Q&A: How do plants respond to cytokinins and what is their importance?

Cytokinins were originally defined as chemicals that induce cell proliferation and trigger callus differentiation to shoot when applied with auxins, but now it is known that cytokinins play a key role in many aspects of plant growth and development [3], including embryogenesis, maintenance of root and shoot meristems, and vascular development. They also modulate root elongation, lateral root number, nodule formation, and apical dominance in response to environmental stimuli. Thus, cytokinins are important signaling molecules for regulating growth and development throughout the life of the plant.

The recent development of mass spectrometry technology has enabled us to quantify the concentration of phytohormones and their conjugates at the organ level. Based on such analyses, for which the values are generally 0.1 to 10 pmol g⁻¹ fresh weight, it may be estimated that the in vivo concentration of cytokinins is at the nanomolar level, and that concentration may vary between different organs and growth conditions. Notably, if cytokinins are unevenly distributed in the organs at either the tissue or cellular level, their local concentration could be higher or lower. The affinity (apparent K _D) of cytokinin receptors to their ligands is around 1–40 nM [4], which would make the estimated nanomolar concentration of cytokinins physiologically relevant.

Cytokinins mainly trigger physiological responses through the regulation of gene expression. A two-component system (TCS) is employed to transduce the cytokinin signal to the target genes. The TCS, which typically consists of a sensory histidine kinase and response regulator, was first discovered in prokaryotes, where it serves to enable response to environmental stimuli. Since its first mention in 1996 [5], extensive biochemical and genetic studies have been conducted with Arabidopsis to identify and characterize the cytokinin TCS. Today, it is widely accepted that the TCS of cytokinin signaling consists of three groups of proteins in Arabidopsis: three histidine kinases (AHKs; AHK2, AHK3 and AHK4/WOL1/CRE1), five histidine-containing phosphotransfer proteins (AHPs; AHP1–AHP5) and eleven type-B response regulators (type-B ARRs; ARR1, ARR2, ARR10–ARR14 and ARR18–ARR21) (Fig. 2) [3].

AHKs are membrane-localized cytokinin receptors which consist of three domains: CHASE, histidine kinase, and receiver [2, 3, 6]. The binding of cytokinins to the AHK CHASE domain triggers autophosphorylation of a His residue in its histidine kinase domain, and the phosphoryl group is internally transferred to an Asp residue in its receiver domain. Recent studies suggest that the cytokinin receptors are localized in both the plasma membrane and the endoplasmic reticulum [7, 8], but it is still unclear whether both are functional or not.

AHPs mediate the transfer of a phosphoryl group from cytosolic AHKs to nuclear-localized type-B ARRs [3, 9]. Type-B ARRs possess a receiver domain for the phosphoryl group, a DNA-binding domain (GARP domain), and a glutamine-rich domain for transcriptional activation [10]. Transcriptional activation by ARRs is repressed in the non-phosphorylated state [10]. When phosphorylated, type-B ARRs can bind target DNA sequences and activate transcription of target genes.

Since each member of the AHK, AHP and type-B ARR gene families is functionally redundant, clear phenotypic differences are not observed in their single mutants. However, responsiveness to cytokinins is severely reduced and growth phenotypes could be observed in multiple mutants within each gene family [11–13], indicating that the cytokinin TCS plays a central role in cytokinin responses in plants.

In order to finely regulate cytokinin signaling, multiple feedback loops of the cytokinin TCS are employed to ensure the appropriate signaling flux (Fig. 2). In addition to the 11 type-B ARRs, the Arabidopsis genome has ten type-A ARR genes (ARR3–ARR9, ARR15–ARR17) [3], and some of them are direct targets of type-B ARRs. Like type-B ARRs, type-A ARRs possess a receiver domain and receive a phosphoryl group from AHPs, but they lack a DNA-binding domain. Thus, type-A ARRs can potentially inhibit the cytokinin signaling flux by competing with type-B ARRs for phosphate transfer [14].

Differential regulation of posttranslational stability of type-B ARRs is involved in the modulation of the TCS signaling flux. Cytokinins promote degradation of ARR2 via the 26S proteasome [15] while they stabilize ARR1 by preventing degradation by 26S proteasome [16]. On the other hand, KISS ME DEADLY proteins (KMDs), a family of F-box proteins, are involved in degrading ARR1, ARR2 and ARR12, but the detailed mechanisms have yet to be discovered [17].

It has also been reported that an AHP homologue, AHP6, which lacks the conserved histidine for phosphoryl group transfer, physically interacts with AHKs but does not receive a phosphate, suggesting that AHP6 inhibits cytokinin response through competition with canonical AHPs. Cytokinins repress AHP6 expression [18], which suggests that the promotion of cytokinin signaling flux is in part mediated by the down-regulation of AHP6. It was reported that cell-specific expression of AHP6 serves spatial specification of cytokinin signaling [19, 20]. In addition, cytokinins induce the expression of a cytokinin receptor gene, AHK4/WOL1/CRE1, which might lead to increased sensitivity to cytokinins [21]. It has been proposed that occurrence of these feedback loops in an organ- or a cell-specific manner is important for regulation of cytokinin signaling flux, as described later.

Cytokinin levels in vivo are determined by the balance between biosynthesis and catabolism. Biosynthesis of cytokinins is regulated by three key enzymes (Fig. 3). First, adenosine-phosphate isopentenyltransferase (IPT) catalyzes the formation of iP ribonucleotides from dimethylallyl diphosphate (DMAPP) and adenine nucleotides, with preferential use of ATP or ADP [22, 23]. Cytochrome P450 monooxygenase 735A (CYP735A) hydroxylates the end of the prenyl side chain of iP ribonucleotides, with a preference for iP-riboside 5’-monophosphate (iPRMP) or diphosphate (iPRDP), to produce tZ ribotides [24]. Finally, LONELY GUY (LOG), a phosphoribohydrolase, converts the precursors iPRMP and tZ-riboside 5’-monophosphate (tZRMP) to their active forms, iP and tZ, respectively [25]. Although cytokinin ribonucleosides, such as iP ribonucleoside and tZ ribonucleoside, are widely found in plant tissues, they are considered to be precursors and mainly converted to their active form via LOG after phosphorylation, because log septuple mutants, in which the LOG-dependent pathway is impaired, exhibit severe cytokinin-deficient phenotypes in Arabidopsis [26]. The biosynthesis of cytokinins is regulated through IPT genes by various internal and external environmental stimuli, such as nitrogen sources [27, 28], and other phytohormones [29].

For the catabolism of cytokinins, cytokinin oxidase/dehydrogenase (CKX) irreversibly degrades cytokinins by cleaving the unsaturated isoprenoid side chain, which results in the formation of adenine and the corresponding aldehyde [3]. Glycosylation of cytokinins also plays an important role in modulating cytokinin activity: N-glucosylation of the N7 and N9 positions of the adenine moiety and O-glucosylation of the zeatin side chain are catalyzed by UGT76C1 and C2, and 85A1, respectively, in Arabidopsis [3, 30]. Because of their biological stability, the glucosides are the most abundant components of cytokinin derivatives, making up 80 % or more of total cytokinin-related compounds in plants (Fig. 1b). These glucosides are thought to be sequestered in the vacuole. In addition, ribosidation and ribotidation by purine salvage pathway enzymes contribute to the cycling of cytokinins, through reverting active forms back to inactive precursors (Fig. 3). Thus, homeostasis of cytokinin activity is maintained by multiple metabolic systems.

The importance of cytokinins’ local action is best characterized by their role in maintenance of shoot apical meristem and root vascular development (Fig. 5). In Oryza sativa (rice), a single log mutation confers reduced shoot apical meristem size and abnormal floral organ development, although ten LOG-like genes are expressed in various organs [25]. Similarly, although the genes of the Arabidopsis LOG family are expressed in various types of tissues and organs and thought to be functionally redundant [32], the single log7 mutation causes reduction of shoot apical meristem size [26]. Rice LOG and Arabidopsis LOG7 are expressed in a restricted region of the upper part of the shoot apical meristem. Thus, loss of function of certain LOG genes expressed in this specific site is not fully complemented by other LOG genes expressed in other parts of the plant. These facts indicate that maintenance of shoot apical meristem activity requires site-specific activation of cytokinins.

In order to maintain shoot apical meristem, it is necessary to channel the cytokinin signal towards the organizing center [34]. In shoot apical meristem, active cytokinins are produced in the L1 layer by LOG4 and in the central zone by LOG7 [35, 36]. The organizing center and rib meristem are responsive to cytokinins via AHK4/WOL1/CRE1 (Fig. 5a) [36]. Endogenous cytokinins activate cytokinin signaling only in the organizing center but not in the rib meristem, although exogenously applied cytokinins activate both of these tissues (Fig. 5a) [36]. This indicates that the longitudinal diffusion of cytokinins to adjacent regions and local action in the restricted cells are important for proper meristem function.

The importance of local activation of cytokinins is also shown in the early stage of root vascular development. In root tip vasculature, which consists of procambium, protoxylem, metaxylem and phloem (Fig. 5b), cytokinin-activating LOG3 and LOG4 are specifically expressed in protoxylem cells, while cytokinin-perceiving AHK4/WOL1/CRE1 is expressed in the procambium [19, 20]. Analysis using marker genes for cytokinin response revealed that procambium cells adjacent to protoxylem respond to cytokinins more strongly than distant procambium cells [19, 20], indicating that local activation of cytokinins in protoxylem specifies cytokinin perception cells in procambium. Since cytokinins inhibit procambium-to-protoxylem differentiation, this regulation may be necessary for proper positioning of those cells during vascular development.

Root morphogenesis is abnormal in the log3 log4 log7 triple mutant, possibly due to reduced levels of active cytokinin [19]. Expression of LOG3 under a xylem precursor cell-specific promoter rescues this phenotype. However, ectopic expression of LOG3 in the phloem, through which cytokinins are systemically transported, does not. Therefore, it is suggested that the locally synthesized cytokinins are necessary for root morphogenesis.

Movement of cytokinins between organs has been shown by tracer experiments using isotope-labeled cytokinins [31, 37–39] and detection of cytokinins in vascular saps also supports translocation of cytokinins [40]. In general, long-distance signals are translocated through xylem and phloem, two major conduits for material transfer in plant vasculature. Intriguingly, these studies show that cytokinin species are unevenly distributed: tZ-type cytokinins are more abundant than iP-type in xylem sap and vice versa in phloem sap [40] (Fig. 4). These studies imply that iP-type and tZ-type cytokinins are directionally translocated and transmit different biological messages between organs.

Grafting experiments support the importance of cytokinins as long-distance signals. The Arabidopsis ipt1 ipt3 ipt5 ipt7 quadruple mutant shows severely reduced cytokinin content and shoot and root growth-deficient phenotypes [41]. Reciprocal grafting between the quadruple mutant and wild type rescued the growth-deficient phenotypes in shoot and root with accompanying recovery of cytokinin content [41], indicating a role of root-borne cytokinins in shoot and vice versa.

Grafting experiments also show the importance of root-borne tZ for normal shoot growth. In the cyp735a1 cyp735a2 double mutant, tZ-type cytokinin content is severely reduced without affecting total cytokinin quantity, and shoot growth is retarded [31]. When the shoot of the double mutant is grafted onto wild-type stock, the shoot phenotype is complemented with accompanying recovery of tZ-type cytokinin content, suggesting that side chain modulation of cytokinins has a specialized role in their long-distance translocation for shoot growth regulation.

As for the role of shoot-borne cytokinin in roots, impairing phloem transport destabilizes root vascular development with an accompanying reduction of basipetal cytokinin translocation in Arabidopsis [37]. Thus, shoot-borne cytokinins, via phloem, could participate in normal development of root vasculature in coordination with locally produced cytokinins.

In Lotus japonicas, root nodule number is reduced in the Ljipt3 cytokinin biosynthesis gene mutant [38], as well as in a Ljipt3 knockdown transgenic line [42]. Grafting between Ljipt3 shoot and wild-type root stock represses nodulation, while reverse grafting does not [38]. This suggests that shoot-derived cytokinins function as a repression signal of nodulation. It is notable that cytokinins synthesized in the root are not significantly involved in regulation of nodule number, although the mechanism through which the origin of cytokinins is determined remains to be elucidated.

So far, studies have shown that both xylem and phloem sap contain vast amounts of cytokinin ribonucleosides [40], suggesting that cytokinin ribonucleosides are translocated through the vasculature. It is believed that the cytokinin ribonucleosides are converted to their ribonucleotides followed by activation via LOG where they function.

An elaborate translocation system is necessary for the regulation of organ-to-organ communication via cytokinins. Recent studies identified ATP-binding cassette transporter subfamily G14 (ABCG14) as a key gene for appropriate root-to-shoot cytokinin translocation [39, 43]. In abcg14, tZ-type cytokinin contents are greatly reduced in xylem sap, and the dwarf phenotype of abcg14 is rescued in grafted plants between abcg14 shoot and wild-type root stock [43], indicating that ABCG14 is an essential gene for root-to-shoot translocation of cytokinins. Since the biochemical properties of ABCG14 have not been well characterized, the substrate of ABCG14 has not been identified. In addition to ABCG14, purine permease 1 and 2 (PUP1 and PUP2) and equilibrate nucleoside transporter (ENT) have been shown in in vitro studies to be involved in transport of cytokinins [44, 45]. However, their functions in planta, especially in long-distance translocation of cytokinins, are still poorly characterized.

Spatial regulation of LOG expression is one of the determining factors that specify the functional sites of cytokinins, and expression of each LOG family gene is regulated in a site-specific manner [25, 32]. Recently, a mechanism for this site-specific expression of LOG genes has been reported in flower development in Arabidopsis [46]. In a mutant of APETALA1 (AP1), a MADS-BOX transcriptional factor, abnormal floral organs are observed. AP1 directly represses LOG1 expression, and possibly in sepals where their expression overlaps. LOG1 repression, under control of the AP1 promoter, partially rescued the ap1 phenotype, indicating that sepal-specific LOG1 repression is required for normal flower development. In addition, expression of a LOG homolog is directly activated by KNOTTED1 (KN1) in maize [47]. It is thought that KN1 might provide site-specific regulation of the LOG homolog’s coordination with certain BEL1-LIKE HOMEOBOX (BLH) gene products, which interact with KN1 to bind DNA.

Recent studies have also revealed the importance of cytokinin oxidase/dehydrogenase (CKX) function as a metabolic attenuator of local cytokinin action. As well as LOG, the expression of each CKX family gene is regulated in a site-specific manner [33]. Loss-of-function mutants of specific CKX genes in rice (OsCKX2) and Arabidopsis (CKX3 and CKX5), which are expressed in reproductive meristem, cause increased cytokinin levels, leading to larger meristem size and an increase in reproductive organ number [48, 49], clearly showing that the CKXs fine-tune active cytokinin levels at the expression site. CKX3 expression in the organizing center and CKX5 broadly in the meristematic domain [49] may be particularly important for ensuring the site-specific cytokinin response (Fig. 5a). It is also intriguing that Arabidopsis AP1, which negatively regulates LOG1 expression, positively regulates CKX3 in the sepal [46]. This regulation is important for development of a determinate floral meristem.

Another candidate to control functionality of cytokinins is the family of AHK proteins. AHK promoter:reporter gene analyses revealed that the cytokinin receptors are expressed in various tissues in the plant, but that AHK3 is preferentially expressed in aerial organs, such as rosette leaves, AHK4/WOL1/CRE1 is expressed in root, and AHK2 is expressed in both [11]. Interestingly, it is reported that their affinities to ligands are different: AHK2 and AHK4/WOL1/CRE1 bind with similar affinity to tZ as well as iP, while AHK3 has less affinity to iP than to tZ [4, 50]. Thus, it is expected that tZ plays a major role as a cytokinin in shoots. This supports previous work with cyp735a1 cyp735a2 identifying a specialized function of tZ in shoot development [31].

Some other components of the cytokinin TCS also contribute to determining the sites of action of cytokinins. Cytokinins induce expression of WUSCHEL (WUS), a transcriptional factor expressed in the organizing center of the shoot apical meristem. WUS directly binds to the promoters of ARR5, ARR6, ARR7, and ARR15, and represses their expression (Fig. 5) [34]. This regulation establishes a local spatial domain for the organization of a stem cell niche in the shoot apex. (Fig. 5) [51].

Signaling systems of phytohormones build a network and mutually regulate signaling, transport, and metabolic systems. Interplay between cytokinins and auxin is one of the best-characterized cases of hormone–hormone interaction. Recent studies shed light on the importance of cytokinin–auxin interaction for auxin traffic and specification of cytokinin action sites in root vascular development. Cytokinin signaling in the procambium up-regulates expression of PINs, a family of auxin efflux carriers, and promotes their distribution in the plasma membrane from anticlinal to periclinal [52]. The bisymmetric distribution of PINs channels basipetally translocates auxin to protoxylem via the procambium. The auxin is perceived in the protoxylem and induces LONESOME HIGHWAY (LHW) and TARGET OF MONOPTEROS5 (TMO5). They directly induce LOG3 and LOG4, which activate cytokinins, and AHP6, which inhibits cytokinin responses in the protoxylem [19, 20]. Thus, the auxin-induced LOG3 and LOG4 in protoxylem provides cytokinins to adjacent procambium for proper root vascular development while the induced AHP6 inhibits cytokinin response (Fig. 5b) [19, 20]. Consequently, the procambial cytokinin response regulates PIN expression and distribution.

Cytokinins positively regulate agriculturally important traits such as grain size and biomass [3] but they also promote unfavorable phenotypes such as inhibition of root elongation. Therefore, spatio-temporal regulation of cytokinins is required for appropriate function in specific organs. In this article, we have discussed the spatial regulation of cytokinin biosynthesis and the role of these hormones in signaling. In order to better utilize cytokinin action to enhance beneficial traits of crops, a deeper understanding of their temporal regulation will be necessary.