Institute of Experimental Botany, AS CR, Prague, Czech Republic). Plant material and growth conditions Arabidopsis plants harboring the ARR5:GUS reporter gene construct and the Arabidopsis receptor double mutants have been described previously [22,24]. For analysis of ARR5:- GUS activity in the ahk mutant background, the ARR5:- GUS marker gene was introgressed into the three different ahk Arabidopsis double mutant lines. Arabidopsis thaliana ecotype Columbia (Col-0) was used as a wild-type control. For in vitro assays, seeds were surface-sterilized and sown on half-strength Murashige & Skoog medium including vitamins (Duchefa, Haarlem, The Netherlands) supple- mented with 1% (w ⁄ v) sucrose, 0.05% (w ⁄ v) MES-KOH (pH 5.7) and 1.1% agar. After cold pretreatment at 4 ° C for 3 days in the dark, the seedlings were grown under long-day conditions (16 : 8 h light ⁄ dark photoperiod) at 22 °C in a growth chamber unless otherwise stated. In the classical cytokinin bioassays, cytokinin-dependent tobacco (N. tabacum L. cv. Wisconsin 38) callus, seeds of winter wheat (T. aestivum L. cv. Hereward) and Amaranthus (A. caudatus L. var. atropurpurea) were used as described below and as described previously [20]. Bacterial cytokinin assays E. coli KMI001 strains harboring plasmids pIN-III-AHK4 and pSTV28-AHK3 [3,18] were used in the bacterial cytoki- nin assays, which were performed as described previously [19]. Fractionation of E. coli and microsome-binding assays CRE1 ⁄ AHK4- and AHK3-expressing E. coli strains [3,18] were grown until D 600 of  1 was reached at 25 °C and then fractionated into periplasmic, cytoplasmic and mem- brane fractions. Fractionation and binding assays with membranes were carried out as described previously [34]. Arabidopsis ARR5:GUS reporter gene assay This assay was carried out as described previously [23], with slight modifications. For quantitative assays, ARR5:GUS seedlings were grown for 2–3 days (22 °C, 16 : 8 h light ⁄ dark photoperiod) in six-well plates (Techno Plastic Prod- ucts, Trasadingen, Switzerland) and then cytokinin, PI-55 or control solvent (dimethylsulfoxide, final concentration 0.1%) were added to the desired final concentration. The seedlings were then incubated for 17 h at 22 °C in the dark. Histochemical b-glucuronidase analysis was performed with 9-day-old seedlings as described previously [25]. Root assay Arabidopsis wild-type and receptor double-mutant seedlings were grown on vertical plates in the presence of BA and ⁄ or PI-55 or control solvent (0.1% dimethylsulfoxide). After 6 DAG, plates were photographed and the length of the primary roots was scored using scion image software (Scion Corp., Frederick, MD, USA). The number of lateral roots was scored 11 DAG under a stereomicroscope. Twenty plants for each genotype and treatment were analyzed in the both cases. Seed germination assay This assay was performed under long-day conditions (16 : 8 h light ⁄ dark photoperiod) with no sucrose in the L. Spı ´ chal et al. PI-55 blocks cytokinin action FEBS Journal 276 (2009) 244–253 ª 2008 The Authors Journal compilation ª 2008 FEBS 251 medium as described previously [24]. Twenty seeds for each treatment were analyzed in three replicates. Cytokinin bioassays Standard bioassays based on stimulation of cytokinin- dependent tobacco callus growth, the retention of chloro- phyll in excised wheat leaves and the dark induction of betacyanin synthesis in Amaranthus cotyledons were carried out as described previously [20]. The only exception to the published protocols was that the incubation time was short- ened to 17 h in the Amaranthus bioassay. The final concen- tration of dimethylsulfoxide in the media did not exceed 0.2%. Five replicates were prepared for each cytokinin con- centration and the entire tests were repeated at least three times. Acknowledgements We thank Jarmila Balonova ´ and Miloslava S ˇ ubova ´ for skillful technical assistance; Dr Vladimı ´ r Krys ˇ tof for testing of PI-55 cytotoxicity with human cell lines; Katharina Achazi for help with introgression of the ARR5:GUS marker gene into the ahk Arabidopsis double mutant lines; and Professor David Morris for his helpful suggestions and critical reading of the man- uscript. This work was supported by the Grant Agency of the Czech Republic (522 ⁄ 07 ⁄ P197 to L.S.); the Czech Ministry of Education (MSM 6198959216 and MSM LC06034 to M.S.); and the Deutsche Fors- chungsgemeinschaft (Sfb 449 to T.S.). References 1 Mok DW & Mok MC (2001) Cytokinin metabolism and action. Annu Rev Plant Physiol Plant Mol Biol 52, 89–118. 2 Inoue T, Higuchi M, Hashimoto Y, Seki M, Kobayashi M, Kato T, Tabata S, Shinozaki K & Kakimoto T (2001) Identification of CRE1 as a cytokinin receptor from Arabidopsis. Nature 409, 1060–1063. 3 Suzuki T, Miwa K, Ishikawa K, Yamada H, Aiba H & Mizuno T (2001) The Arabidopsis sensor His-kinase, AHK4, can respond to cytokinins. Plant Cell Physiol 42, 107–113. 4 Yonekura-Sakakibara K, Kojima M, Yamaya T & Sak- akibara H (2004) Molecular characterization of cytoki- nin-responsive histidine kinases in maize. Differential ligand preferences and response to cis-zeatin. Plant Physiol 134, 1654–1661. 5 Gonzalez-Rizzo S, Crespi M & Frugier F (2006) The Medicago truncatula CRE1 cytokinin receptor regulates lateral root development and early symbiotic interaction with Sinorhizobium meliloti. Plant Cell 18, 2680–2693. 6 Murray JD, Karas BJ, Sato S, Tabata S, Amyly L & Szczyglowski K (2007) A cytokinin perception mutant colonized by Rhizobium in the absence of nodule organ- ogenesis. Science 315, 101–104. 7 Tirichine L, Sandal N, Madsen LH, Radutoiu S, Albrektsen AS, Sato S, Adamitu E, Tabata S & Stougaard J (2007) A gain-of-function mutation in a cytokinin receptor triggers spontaneous root nodule organogenesis. Science 315, 104–107. 8 Du L, Jiao F, Chu J, Jin G, Chen M & Wu P (2007) The two-component signal system in rice (Oryza sativa L.): a genome-wide study of cytokinin signal perception and transduction. Genomics 89, 697–707. 9 Iwamura H (1994) Cytokinin antagonists: synthesis and biological activity. In Cytokinins. Chemistry, Activity and Function (Mok DWS & Mok MC, eds), pp. 43–55. CRC Press, Boca Raton, FL. 10 Hecht SM, Bock RM, Schmitz RY, Skoog F & Leon- ard NJ (1971) Cytokinins: development of a potent antagonist. Proc Natl Acad Sci USA 68, 2608–2610. 11 Skoog F, Schmitz RY, Bock RM & Hecht SM (1973) Cytokinin antagonists: synthesis and physiological effects of 7-substituted 3-methylpyrazolo (4,3-d)pyrimi- dines. Phytochemistry 12, 25–37. 12 Iwamura H, Ito T, Kumazawa Z & Ogawa Y (1974) Anticytokinin activity of 4-furfurylamino-7-(b-D-ribo- furanosyl)pyrrolo[2,3-d]pyrimidine. Biochem Biophys Res Commun 57, 412–416. 13 Iwamura H, Ito T, Kumazawa Z & Ogawa Y (1975) Synthesis and anticytokinin activity of 4-substituted-7- (b-D-ribofuranosyl)pyrrolo[2,3-d]pyrimidines. Phyto- chemistry 14, 2317–2321. 14 Skoog F, Schmitz RY, Hecht SM & Frye RB (1975) Anticytokinin activity of substituted pyrrolo[2,3-d]pyr- imidines. Proc Natl Acad Sci USA 72, 3508–3512. 15 Kefford NP, Zwar JA & Bruce MI (1968) Antagonism of purine and ureas cytokinin activities by derivatives of benzylurea. In Biochemistry and Physiology of Plant Growth Substance (Whitman F & Setterfield G, eds), pp. 61–69. Runge Press, Ottawa. 16 Shimizu R, Iwamura H, Matsubara S & Fujita T (1989) Development of s-triazine anticytokinins and their quantitative structure-activity relationship. J Agric Food Chem 37, 236–240. 17 Spı ´ chal L, Krys ˇ tof V, Paprska ´ r ˇ ova ´ M, Lenobel R, Sty´ skala J, Binarova ´ P, Fenolova ´ V, De Veylder L, Inze ´ D, Kontopidis G et al. (2007) Classical anticytokinins do not interact with cytokinin receptors but inhibit cyclin-dependent kinases. J Biol Chem 282, 14356– 14363. 18 Yamada H, Suzuki T, Terada K, Takei K, Ishikawa K, Miwa K, Yamashino T & Mizuno T (2001) The Arabid- opsis AHK4 histidine kinase is a cytokinin-binding receptor that transduces cytokinin signals across the membrane. Plant Cell Physiol 42, 1017–1023. PI-55 blocks cytokinin action L. Spı ´ chal et al. 252 FEBS Journal 276 (2009) 244–253 ª 2008 The Authors Journal compilation ª 2008 FEBS 19 Spı ´ chal L, Rakova NY, Riefler M, Mizuno T, Roma- nov GA, Strnad M & Schmu ¨ lling T (2004) Two cytoki- nin receptors of Arabidopsis thaliana, CRE1 ⁄ AHK4 and AHK3, differ in their ligand specificity in a bacterial assay. Plant Cell Physiol 45, 1299–1305. 20 Holub J, Hanus ˇ J, Hanke DE & Strnad M (1998) Bio- logical activity of cytokinins derived from ortho- and meta-hydroxybenzyladenine. Plant Growth Reg 26, 109– 115. 21 Dolezˇ al K, Popa I, Krys ˇ tof V, Spı ´ chal L, Fojtı ´ kova ´ M, Holub J, Lenobel R, Schmu ¨ lling T & Strnad M (2006) Preparation and biological activity of 6-benzylaminopu- rine derivatives in plants and human cancer cells. Bioorg Med Chem 14, 875–884. 22 D’Agostino IB, Deruere J & Kieber JJ (2000) Charac- terization of the response of the Arabidopsis response regulator gene family to cytokinin. Plant Physiol 124, 1706–1717. 23 Romanov GA, Kieber JJ & Schmu ¨ lling T (2002) A rapid cytokinin response assay in Arabidopsis indicates a role for phospholipase D in cytokinin signalling. FEBS Lett 515, 39–43. 24 Riefler M, Nova ´ k O, Strnad M & Schmu ¨ lling T (2006) Arabidopsis cytokinin receptor mutants reveal functions in shoot growth, leaf senescence, seed size, germination, root development, and cytokinin metabolism. Plant Cell 18, 40–54. 25 Werner T, Motyka V, Laucou V, Smets R, Onckelen HV & Schmu ¨ lling T (2003) Cytokinin-deficient trans- genic Arabidopsis plants show multiple developmental alterations indicating opposite functions of cytokinins in the regulation of shoot and root meristem activity. Plant Cell 15, 2532–2550. 26 Laplaze L, Benkova ´ E, Casimiro I, Maes L, Vanneste S, Swarup R, Weijers D, Calvo V, Parizot B, Herrera- Rodriguez MB et al. (2007) Cytokinins act directly on lateral root founder cells to inhibit root initiation. Plant Cell 19, 3889–3900. 27 Dello Ioio R, Linhares FS, Scacchi E, Casamitjana-Mar- tinez E, Heidstra R, Costantino P & Sabatini S (2007) Cytokinins determine Arabidopsis root-meristem size by controlling cell differentiation. Curr Biol 17, 678–682. 28 Miller CO, Skoog F, Vonsaltza MH & Strong FM (1955) Kinetin, a cell division factor from deoxyribonu- cleic acid. J Am Chem Soc 77, 1392. 29 Biddington NL & Thomas TH (1973) Modified Ama- ranthus betacyanin bioassay for rapid determination of cytokinins in plant extracts. Planta 111, 183–186. 30 Hutchison CE, Li J, Argueso C, Gonzalez M, Lee E, Lewis MW, Maxwell BB, Persie TD, Schiller GE, Alfonso JM et al. (2006) The Arabidopsis histidine phosphotransfer proteins are redundant positive regula- tors of cytokinin signaling. Plant Cell 18, 3073–3087. 31 Miyawaki K, Tarkowski P, Matsumoto-Kitano M, Kato T, Sato S, Tarkowska ´ D, Tabata S, Sandberg G & Kakimoto T (2006) Roles of Arabidopsis ATP ⁄ ADP isopentenyltransferases and tRNA isopentenyltransfe- rases in cytokinin biosynthesis. Proc Natl Acad Sci USA 103, 16598–16603. 32 Werner T, Motyka V, Strnad M & Schmu ¨ lling T (2001) Regulation of plant growth by cytokinin. Proc Natl Acad Sci USA 98, 10487–10492. 33 Serek M, Woltering EJ, Sisler EC, Frello S & Sriskand- arajah S (2006) Controlling ethylene responses in flow- ers at the receptor level. Biotech Adv 24, 368–381. 34 Romanov GA, Spı ´ chal L, Lomin SN, Strnad M & Schmu ¨ lling T (2005) A live cell hormone-binding assay on transgenic bacteria expressing a eukaryotic receptor protein. Anal Biochem 347, 129–134. Supporting information The following supplementary material is available: Doc. S1. Synthesis, elemental and melting point analy- sis, analytical thin layer chromatography, MS and NMR analyses of PI-55. Fig. S1. Effect of PI-55 on cytokinin binding to AHK3. Fig. S2. Comparison of PI-55 and BA effect of on pri- mary root length of Arabidopsis wild-type seedlings. Table S1. Effect of compound PI-55 on proliferation of human cell lines and tobacco callus. This supplementary material can be found in the online version of this article. Please note: Wiley-Blackwell is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corre- sponding author for the article. L. Spı ´ chal et al. PI-55 blocks cytokinin action FEBS Journal 276 (2009) 244–253 ª 2008 The Authors Journal compilation ª 2008 FEBS 253 . The purine derivative PI-55 blocks cytokinin action via receptor inhibition Luka ´ s ˇ Spı ´chal 1 , Toma ´ s ˇ Werner 2 ,. study the mechanism of cytokinin perception and signal transduction. Furthermore, they Keywords anticytokinin; cytokinin; cytokinin receptor; cytokinin receptor