The role of cis-zeatin-type cytokinins in plant growth regulation and mediating responses to environmental interactions

Abstract

Cytokinins (CKs) are well-established as important phytohormonal regulators of plant growth and development. An increasing number of studies have also reveal the function of these hormones in plant responses to biotic and abiotic stresses. While the function of certain CK classes, including trans-zeatin and isopentenyladenine-type CKs, have been studied in detail, the role of cis-zeatin-type CKs (cZs) in plant development and in mediating environmental interactions is less well defined. Here we provide a comprehensive summary of the current knowledge about abundance, metabolism and activities of cZs in plants. We outline the history of their analysis and the metabolic routes comprising cZ biosynthesis and degradation. Further we provide an overview of changes in the pools of cZs during plant development and environmental interactions. We summarize studies that investigate the role of cZs in regulating plant development and defense responses to pathogen and herbivore attack and highlight their potential role as “novel” stress-response markers. Since the functional roles of cZs remain largely based on correlative data and genetic manipulations of their biosynthesis, inactivation and degradation are few, we suggest experimental approaches using transgenic plants altered in cZ levels to further uncover their roles in plant growth and environmental interactions and their potential for crop improvement.

A brief history of the analysis of cis-zeatin derivatives

cis-Zeatin-type cytokinins (cZs) are a group of cytokinins (CKs) that have largely been ignored when compared to trans-zeatin (tZ) isomers or other highly active CKs. The lack of interest in cZs is mainly based on their lower activity in the classical CK bioassays. However, the research on cZs has also been restricted by the availability of appropriate methods to analyze their levels in plant tissues. From an analytical perspective, tZ and cZ-derivatives are mainly distinguishable by their chromatographic behavior. Therefore, the isolation and identification of cZs was tightly linked to the development of new methods in analytical chemistry that have been able to separate the different Z isomers. Although difficult to reconstruct, it’s likely that previous analyses based on low-resolution chromatographic methods likely reported the combination of both isomers in their reported zeatin (Z) levels, whereas some analyses might have also neglected them.

More than half a century ago, the main procedure to determine CKs was based on bioassays (reviewed in Letham, 1978). Tissue extracts were embedded in auxin-containing solid medium on which plant tissues were allowed to grow. The mass increase of the plant material in comparison to treatments with known substances (often kinetin) revealed cell division and therefore CK activity (Gyulai and Heszky, 1994). Zea mays caryopses, for example, were shown to contain high cell division activity and therefore one of the first CKs purified from these tissues was called Z (Letham, 1963). Mass spectrometry (MS) and nuclear magnetic resonance techniques allowed much better structural resolution and contributed to the identification of cis-zeatin riboside (cZR) in RNA extractions of plant tissues (Hall et al., 1967). In 1971, the chemical synthesis of cZ allowed comparisons between biological activities of cZ and tZ (Leonard et al., 1971). Soon thereafter, cZ was identified as the first bacterial Z isomer purified from cultures of the plant pathogen Corynebacterium fascians (Scarbrough et al., 1973). Trimethylsilyl derivatives of CKs for gas chromatography (GC)-MS improved the separation and quantitation of cZ from various tissues, such as wheat (Triticum aestivum) caryopses (Armstrong and Skoog, 1975) and hop (Humulus lupulus) cones (Watanabe et al., 1978). With the establishment of selected ion monitoring (SIM) in MS, it was possible to more accurately quantify different CKs from plant extracts, including cZR (e.g. Dauphin et al., 1979; Hashizume et al., 1979). Tay et al. (1986) noted that measurements of cZR could have resulted from enzymatic breakdown of tRNA during extraction. Using a modified extraction method, the authors concluded that cZR does not occur as free CK in tobacco shoots. Although it is now well established that cZR occurs as free CK in tobacco and many other plant species, this publication motivated various groups to re-evaluate their extraction procedures. As a result, extraction buffer systems, which minimize enzymatic reactions (Bieleski, 1964) were widely applied and are still used today.

In the late 80’s a significant improvement for the quantification of cZs and other CKs was achieved by using deuterium labeled internal standards. These internal standards demonstrated high concentrations of cZ, cZR, cis-zeatin-O-glucoside (cZOG) and cis-zeatin riboside monophosphate (cZRMP) occur in rice tissues (Takagi et al., 1989). The accuracy of CK analysis was further enhanced by the development of different derivatization strategies (e.g. Hocart et al., 1986; Letham et al., 1991). These methods also helped to characterize the first cZ-specific O-glucosyltransferase (cZOGT, Martin et al., 2001) and contributed to the identification of cZ derivatives as the most abundant CKs in chickpea (Cicer arietinum) tissues (Emery et al., 1998), or in specific organs, such as male flower buds of Mercurialis (Durand and Durand, 1994).

A major breakthrough in the analysis of CKs, including cZ derivatives was the establishment of highly selective and sensitive tandem MS techniques. While the chromatographic conditions in some reports do not distinguish between cis and trans-zeatins (Prinsen et al., 1995; Van Meulebroek et al., 2012), this was achieved by others (e.g. Dobrev et al., 2002; Mader et al., 2003; Novak et al., 2003; van Rhijn et al., 2001). Simplifications of the CK extractions (Dobrev and Kaminek, 2002) and the development of a high throughput analysis (Kojima et al., 2009; Schäfer et al., 2014a; Schäfer et al., 2014b) has added new impetus to the field and increased the number of publications that report the levels of cZ derivatives in plant and microbial sources.

Immunoassays were also developed to detect CKs, including cZs (Weiler, 1980). However, the practicability of these assays was often compromised by the relative cross reactivities of the antibodies to different CKs, which required chromatographic separation of extracts prior to the analysis (Wagner and Beck, 1993). Immunoaffinity co-purification of CKs provides a useful step in various CK extraction protocols. Such methods were coupled with other chromatographic methods to analyze cZ derivatives from various sources, including potato tuber sprouts (Nicander et al., 1995; Nicander et al., 1993). The first monoclonal antibodies against a cis-derivative of a plant CK were developed by Banowetz (1993). Immunopurification methods coupled with LC-MS analysis of CKs are still commonly used today (Novak et al., 2003) and contribute to the increasing number of publications regarding the levels of cZ and its derivatives as well as other CKs in tissues from various sources, including pea (Pisum sativum) roots (Stirk et al., 2008); algae (Chlorella minutissima, Stirk et al., 2011); moss (Physcomitrella patens) (Lindner et al., 2014; von Schwartzenberg et al., 2007) and the plant pathogen Rhodococcus fascians (Pertry et al., 2009).

Clearly the recent development of sensitive, rapid, high-throughput analytical methods has largely been responsible for our current understanding on the distribution and function of cZ-type CKs in plants. In this review, we provide an overview of the recent findings on the distribution, metabolism and activities of cZs in plants and interacting organisms.

Distribution of cZ type CKs

As above cZ-type CKs have been detected in plant species throughout the plant kingdom, but also in interacting organisms, such as bacteria (Scarbrough et al., 1973) and fungi (Strzelczyk et al., 1989). Gajdošová et al. (2011) analyzed the abundance of the cis and trans forms of Z and its derivatives in shoots and leaves of more than 150 representative species of different plant groups (Angiosperms, Bryophytes, Eudicots, Ferns, Gymniosperms, Lycophyta, Magnolliids and Monocots). Some plants such as Cryptomeria japonica (Gymniosperm, Cupressaceae) and Quercus robur (Angiosperm, Fagaceae) contain similar amounts of cis- and trans- Zs, but also plants with strong preferences for one of the isomers were found. In Ginkgo biloba (Gymniosperm, Ginkgoaceae) and Oenothera biennis (Angiosperm, Onagraceae) for example more than 90% of the Z occurs as trans-isomer, whereas in Pinus sylvestris (Gymniosperm, Pinaceae) and Urtica dioica (Angiosperm, Urticaceae) the cis-isomer dominated the CK spectrum. In maize and oat cZ-type CKs even strongly exceed the combined amounts of isopentenyladenine (IP)-, tZ- and dihydrozeatin (DHZ)-type CKs. Interestingly, these patterns are not associated with the evolutionary history of the plants. At present it is not clear, which properties are related to the abundance of the cis-isomer in particular plant species. It might be possible that specific environmental conditions (e.g., water and temperature regime, nutrient availability), biotic interactions (e.g., pathogen/herbivore pressure, plant-plant competition, mycorrhiza formation, interaction with nodulating bacteria) or the lifestyle (e.g., slow/fast growing, annual/perennial) of the plants are associated with the abundance of cZs. It is also possible that the high levels of cis-isomers in many crops (rice, Takagi et al., 1985; maize, Veach et al., 2003; pea, Quesnelle & Emery, 2007; potato, Nicander et al., 1995; chickpea, Emery et al., 1998; sweet potato, Hashizume et al., 1982) may be a result of the plant breeding process itself.

cZ metabolism

The biosynthesis and metabolism of CKs was described in detail elsewhere (e.g. (Frebort et al., 2011; Sakakibara, 2006). Here we will mainly focus on the specific requirements for the metabolism of cZ-type CKs (Fig.1). The rate-limiting step for CK biosynthesis is the prenylation of adenine nucleotides by isopentenyltransferases (IPTs). In plants, there are two possible isoprene sources: the cytosolic mevalonate (MVA) and the plastidic methylerythritol phosphate (MEP) pathway. Additionally, there are two alternative adenine substrates: ATP/ADP and tRNA; and the IPTs are accordingly classified as adenylate-IPTs and tRNA-IPTs. Evidence from Arabidopsis thaliana suggests a preference of IPTs to specific nucleotide substrates in combination with one specific isoprene source (Kasahara et al., 2004; Miyawaki et al., 2006). cZ-type CKs are synthesized via tRNA-IPTs (tRNA delta2 isopentenylpyrophosphate transferases; also known as IPPTs; EC 2.5.1.75, Kasahara et al., 2004). These enzymes can be found in all organisms except Archaea, while adenylate-IPTs have only been found in higher plants (Frebort et al., 2011). Prenylation by IPTs is not a random modification observed on all tRNAs. It targets a specific base, adenine 37, on the anti-loop of tRNAs of codons starting with uracil (i⁶A37; Persson et al., 1994; Taller, 1994 ). Studies in other organisms, including yeast, bacteria or mammalian cell culture have shown that prenylation of tRNA is important for translation, avoiding frameshifts and nonsense suppression of UAA (Guy et al., 2012; Laten et al., 1978; Waas et al., 2007). Interestingly, the role of tRNA modifications is often reported to be particularly apparent under specific conditions, such as stress or during interactions with other organisms (El Yacoubi et al., 2012). However, it is not clear whether the functions of prenylated tRNA are associated with those of free cZs in plants. With the exception of AtIPT4 and AtIPT7, which are localized in the cytosol and mitochondria, respectively, all Arabidopsis adenylate-IPTs are localized in plastids where MEP biosynthesis occurs. In contrast, the functional tRNA-IPT in Arabidopsis (AtIPT2) is localized in the cytosol (Kumari et al., 2013; Miyawaki et al., 2004; Miyawaki et al., 2006). Nevertheless, in silico analysis of tRNA-IPTs in monocot species, such as maize and rice, predict plastid localizations (Gramene, Phytozome v9.1, UNIPROT and chlorop v1.1; Emanuelsson et al., 1999). Similarly, all known IPTs in moss are tRNA-IPTs and PtIPT1 was shown to be localized in the chloroplast (Lindner et al., 2014). Knocking down PpIPT1 decreased cZ-type CK levels in moss, consistent with its role in cZ biosynthesis (Lindner et al., 2014). Future studies will reveal if the different localizations of tRNA-IPTs in different plant species are correlated with the amount and the importance of cZ-type CKs in these species.

Figure 1. Overview of cis-zeatin metabolism.

tRNA-isopentenyltransferases (tRNA-IPTs) catalyze the prenylation of adenine (A) 37 on specific (UNN-)tRNAs leading to the formation of isopentenyl adenine (IP)-containing tRNA. In A. thaliana, the isopentenyl group is derived from the mevalonate (MVA) pathway in the cytosol; but predicted localization of enzymes in other plants suggest the use of isoprene moieties derived from the methylerythritol phosphate (MEP) pathway (broken arrow). The MEP pathway also contributes to IP and trans-zeatin (tZ) biosynthesis. Once isoprenylated tRNA is synthesized, it can be further modified and the CKs can be released by unknown enzymes (“?” in a black box; e.g. tRNA degrading enzymes). Hydroxylation of the prenyl side chain is suggested to occur on the IP-containing tRNA (i⁶A37-tRNA), leading to the formation of cZ-containing tRNA (c-io⁶A37-tRNA). Inactivation and catabolism of cZ is mediated by cZ-O-glucosyltransferases (cZOGT), cZ-N-glucosyltransferases (cZNGT), and cytokinin oxidase/dehydrogenase (CKX). Glucosylated forms of cZ-type CKs: cZOG (cZ-O-glucoside), cZROG (cZ-riboside-O-glucoside) and cZ7G/cZ9G (cZ-N9/N7-glucosides). Multiple arrows indicate multiple biochemical steps; dotted-lines show unexplored metabolic flow. DMAPP (dimethylallyl-diphosphate), IPR (IP-riboside), cZR (cZ-riboside), A (adenine).

In addition, it is likely that isoprenylation by IPTs is not the rate-limiting step of cZ biosynthesis but rather the degradation of specific tRNAs. Consistent with this observation, tRNA-IPTs are constitutively expressed and not affected by plant hormones or nutrient status in Arabidopsis (Miyawaki et al., 2004; Miyawaki et al., 2006), even though cZ levels change under stress conditions (see about biotic and abiotic stress). Increased tRNA turnover has been found under various stress conditions (Fournier et al., 1976; Hopper et al., 2010; Lee and Collins, 2005; Phizicky and Hopper, 2010). Such stress conditions can also lead to increased cZs levels in various plant species (Fournier et al., 1976; Hopper et al., 2010; Lee and Collins, 2005; Phizicky and Hopper, 2010, see the following paragraphs about biotic and abiotic stress). Two of three known tRNA turnover pathways, rapid tRNA decay surveillance pathway (RTD) and endonucleolytic cleavage (EC) are induced upon stress, and are known to act specifically upon modified tRNAs (Alexandrov et al., 2006; Lee and Collins, 2005; Persson et al., 1994), such as isoprenylated tRNAs. Alternatively, stress-induced ribonucleases, as described for angiogenin and tiRNAs, could lead to the release of tRNA-derived CKs (Yamasaki et al., 2009). As mentioned before, ribonuclease activity in plant extracts can also contribute to the levels of free cZs. Future research should focus on the specific tRNA turnover pathways possibly involved in cZ release, and on the timing and mechanisms of hydroxylation of the isoprene moiety (Fig. 1).

Active CKs can be metabolized via oxidation by cytokinin oxidase/dehydrogenase (CKX; EC 1.4.3.18/1.55.99.12), or by the activity of glycosyltransferases. While the O-glucosylation of CKs, including cZ(R) is reversible and O-glucosides are generally considered as storage products (Mok et al., 1992), N-glucosides are thought to represent deactivation products (Letham et al., 1983; Vankova, 1999). The detection of cZ(R)-O-glucosides (cZOG, cZROG) and N-glucosides (cZ7G, cZ9G) indicated that cZ-type CKs are not mere tRNA degradation products, because glucosylated forms are not found in tRNA (Nicander et al., 1995; Takagi et al., 1989; Wagner and Beck, 1993). In 1979, Entsch et al. characterized a pea glucosyltransferase enzyme that conjugated cZ (although with lower activity than tZ, Entsch et al., 1979). O-glucosyltransferases with affinity for cZ are characterized in maize (Martin et al., 2001; Veach et al., 2003) and rice (Kudo et al., 2012). 7 and 9-N-glucosyltransferases with affinity for tZ and other CKs were identified in Arabidopsis (Hou et al., 2004), however their activity towards cZ was not tested. Limited knowledge also exists about the function of cZ degradation pathways via CKX. Recently, the Arabidopsis CKX1 and 7 were shown to have high preference for cZ (Gajdošová et al., 2011). Accordingly, overexpressing CKX7 highly decreased levels of free cZ(R) in Arabidopsis (Köllmer et al., 2014). In summary, this information illustrates that although many cZ-derived metabolites are commonly measured in various plant species, our knowledge of the genes that contribute to the regulation of cZ and derivatives is very limited.

It was proposed by Bassil et al., 1993 that cZ(R) and tZ(R) could be converted by a cis-trans isomerase as observed in beans. Although other studies could find no or only negligible conversion between the Z isomers in tobacco BY-2 cultures, oat leaves, rice seedlings and maize cultured cells (Yonekura-Sakakibara et al., 2004; Gajdošová et al., 2011; Kudo et al., 2012), it cannot be excluded that isomerization might occur under specific conditions or in particular tissues or plants. Additionally, it remains an open question if cZ regulation also relies on within-plant transport. cZR were reported as an abundant CKs in phloem sap of Arabidopsis (Hirose et al., 2008), but were also reported to occur in the xylem sap of Arabidopsis, wheat and oat (Parker et al., 1989; Hirose et al., 2008). Additionally, Arabidopsis purine permease 1 (AtPUP1) and AtPUP2 were proposed as potential transporters for various CKs including cZ (Burkle et al., 2003) and they may play a role in the loading and unloading required for long-distance transport. However, definitive functional studies on a potential role for long-distance transport of cZs remain to be done.

cZ perception and signaling

To activate the CK-specific phosphorelay, cZs should be able to bind and activate the CHASE-domain containing histidine kinases (CHKs), which serve as CK receptors. Indeed, it could be shown that cZs can bind to CHKs and activate downstream elements of the signaling cascade, although with different sensitivity depending on the plant species and the specific receptor (Lomin et al., 2011; Romanov et al., 2006; Spichal et al., 2004; Stolz et al., 2011; Yonekura-Sakakibara et al., 2004). The Arabidopsis CK receptors AHK2 and AHK3, for example showed higher cZ affinity when compared to its paralog AHK4/CRE1, however in all cases cZ affinity was several fold lower than of its trans-isomer (Romanov et al., 2006; Stolz et al., 2011). Crystal structure analysis of the AHK4-CHASE domain in complexes with CKs revealed that the hydroxyl-group of cZ in contrast to tZ cannot form an additional hydrogen bond with Thr294, which is likely the reason for the lower cZ affinity of this receptor (Hothorn et al., 2011). In contrast, the maize receptor ZmHK1, a closely related homolog of AHK4, was shown to have a similar sensitivity to cZ compared to tZ (Lomin et al., 2011; Yonekura-Sakakibara et al., 2004). Also the rice CK receptors OsHK3 and OsHK4 have a cZ affinity, similar to other tested CKs (Choi et al., 2012). Activity measurements with the P_ARR5::GUS reporter construct verified that cZ can activate the CK signaling cascade in Arabidopsis (Spichal et al., 2004). This was also confirmed by showing strong, tZ-comparable activity of cZ in eliciting the transcript accumulation of the maize response regulator ZmRR1 (Yonekura-Sakakibara et al., 2004) and of the rice response regulators OsRR1, OsRR2, OsRR6, and OsRR9/10 (Kudo et al., 2012). Interestingly, also PpHCK4 (from P. patens), a member of a recently discovered subgroup of CHKs, which was only found in the early diverging land plant Marchantia polymorpha and the moss P. patens, strongly responds to cZ (Gruhn et al., 2014). Additional differences were reported for the receptor affinity to CK-ribosides. While AHK4 does not respond to CK-ribosides (Yamada et al., 2001), ZmHK2 showed similar sensitivities to free bases and ribosides (Yonekura-Sakakibara et al., 2004). CHKs were reported to have a high degree of redundancy, but specific functions can also be mediated by a single receptor (Riefler et al., 2006). The differential receptor affinities of cZs might therefore allow functional specialization, as indicated already for tZ and IP (Stolz et al., 2011). Alternatively, cZs might also only function as modulator to fine-tune ‘general’ CK-pathway activity under specific conditions, but otherwise be functionally redundant with other CKs. How far species-specific differences in the cZ-affinities of CK receptors are related to a functional differentiation of cZs is currently unknown. The presence of receptors with a high cZs affinity could for example indicate a broader physiological role, whereas low affinity receptors, especially in combination with a low cis/trans ratio in some plants might indicate only subsidiary functions for cZs. cZ can compete with tZ for receptor binding in the bacteria assay (Romanov et al., 2006) and can partially antagonize tZ-induced chlorophyll accumulation in squash (Cucurbita maxima; Kuraishi et al., 1991). Moreover it was suggested that cZs might also play a role as a competitor to the more active CKs, thereby preserving specific CK functions that only require a low CK threshold (Gajdošová et al., 2011).

Roles of cZ in plant growth

CKs are well-known for their essential function in plant development and growth. cZs have long been thought to be biologically inactive and were considered as possible remnants of tRNA degradation (Skoog et al., 1966; Tay et al., 1986; Vreman et al., 1972; Vreman et al., 1978). Comparing the activity of cZ-type CKs with tZ- or IP-type CKs in classical activity assays (summarized in Gyulai and Heszky, 1994), such as Phaseolus (Mok et al., 1978) or tobacco cell-culture assay (Gajdošová et al., 2011; Leonard et al., 1971; Schmitz et al., 1972), it was revealed that cZs have little or no activity compared to IP and tZ, which are generally considered to be the most active natural CKs. Comparing the activities of cZs with their trans counterparts in various bioassays, Gajdošová et al. (2011) report in general between 3 and >50 times higher activities of the tZs (in accordance with their EC50 values), but the activities strongly depend on the particular bioassay. Several recent studies showed a developmental regulation of cZs in different model plants. In Arabidopsis, cZs are high in seeds and after imbibition (24h), low in growing young plants and increase again when plants stop growing and start to senesce (Gajdošová et al., 2011). Similarly, cZs concentrations are high during seed development in specific chickpea cultivars (Lulsdorf et al., 2013). Micropropagated plantlets of Musa have high levels of cZ, which were replaced by IP upon acclimatization (Aremu et al., 2014). In addition, cZs levels change significantly during development in the maize grain, as well as shoot and root tissues (Saleem et al., 2010; Zalabák et al., 2014). Dwarf hop varieties contain significantly higher amounts of cZs (Patzak et al., 2013) and cZR is a major CK in unfertilized hops (Watanabe et al., 1981). These results reveal that cZ-type CKs tend to accumulate under the particular circumstances associated with limited growth. However, the accumulation of cZ(R) during radicle emergence and early seedling establishment in Tagetes minutia (Stirk et al., 2005) also shows that cZs can be associated with fast-growth developmental stages. More data on the levels of specific CKs, including cZs, during the entire developmental phase of plants, instead of levels in very specific growth stages, are needed to draw general conclusions about their levels during plant growth.

Physiological processes influenced by CKs also include the inhibition of senescence (Gan and Amasino, 1995). cZ, also suppressed senescence-induction in maize leaves (Behr et al., 2012), and in an oat-leaf assay (Gajdošová et al., 2011), although with lower activity than tZ, but did not inhibit senescence of detached flowers of Dianthus (Upfold and Van Staden, 1990). cZ only slightly affected fruit development in Cucumis sativus (Ogawa et al., 1990). Decreasing the amount of cZ by overexpressing a cZOGTs in rice delayed leaf-senescence and led to short root phenotypes and bigger number of crown roots (Kudo et al., 2012). Suppressing cZ levels by overexpressing CKX7 also affected root development in Arabidopsis (Köllmer et al., 2014). Additional support for cZ functions in plants came from Arabidopsis T-DNA insertion lines with impaired cZ biosynthesis (atipt2 9; Miyawaki et al., 2006; Köllmer et al., 2014) which showed chlorotic phenotypes, a shortened primary root, likely a result of the reduced root meristem size and ectopic protoxylem formation. This suggests an active role of cZ- CKs in Arabidopsis and rice growth and development. However, the phenotypic changes in CKX7 overexpression lines might be more related to the reduction of cytosolic CK levels than to that of cZs in particular and the atipt2 9 plants might have been compromised by the reduced level of prenylated tRNA (Köllmer et al., 2014). In the moss P. patens, in which cZs are the major CK-type, PpIPT1 (tRNA-IPT) knockout plants with reduced levels of cZs also show developmental disturbances (Lindner et al., 2014), however, these might be caused by concomitant changes in levels of other active CKs.

Some studies suggest a role of cZs in dormancy and seed germination. cZR decreases after decapitation in released buds of Cicer arietinum and might be a possible inhibitor of lateral bud growth (Mader et al., 2003). In Brassica napus, cZR concentrations increased greatly after the onset of vernalization (Tarkowska et al., 2012). In the seeds of oat, lucerne, Tagetes and pea, cZs dominate the CK profile, suggesting a role in seed physiology (Stirk et al., 2005; Stirk et al., 2008; Stirk et al., 2012). In Lolium perenne, highly dormant seeds have been shown to have higher levels of cZR compared to seeds at a less dormant stage (Goggin et al., 2010). cZs were also found at higher levels in dormant potato tubers compared to non-dormant ones and injection of cZ induced premature sprout formation (Mauk and Langille, 1978; Suttle and Banowetz, 2000).

In summary, these reports suggest that cZs might play a role during times of limited growth or dormancy, as is found in buds, tubers and seeds. It was proposed by Gajdošová et al. (2011) that cZs could help to maintain a basal level of CK activity under these conditions, but experimental proof remains lacking.

Roles of cZ in abiotic stress

As mentioned before, the function of cZ(R) in plants was assumed to be the maintenance of minimal CK activity (e.g., ensuring resource supply by minimal sink activity or a suppression of premature leaf-senescence) under growth-limiting conditions, including abiotic stresses. Due to the high energy requirements during stress adaptation, effective stress responses (at least at the early stage) are associated with suppression of growth and reallocation of resources to the formation of stress-protective compounds. Probably thus, in this period, tZ(R), which exhibits very high cell division promoting activity, is often replaced by the much less active cZ(R). These CK dynamics may be illustrated by various temperature stress experiments, but also in the case of biotic stresses (see below). Adjustment of the CK pool during the cold stress is tissue specific. Transfer of winter wheat to 4°C was associated with a rapid increase of cZ(R) in the main meristematic tissues crucial for over-wintering, namely crowns, in which maximum levels were attained already during the early stress response (after one day of cold exposure) (Kosova et al., 2012). Simultaneously, the majority of the main protective proteins – dehydrins (especially WCS120) accumulated. In leaves, mild, gradual elevation of cZ(R) was detected, attaining maximum concentrations after three days of cold. The levels of tZ(R), however, dropped almost immediately upon cold exposure. During the subsequent acclimation phase, a moderate increase of tZ(R) was detected, associated with the plant’s adaptation to low temperature. A peak of cZR (representing approximately 88% of the active CKs) was found in shoot apices of Brassica napus plants also after prolonged incubation at a low temperature (Tarkowska et al., 2012).

This cZR peak, however, seems to be associated with the onset of the transition between vegetative to generative developmental phases, as mentioned in previous sections. Similarly, the peak of predominantly cZR was found at the very beginning of the vegetative to reproductive developmental transitions in cold treated Triticum monococcum (Vankova et al., 2014). Association of the cZ(R) peak with developmental changes was confirmed by comparison of the spring and the winter lines. In the spring wheat line, without vernalization requirement, a peak of cZR was detected in leaves, crowns and roots after 21 days at 4°C, while in the winter line, the maximum of cZR was delayed, occurring after the fulfilment of vernalization requirement (42 days).

Strong increases in cZR, together with moderately active IP and IPR, was found in young pea leaves after 4-day cold stress as well as after prolonged heat stress (Vaseva et al., 2009). Elevation of cZ(R) concentrations with simultaneous down-regulation of tZR was reported at the early phase of heat stress response in leaves as well as in roots of tobacco plants (Dobra et al., 2010). This change in CK pool coincided with high expression of heat shock factors and heat stress associated proteins. Thus, response to both temperature extremes seems to affect cZs/tZs profiles.

The effect of salt stress on the dynamics of cZ/tZ CKs was tested in maize plants (Vyroubalova et al., 2009). In roots, tissues directly exposed to the salt stress, rapid elevation of the cZ precursor, cZRMP, was followed by elevation of cZ and especially of cZR, which reached maxima after 3 h (coinciding with minimum tZ and tZR levels). In leaves, the peak of cZR also coincided with the minimum of tZ. After 3 days, when acclimation took place, levels of cZ-type CKs were down-regulated, while peaks of tZ and tZR were detected. Thus in salt stress, rapid down-regulation of growth is also associated with cZ/tZ changes. Response to salt stress may be also affected by other external factors, e.g. by the CO₂ levels (Pinero et al., 2014), sweet pepper (Capsicum annuum) plants at lower CO₂ content exhibited almost doubling cZR levels in comparison with high CO₂ supplementation.

During drought stress, increase of cZ levels was detected in roots (Havlova et al., 2008; Mackova et al., 2013). The levels of cZ(R) were highly up-regulated in tobacco roots also in response to combined drought and heat stress. After re-watering, cZs were down-regulated, with a simultaneous increase of tZ-type CKs and a stimulation of growth, even to a higher extent than in control (non-stressed) plants. Up-regulation of cZs was also found as a crucial response of Plectranthus ambiguus to nitrogen deficiency (Papparozzi, personal communication).

The above mentioned patterns of cZ and its riboside suggest a role in maintenance of certain physiological functions of CKs under stress or growth-limiting conditions. The described mechanism should, however, be treated with caution, as some plant species do not respond with an elevation of cZs (e.g. soybean, Le et al., 2012) and until now, experimental tests that analyse abiotic stress resistance of plants with manipulated cZ levels are lacking. Additionally, plants with generally high levels of cZs (e.g., maize, Veach et al., 2003) indicate that cZs are not only involved in stress responses and during periods of growth limitations.

Roles of cZ in pathogen resistance

Recently, a function of CKs in the regulation of plant immunity against pathogens has been identified. Several active CKs, including 6-benzylaminopurine (6-BAP), kinetin and tZ have been demonstrated to efficiently increase resistance against hemi- /biotrophic pathogens in Arabidopsis and tobacco (Argueso et al., 2012; Choi et al., 2010; Großkinsky et al., 2013; Großkinsky et al., 2011). In these CK-mediated resistance phenotypes, interactions with other phytohormones, such as abscisic acid (Großkinsky et al., 2014) or salicylic acid (Argueso et al., 2012; Choi et al., 2010; Großkinsky et al., 2011) have been shown. In contrast, information on the role of cZ in plant immunity is limited. Pre-treatment with cZ can considerably suppress symptom development of Pseudomonas syringae infection in cultivated tobacco (Großkinsky et al., 2013). Similar effects were observed in cZ pre-treated or AtIPT2 expressing (SAG-IPT2) plants of the wild tobacco (N. attenuata, Fig. S1). The cZ effect on P. syringae symptoms in N. tabacum was, however, significantly lower compared to the highly active tZ. Concomitantly, cZ had no effect on the in planta proliferation of the pathogen as it had been shown for more active CKs including tZ (Großkinsky et al., 2013; Großkinsky et al., 2011). These data indicate that cZ does not directly activate anti-pathogen defense in these plant species, but mainly suppress symptom development, e.g., by suppressing the pathogen-induced cell death response similar as described by Barna et al. (2008) for Z and thereby maintain tissue integrity during an infection.

Interestingly, as mentioned before, the ability to produce cZs (among others) has been identified in several pathogens. No specific biological role has been attributed to cZs found in Magnaporthe grisea hyphae and its culture filtrates as well as in rice tissue post M. grisea infection (Jiang et al., 2013). However, they correlated with the proliferation and symptom maintenance of R. fascians infection in Arabidopsis (Pertry et al., 2009). This was consistent with the considerably lower amounts of cZs (in addition to others) produced in culture by the non-pathogenic R. fascians strain D188-5 compared to the virulent strain D188. Furthermore, production as well as interconversion of CKs, including cZ and derivatives, has been described for Colletotrichum graminicola (Behr et al., 2012). The biological relevance of the modulation of cZ levels by fungal production and conversion has further been related to the infection process. Comparable to C. graminicola infection, treatment of maize leaves with cZ (but also tZ or 6-BAP) resulted in delayed senescence, as evidenced by the formation of photosynthetically active green islands (Behr et al., 2012) indicating the potential modulation of the host physiology by the fungus via cZ production. This physiological modulation could also include the regulation of the host’s carbohydrate metabolism as the CK-related delay of senescence is mediated by invertase activity (Balibrea Lara et al., 2004). Since invertases play important roles for the tolerance/resistance against biotic (Roitsch et al., 2003) as well as abiotic (Albacete et al., 2014) stress situations, they could be an important target for the specific modulation of host physiology via cZ in plant-microbe interactions.

Role of cZ in herbivore resistance

In addition to their potential role in abiotic stress responses and pathogen resistance, cZs are also indicated to play a role in plant-herbivore interactions. Very high cZ levels in the larval body of the aphid, Pachypsylla celtidis indicate that they might be involved in induced gall formation in hackberry (Celtis occidentalis, Straka et al., 2010). Insects and their endosymbiotic bacteria were reported to utilize CKs to manipulate a plant’s physiology to their advantage and it seems possible that cZs could also be used in combination with other CKs for this purpose (Giron and Glevarec, 2014). In addition to herbivore-mediated manipulations plant-mediated responses to herbivory were shown to involve cZs. Conrad and Kohn (1975) showed a wound-induced formation of Z-containing tRNA, potentially the cis-isomer; however the cis/trans conformation was not further specified in this study. Recently, Schäfer et al., (2014a) showed that cZs were upregulated in N. attenuata and Arabidopsis by wounding and application of oral secretions from the tobacco hornworm (Manduca sexta) or the grasshopper, Schistocerca gregaria, respectively. cZR levels responded in both species, while cZ was much more responsive in Arabidopsis. Even 4 h after treatment the cZR levels were still highly upregulated in N. attenuata, while the levels of another herbivory-induced, bioactive CK, IPR, already started to decline. Jasmonic acid (JA) is a key player in the plant response to chewing insect herbivores. Although methyljasmonate (MeJA) application to N. attenuata leaves reduced IPR levels and suppressed the herbivory-induced CK-pathway signaling (indicated by NaARR5 transcripts), the cZR levels were elevated by this treatment. Accordingly, silencing JA biosynthesis and signaling components reduced the herbivory-induced accumulation of cZR. It seems likely that the contrasting regulation of cZR and IPR by JA could be related to their distinct metabolic origin (Kasahara et al., 2004; Miyawaki et al., 2004). Herbivory, as well as jasmonates were reported as potent suppressors of plant-growth (Attaran et al., 2014; Hummel et al., 2007; Meldau et al., 2012), which is consistent with the observation that cZ are particularly associated with growth-limiting conditions.

In recent years, CKs were shown to amplify plant defense responses against herbivores (Dervinis et al., 2010; Smigocki et al., 1993). Therefore cZs might also be involved in the regulation of anti-herbivore defenses. JA-mediated defense responses, such as proteinase inhibitor accumulations, were found to be promoted by CKs (Dervinis et al., 2010). When we applied cZR to N. attenuata leaves we found an increase in the MeJA-mediated induction of the phenolamide pathway and trypsin proteinase inhibitor activity (Fig. S2). These data suggest that cZs are potentially involved in defense metabolite accumulations after herbivore attack. However, external applications do not allow for the precise regulation of intracellular CK levels and additional work is necessary to ensure that the observed effects can be triggered by physiologically relevant levels of cZs.

Secondary metabolites are known to play an important role in various stress responses (Bennett and Wallsgrove, 1994; Rangan et al., 2014) and it should be tested if cZs can also amplify the accumulation of other secondary metabolites, potentially to improve plant survival under stress conditions. The frequently observed mild CK activity of cZs (Kamínek et al., 1979; Schmitz and Skoog, 1972) might help plants to retain the resources required for the induction of the defense and stress-resistance metabolites.

Outlook

The CK pathway has frequently been predicted to harbor opportunities for future crop improvement (Qin et al., 2011; Werner et al., 2010; Wilkinson et al., 2012; Yang et al., 2000). However, most of these studies gloss over the differences in specific CK profiles that exist between model and crop plants. cZs are highly abundant in many crop plants. In contrast, in Arabidopsis the trans-isomer is the most abundant in most growth stages (Gajdošová et al., 2011) and its CK receptors have a lower affinity to cZs than has been reported for some of the CHKs of crop plants (Yonekura-Sakakibara et al., 2004; Romanov et al., 2006; Lomin et al., 2011; Stolz et al., 2011; Choi et al., 2012). Therefore the functional predictions based on previous investigations in low cZ-containing plants might not reflect the role of cZs in many crops.

Environmental factors can have a tremendous effect on the agricultural productivity (Boyer, 1982). Various references indicate that cZs are part of the plant response to growth under limiting conditions and hence they might be potential targets to improve the biotic and abiotic stress resistance of crop plants (Fig. 2). Moreover their role during plant development could also help improve crop properties. For example, cZs were proposed to play a role in regulating potato tuber dormancy (Suttle & Banowetz, 2000), which makes cZs a potential breeding targets to produce plants with improved tuber storage characteristics.

Figure 2. cis-zeatins as potential regulators of plant development and stress responses.

cis-zeatin (cZ) and its riboside (cZR) are reported to accumulate in particular under various conditions characterized by limited growth, during particular developmental stages, but also in response to abiotic and biotic stresses. cZ/cZR were shown to be involved in the regulation of the plant development and to be able to modulate plant defense responses. Based on their distribution patterns, they were additionally proposed to sustain a minimum cytokinin (CK) activity under growth-limiting conditions, to prevent the redirection of resources e.g., from stress adaptation processes to plant growth and for a means by which phytophagous organisms could manipulate the plant’s physiology and morphology for their bbenefit. However, an experimental confirmation of this hypothesis is still missing.

Unfortunately, experimental proof of cZ functions remains rare and many assumptions need deeper and more rigorous experimental examination. Experiments that use plants with manipulated cZ levels, e.g., by external application of cZs (e.g., Großkinsky et al., 2013), impaired cZ-biosynthesis (atipt2 9 mutants, Miyawaki et al., 2006) or increased cZ-degradation/ inactivation (AtCKX7 overexpression, Köllmer et al., 2014; OscZOGT1 and OscZOGT2, Kudo et al., 2012) are sorely needed to unravel the role of cZs in plant-growth and stress responses. Additionally, plants overexpressing cZ-biosynthetic genes (e.g., AtIPT2) and forward genetic approaches to identify regulatory elements of the cZ metabolism could also be illuminating. Importantly, the experimental side-effects of the manipulations, such as the changes in prenylated tRNA or the effects on other CKs must be taken into consideration (e.g. Köllmer et al., 2014). For the analysis of stress-specific functions, the use of conditional expression systems (reviewed in Corrado and Karali, 2009) can be very useful to disentangle stress responses from changes in development. The knowledge gained from these experiments about this widely distributed, but often neglected hormone will help us understand if it plays a role as ‘stress hormone’ under growth-limiting conditions and as a mediator of responses to a plant’s interactions with other organisms, including attackers, as well as mutualists.

Summary.

cis-Zeatin is often characterized as low-active cytokinin, but its widespread distribution, complex metabolism and specific regulation, suggests many possible functions. We summarize and discuss these observations and recent experimental evidences.