Mutual interactions of the phytohormones, cytokinins and auxin determine root or shoot identity during postembryonic de novo organogenesis in plants. However, our understanding of the role of hormonal metabolism and perception during early stages of cell fate reprogramming is still elusive. Here we show that auxin activates root formation, whereas cytokinins mediate early loss of the root identity, primordia disorganisation and initiation of shoot development. Exogenous and endogenous cytokinins influence the initiation of newly formed organs, as well as the pace of organ development. The process of de novo shoot apical meristem establishment is accompanied by accumulation of endogenous cytokinins, differential regulation of genes for individual cytokinin receptors, strong activation of AHK4-mediated signalling and induction of the shoot-specific homeodomain regulator WUSCHEL. The last is associated with upregulation of isopentenyladenine-type cytokinins, revealing higher shoot-forming potential when compared with trans-zeatin. Moreover, AHK4-controlled cytokinin signalling negatively regulates the root stem cell organiser WUSCHEL RELATED HOMEOBOX 5 in the root quiescent centre. We propose an important role for endogenous cytokinin biosynthesis and AHK4-mediated cytokinin signalling in the control of de novo-induced organ identity.
The ability to undergo postembryonic de novo organogenesis is one of the most important adaptation strategies of plants. The capacity to form new tissues and organs during postembryonic development is present in cells with high regeneration potential. These cells are under the control of regulators that mediate specific spatiotemporal changes in their highly responsive developmental programmes. Phytohormones, auxin and cytokinins have long been known to be key factors controlling de novo organ formation from plant explants (Skoog and Miller, 1957). In the presence of auxin alone or in media with a high auxin-to-cytokinin concentration ratio, growth of the roots is initiated from various plant tissues (Atta et al., 2009; Pernisová et al., 2009; Sugimoto et al., 2010). However, if the auxin-to-cytokinin ratio is reversed (i.e. if the cytokinin concentration is higher than that of auxin), shoots are formed (Skoog and Miller, 1957).
Recently, substantial progress has been achieved in understanding the molecular mechanisms underlying the ability of plants to induce and respecify the identity of newly formed organs. In Arabidopsis, the two-step protocol for de novo organogenesis is well established (Che et al., 2006; Sugimoto and Meyerowitz, 2013; Valvekens et al., 1988). In the first step, the competence of an explant to regenerate organs is acquired by cultivation on auxin-rich callus-inducing medium (CIM). Importantly, the organ primordia induced during the cultivation on CIM from the explants of the shoot origin share plenty of anatomical and molecular determinants with roots and have been shown to be induced via the root-specific developmental pathway (Atta et al., 2009; Sugimoto et al., 2010). Subsequent cultivation of the organogenesis-competent calli on auxin-supplemented root-inducing medium (RIM) promotes growth of roots. Compared with this, the high cytokinin content in the shoot-inducing medium (SIM) promotes shoot regeneration (Atta et al., 2009; Kareem et al., 2015; Sugimoto et al., 2010; Valvekens et al., 1988). Alternatively, de novo shoot apical meristem formation can occur directly via redifferentiation of auxin-induced lateral root primordia into shoot meristems by cytokinin application without intermediate callus formation (Chatfield et al., 2013; Kareem et al., 2016; Rosspopoff et al., 2017).
In plant postembryonic development, growth is largely controlled by the activity of apical (shoot and root) and lateral (procambium, cambium and axillary) meristems. Meristems maintain a pool of totipotent (stem) cells that divide and differentiate, allowing for the formation of new plant organs and tissues (reviewed by Greb and Lohmann, 2016). In the shoot apical meristem, the localisation of undifferentiated cells in the stem cell niche is controlled by the expression of homeodomain transcription factor WUSCHEL (WUS) in the organising centre (Mayer et al., 1998; Schoof et al., 2000). WUS function is essential for shoot apical meristem development (Gallois et al., 2004; Gordon et al., 2007) and wus mutants fail to regenerate shoots in vitro (Zhang et al., 2017). Similarly, the WUS homologue WUSCHEL RELATED HOMEOBOX 5 (WOX5) regulates root stem cell maintenance in the stem cell niche adjacent to the quiescent centre, an analogue of the shoot organising centre in the root apical meristem and the location of WOX5 expression (Blilou et al., 2005; Pi et al., 2015; Sarkar et al., 2007). Expression and differential regulation of the WUS gene family starts early during embryogenesis (Haecker et al., 2004), and WUS and WOX5 can be considered as shoot- and root-specific reporters, respectively.
Cytokinins control plant development by activating a multistep phosphorelay signalling pathway. Cytokinins recognise the CHASE domain of the sensor histidine kinases ARABIDOPSIS HISTIDINE KINASE (AHK)2, AHK3 or AHK4/WOL/CRE1 (Higuchi et al., 2004; Inoue et al., 2001; Mahonen et al., 2000; Nishimura et al., 2004; Ueguchi et al., 2001). The signal is subsequently transferred via small cytoplasmic ARABIDOPSIS HISTIDINE-CONTAINING PHOSPHOTRANSFER proteins (AHPs) (Hutchison et al., 2006) to ARABIDOPSIS RESPONSE regulators (ARRs) (Suzuki et al., 1998; Tanaka et al., 2004). Type B ARRs (ARRs-B) are GARP transcription factors controlling the expression of cytokinin-regulated genes, including type-A ARRs (ARRs-A) (Mason et al., 2004; Sakai et al., 2001). The target binding sequence of ARRs-B served to develop the cytokinin-responsive two-component signalling (TCS) reporter (Müller and Sheen, 2008; Zurcher et al., 2013). ARRs-A are cytokinin primary response genes that are quickly upregulated by cytokinins even in the absence of translation (Brandstatter and Kieber, 1998; D'Agostino et al., 2000). In parallel, ARRs-A inhibit the cytokinin signalling pathway, generating a negative-feedback loop (Hwang and Sheen, 2001; To et al., 2004).
The levels of the two major types of endogenous cytokinins, N6-(Δ2-isopentenyl)-adenine (iP) and trans-zeatin (tZ) are under control of complex metabolic network. In Arabidopsis, the initial step of cytokinin biosynthesis is catalysed by isopentenyltransferase (IPT), and involves the transfer of a prenyl moiety from dimethylallyl diphosphate to adenosine triphosphate or diphosphate to form iP ribotides (Kakimoto, 2001; Takei et al., 2001). iP ribotides could be hydroxylated by the cytochrome P450 mono-oxygenases CYP735A1/CYP735A2, resulting in the formation of tZ-type cytokinins (Takei et al., 2004). Direct conversion of cytokinin ribotides to active free bases is catalysed by a nucleoside 5′-monophosphate phosphoribohydrolase named LONELY GUY (LOG) (Kurakawa et al., 2007; Kuroha et al., 2009). Besides the biosynthesis, the level of active cytokinins is regulated by glycosylation and other conjugations of cytokinin bases. These metabolites serve as storage, transport and inactivated cytokinin forms. tZ-type cytokinins with a hydroxylated side chain can be reversibly glycosylated by the UGT85A1 enzyme (Hou et al., 2004). O-glycosylation is one of the first reactions to exogenous cytokinin application and produces non-active cytokinins that can be re-activated by β-glucosidases (Brzobohaty et al., 1993). Cytokinin O-glucosides could serve as a storage pool and readily available source of active free bases (Kiran et al., 2012). On the other hand, cytokinin glycosylation at the N7 and N9 positions of the adenine moiety catalysed by UGT76C1 and UGT76C2 enzymes (Hou et al., 2004; Wang et al., 2011) is irreversible, and therefore N-glucosides were proposed to represent a permanent deactivation in the cytokinin lifetime (Parker and Letham, 1973). Irreversible cytokinin degradation is catalysed by the cytokinin oxidase/dehydrogenase (CKX) (Schmülling et al., 2003) that cleaves unsaturated N6-side chains from tZ- and iP-type cytokinins (Jones and Schreiber, 1997). Biosynthesis, modifications and degradation maintain cytokinin homeostasis, balancing optimal phytohormone levels during plant growth and development.
Here, we present a model describing the role of endogenous cytokinin biosynthesis and cytokinin signalling in the regulation of de novo organ formation. Using the one-step hypocotyl explant assay, we demonstrate that cytokinins inhibit auxin-induced organ establishment and mediate the early disorganisation of organ primordia that is associated with switch in the primordia identity. We show that the de novo organogenesis is associated with differential regulation of cytokinin metabolism and perception, and that the effects of endogenous cytokinins are mediated via a specific cytokinin receptor to control de novo organ formation.
Cytokinin induces early root-to-shoot organ respecification
Using a one-step protocol, it has been previously demonstrated that auxin induces de novo root formation along hypocotyl explants, whereas cytokinins alone are not able to induce any organogenic response. However, cytokinins are able to modulate the auxin-induced organogenesis in a concentration-dependent manner (Pernisová et al., 2009), leading to shoot formation after long-term cultivation. Here, we focused on early events (the first 5 days of cultivation), when the decision to form a root or a shoot takes place. Hypocotyls of 5-day-old etiolated Arabidopsis seedlings were excised and placed on media supplemented with a range of phytohormone concentrations; cytokinins were represented by kinetin and auxins by 1-naphthaleneacetic acid (NAA). Two types of cytokinin-containing media were used: K300, containing 100 ng/ml of NAA and 300 ng/ml of kinetin (the lowest tested cytokinin concentration that leads to the loss of root morphology of the auxin-induced organs; Pernisová et al., 2009); and K1000, with the same NAA concentration (100 ng/ml) and 1000 ng/ml of kinetin. Using our protocol, both K300 and K1000 were able to induce shoot formation, the latter with higher efficiency (Fig. S1). The non-natural cytokinin kinetin (Kamínek, 2015) was selected to allow the possible changes in endogenous cytokinin levels to be distinguished during the organogenic response. Moreover, CKX activity has been found to be upregulated in response to exogenous cytokinin application (Gelová et al., 2018; Pernisová et al., 2009; Werner et al., 2006) and kinetin has been described as a poor substrate for CKX enzymes (Frébortová et al., 2004; Galuszka et al., 2007; Popelková et al., 2006). Thus, exogenously applied kinetin is not considered to be cleaved by endogenous CKX activity, or by CKX3 in Pro35S:AtCKX3 line, allowing us to determine functional importance of differences in endogenous cytokinin levels during the organogenic response.
The auxin-induced primordia were recognised as early as the first day of cultivation of hypocotyl explants on auxin-supplemented medium (K0: kinetin 0 ng/ml+NAA 100 ng/ml; Fig. 1). Similar to the results achieved when using a two-step protocol (Sugimoto et al., 2010), the root identity of the auxin-induced primordia in our one-step setup was confirmed by the early activation of the root-specific reporters PLETHORA (PLT) 2, PLT3 or SCARECROW (SCR) at day 1 and WOX5 at day 2 (Fig. 1A, Figs S2, S3). In contrast, on the cytokinin-supplemented shoot-inducing media (K300, K1000), the signal of ProWOX5:GFP as well as the other root-specific markers was downregulated and became undetectable in media with high cytokinin concentration (K1000) latest from the day 3 onwards (Fig. 1A, Figs S2 and S3). Simultaneously, the root primordia changed their morphology, losing their characteristic root primordia pattern. The shoot-specific ProWUS:tdTomato signal appeared solely in disorganised primordia at the day 3 of cultivation (Fig. 1A), suggesting primordia identity respecification. Neither WUS signal outside the organ primordia and/or in the well-organised primordia, nor its co-occurrence with WOX5 signal was detectable in hypocotyl explants carrying both WOX5 and WUS reporters on all media tested. Well-differentiated shoots with newly formed leaves were easily recognisable after 21 days of cultivation on media with both cytokinin concentrations (K300 and K1000, Fig. S1).
To investigate auxin and cytokinin signalling outputs during early events leading to organ identity respecification, we prepared a line carrying the auxin signalling reporter pDR5rev::3XVENUS-N7 (Wabnik et al., 2013) together with the cytokinin-responsive two-component signalling sensor TCSn::GFP (Zurcher et al., 2013) (pDR5rev::3XVENUS-N7/TCSn::GFP). On the medium with auxin alone (K0), auxin signalling was upregulated throughout the newly formed primordia on day 1 (Fig. 1B). In later development, DR5 signal shifted to the columella cells of the newly formed root tip, mimicking the situation in the intact root. However, a broad DR5 expression domain remained detectable at the base of the newly formed roots. The cytokinin signalling-responsive TCS signal was first detectable in the vascular tissue at the base of the primordia at day 2 and became delimited to the vasculature and columella/lateral root cap of the newly formed roots. Cytokinin concentration-dependent downregulation of auxin signalling in the organ primordia was apparent even at the first day of cultivation at K300/K1000. Starting with day 3, the loss of clearly specified DR5 maxima was associated with gradual primordia disorganisation on both cytokinin-rich media types. The cytokinin signalling reporter TCSn::GFP was strongly induced in disorganised primordia from the day 3 onwards (Fig. 1B). Moreover, high level of TCSn::GFP expression colocalised with ProWUS:tdTomato in disorganised primordia, as assayed in the TCSn::GFP/ProWUS:tdTomato line (Fig. 1C).
Altogether, we show that, similar to previous reports employing the two-step protocol (Atta et al., 2009; Kareem et al., 2015; Sugimoto et al., 2010; Valvekens et al., 1988), the one-step procedure allows the regeneration of shoots from hypocotyl explants from the auxin-induced organ precursors of root identity. Exogenous cytokinins quickly downregulate auxin signalling in a concentration-dependent manner and induce early loss of the primordia root identity. The process of cytokinin-induced organ respecification is connected with loss of the primordia organisation and strong activation of cytokinin signalling.
Cytokinins inhibit de novo primordia initiation and differentiation, and promote in vitro primordia disorganisation
The impact of both endogenous and exogenous cytokinins was examined using a detailed morphological analysis of de novo-induced organs. In the wild type, all primordia were primed during the first 24 h of explant cultivation and no increase in the total primordia number was detected during further cultivation up to day 5 (Fig. 2A). Adding exogenous cytokinin into medium (K300) strongly decreased primordia number at day 1, suggesting negative cytokinin effect on primordia initiation (Fig. 2A, Table S1A). To evaluate the further progress of primordia development, individual developmental stages were scored by employing the classification previously introduced for the lateral root formation (Malamy and Benfey, 1997) (Fig. S4). On day 1, the distribution of individual developmental stages per hypocotyl explant was comparable with or without cytokinin (Fig. 2B, Table S1B). Hereafter, on media without cytokinin (K0), approximately half of the primordia reached developmental stages V+ at day 3 and maturated to roots at day 5. In comparison, on the shoot-forming media (K300), most of the primordia stopped growth after reaching stage IV. Only some of them grew further, giving rise to disorganised primordia and very few roots (Fig. 2B).
The role of endogenous cytokinins in the de novo organogenic response was examined in the Pro35S:AtCKX3 stable transgenic line with depleted endogenous cytokinins via overexpression of CKX gene (Pernisová et al., 2009). From day 1 onwards, we detected a higher number of primordia along the hypocotyl explants in Pro35S:AtCKX3 line when compared with wild type, suggesting a negative role for endogenous cytokinins in primordia initiation (Fig. 2A, Table S1C). Furthermore, the higher frequency of older primordia stages on media without cytokinins (K0) indicated accelerated primordia growth in the Pro35S:AtCKX3 line when compared with wild type at all the tested time points (statistically significant differences were scored for category III+ at day 1 and V+ at day 3 and 5 (Fig. 2B, Table S1D). When compared with K0 and K300, similar to the situation observed in wild type, Pro35S:AtCKX3 lines also revealed cytokinin-mediated downregulation in the frequency of primordia at stages V and older (Fig. 2B, Table S1E). Moreover, the percentage of disorganised primordia was decreased (but statistically insignificant) in the Pro35S:AtCKX3 line (Fig. 2B, Table S1F).
Taken together, these data indicate that both exogenous and endogenous cytokinins affect de novo organogenesis by suppressing primordia initiation and primordia differentiation, particularly at transition beyond stage IV. Importantly, endogenous cytokinins also seem to contribute to cytokinin-induced primordia disorganisation.
De novo shoot apical meristem establishment is accompanied by upregulation of iP-type cytokinins
To achieve a more detailed insight into the regulatory role mediated by endogenous cytokinins, the expression and morphology analyses were complemented by measurement of endogenous cytokinin content in hypocotyl explants. The total cytokinin content steadily increased during the first 5 days of cultivation. However, a large increase was apparent particularly at the third day of cultivation (Fig. 3A, Table S2), correlating with upregulated TCSn::GFP expression and activation of the ProWUS:tdTomato reporter (Fig. 1). Cytokinin ribosides followed by cytokinin N-glucosides and ribotides dominantly contributed to the total endogenous cytokinin levels (Fig. S5, Table S2). The partial decrease of endogenous cytokinin levels observed in K300 when compared with hypocotyls grown in K0 might be a result of upregulated endogenous CKX activity, as reported previously (Pernisová et al., 2009; Werner et al., 2006). Interestingly, although the dynamics of tZ-type cytokinins was comparable at all the three media tested (K0, K300 and K1000), the levels of iP-type cytokinins strongly increased, particularly from the day 3 onwards, on shoot-forming media containing a high cytokinin concentration (K1000; Fig. 3A). The correlation of upregulation of iP-type cytokinins with occurrence of disorganised primordia and WUS activation might suggest possible role for iP in the shoot formation. Accordingly, exogenously applied iP was more efficient at de novo shoot regeneration than was tZ and induced shoot formation at a lower concentration (Fig. 3B).
Overall, our results point to a differential regulation of tZ- and iP-type cytokinin homeostasis during de novo shoot formation. Although upregulation of tZ-type cytokinins seems to be connected with organ development in general, the ability of hypocotyl explants to form de novo shoot apical meristems is associated with upregulation of iP-type cytokinins.
Cytokinin receptors differentially control de novo primordia development and organ identity
The specificity of individual cytokinin receptors (AHK2, AHK3, AHK4) in terms of their affinity for individual cytokinins as well as for tissue-specific upregulation of cytokinin signalling has been demonstrated previously (Romanov et al., 2006; Stolz et al., 2011). To uncover the role of AHK-mediated cytokinin perception and individual cytokinin receptors in the early events of de novo organogenesis, we inspected primordia development in double ahk mutants, in which only one cytokinin receptor was functional. As mentioned above, in wild type all the primordia were established on day 1 and no further increase was observed throughout the rest of the cultivation period (Figs 2A and 4A). In contrast, there was an apparent increase in the total number of primordia in ahk3 ahk4 at day 3 and day 5 when compared with day 1; a similar trend, although statistically insignificant, was also observed for ahk2 ahk4 (Fig. 4A, Table S3A). All three double mutant lines possessed an insensitivity to exogenous cytokinin-mediated inhibition of primordia initiation (Fig. 4A, Table S3B). The progress through the primordia developmental path was faster in ahk2 ahk4 and ahk3 ahk4, revealing a higher proportion of primordia at stages III-V on day 1 when compared with wild type (Fig. 4B, blue rectangle; Table S3C), suggesting that AHK4 is a negative regulator of early primordia development. Additionally, disorganised primordia formed in the presence of exogenous cytokinin were detected only in wild type and ahk2 ahk3 lines, where AHK4 is functional (Fig. 4B, green rectangle; Table S3D,E). Accordingly, although shoots were formed in wild type and ahk2 ahk3 mutants after 21 days of explant cultivation, no shoots were recognised in single or multiple mutants carrying the ahk4 allele (Fig. 4C, Figs S6, S7).
To visualise the spatiotemporal distribution of cytokinin perception and signalling, we inspected the expression pattern of individual cytokinin receptors and the primary cytokinin response gene ARR5. Expression of ProAHK2:AHK2-uidA (AHK2-GUS) and ProAHK3:AHK3-uidA (AHK3-GUS) reporters displayed a similar pattern, being recognisable from day 1 of cultivation (Fig. 5). The signal of both receptors appeared at the base of the root primordia and emerged roots, and decreased in disorganised primordia. In comparison with AHK2, the signal of AHK3 was stronger in the provascular tissue of newly developing roots. Primordia on medium with high cytokinin levels (K300, K1000) displayed weaker AHK2 and AHK3 expression when compared with cytokinin-free medium (K0). However, AHK4 and ARR5 reporters (ProWOL::4xYFP and ProARR5:GUS, respectively) were undetectable or very low in root primordia formed in the absence of exogenous cytokinins (K0), with ARR5 and weak AHK4 signal being detectable in provascular tissue of emerging roots at day 4 and 5. However, the signal of both AHK4 and ARR5 was upregulated by exogenous cytokinins (K300, K1000) from day 2. Stronger expression of both reporters concentrated in disorganised primordia – a pattern also corresponding to the spatial specificity of TCSn::GFP and overlapping with shoot apical meristem-specific ProWUS:tdTomato (Fig. 1). With the exception of the columella/lateral root cap on cytokinin-free media (K0), the expression pattern of AHK4 seems to correlate well with the two-component signalling sensor TCSn::GFP in developing roots, as well as in disorganised primordia (Figs 1B and 5). Nonetheless, it should be highlighted here that, in contrast to the AHK2 and AHK3 reporters, AHK4 is not translational but transcriptional fusion. Thus, post-transcriptional regulations that potentially influence the final AHK4 localisation domain cannot be excluded.
These results suggest an important role for the cytokinin receptor AHK4 in the process of in vitro primordia initiation and differentiation, and indicate a central role for AHK4 in cytokinin-induced primordia respecification and de novo shoot formation. Moreover, our findings uncovered differential regulation of individual cytokinin receptors by exogenous kinetin application in the course of the early organogenic response.
AHK4-mediated cytokinin signalling downregulates WOX5 in the root
To investigate a possible role for cytokinin perception and signalling in the root apical meristem of intact seedlings, we crossed single and double ahk mutants with TCSn::GFP, ProARR5:GUS and ProWOX5:GFP reporters. In all combinations carrying the ahk4 mutant allele, the TCSn::GFP and ProARR5:GUS signal in a stele was either very low or undetectable (Fig. 6A). An increase in ProWOX5:GFP expression was detected in ahk3 and particularly in ahk4 single, and ahk2 ahk4 and ahk3 ahk4 double mutant backgrounds. In contrast, ahk2 ahk3 mutants displayed attenuation of the ProWOX5:GFP signal (Fig. 6). In all of the assayed mutant backgrounds, the WOX5 reporter activity appears to be in a negative correlation with the intensity of cytokinin signalling, as assayed by TCSn::GFP and ProARR5:GUS reporters (Fig. 6A). Therefore, AHK4-regulated cytokinin signalling seems to negatively regulate WOX5 expression in the quiescent centre, thus potentially controlling identity and/or activity of the root stem cell niche.
Respecification of organ primordia identity is associated with cytokinin-induced loss of primordia organisation
Several recent studies show that formation of pluripotent callus during the two-step protocol, i.e. during cultivation on auxin-rich CIM, requires activation of key root trait determinants, including members of PLT gene family and/or regulators necessary for the lateral root formation, e.g. ALF4 (Kareem et al., 2015; Sugimoto et al., 2010). During later cultivation on the SIM, cytokinins inhibit the activity of those root determinants and activate the shoot-determining factors, including WUS (Atta et al., 2009; Gordon et al., 2007, 2009; Sugimoto et al., 2010). Our detailed morphological analysis suggests that cytokinins inhibit both de novo organ primordia initiation as well as their later differentiation, as most of the primordia did not succeed in proceeding beyond stage IV on cytokinin-rich medium. In plants, lateral root primordia were reported to be more sensitive to exogenous cytokinin at early developmental stages (Bielach et al., 2012; Laplaze et al., 2007). In addition, endogenous cytokinin overproduction in plants leads to primordia growth inhibition preferentially in stages II-IV (Bielach et al., 2012; Kuderová et al., 2008). Similar observations were obtained during direct conversion of lateral root primordia into shoots (Rosspopoff et al., 2017). In that work, the young root primordia (up to stage V) were shown to be incompetent to the cytokinin-induced identity switch and stopped their growth soon after transplanting to the cytokinin-rich media. Interestingly, the transition of stage IV to stage V coincides with the initiation of WOX5 expression in the root primordia (Goh et al., 2016; Rosspopoff et al., 2017) grown in the cytokinin-free media that is quickly inactivated by exogenous cytokinins in the shoot-inducing media.
In the shooting media from day 3 onwards, we observed the gradual primordia disorganisation that associated with strong upregulation of cytokinin signalling that spatially overlaps the WUS expression domain. It is not clear whether the cytokinin-induced structural disorganisation has any functional meaning in the process of identity switching and/or differentiation status of the tissue or whether it is merely a collateral consequence of cytokinin-induced loss of the auxin patterning, possibly via misregulation of auxin transport (Pernisová et al., 2009). However, the identity switch, exemplified by WUS activation, was observed solely in the disorganising primordia. Thus, in our one-step system, the cytokinin-induced primordia disorganisation seems to be tightly associated with the competence to change their identity.
Nonetheless, although WUS is essential for de novo shoot regeneration (Zhang et al., 2017), activation of WUS alone does not seem to be sufficient for the switch in the organ identity. In the presence of high kinetin concentrations, WUS could also be activated in the non-growing trunk regions of root explants, which are unable to regenerate into shoots (Sugimoto et al., 2010). Similarly, WUS was activated at stage V in the non-competent lateral root primordia that were not developing into the shoots when transplanted to the shoot-inducing media (Rosspopoff et al., 2017). This implies the existence of a developmental- and tissue-specific context necessary for both cytokinin-mediated WUS activation and WUS-dependent induction of a change in organ identity. Recently, the type-B ARRs ARR1, ARR2, ARR10 and ARR12 have been implicated in directly interacting with WUS promoter (Dai et al., 2017; Meng et al., 2017; Wang et al., 2017; Zhang et al., 2017; Zubo et al., 2017). However, similar to our study, showing a much larger TCSn:GFP domain when compared with highly focused WUS activity (Fig. 1C), the expression domains of ARR1, ARR10 and ARR12 were also larger than that of WUS (Meng et al., 2017). Furthermore, WUS activity, although weaker, was detectable and normally positioned in the arr1 arr10 arr12 background, implying the existence of other factors controlling WUS expression (Meng et al., 2017). The regulatory elements that allow cytokinin-induced upregulation of WUS specifically in the disorganising primordia and the tissue-specific determinants that act downstream of WUS in the root-to-shoot identity switch therefore remain to be identified.
Root-to-shoot induction associates with specificity in the endogenous cytokinin production and signalling
The measurement of endogenous cytokinin levels revealed connection between shoot apical meristem establishment and iP-type cytokinin production. A possible explanation could be expression specificity of cytochrome P450 monooxygenase CYP735A2, which mediates the formation of tZ cytokinin types via iP hydroxylation. Expression of CYP735A2 was found to occur dominantly in the roots. Furthermore, activity of CYP735A2 in the roots was strongly inducible by cytokinins, including its substrate iP, whereas a similar response was not observable in the shoot (Takei et al., 2004). Thus, the specific upregulation of iP type cytokinins on the shoot-forming media might reflect low activity of CYP735A2, thus leaving most of the produced cytokinins in their non-hydroxylated form.
Our results showed that exogenously applied iP induced in vitro shoot regeneration at a much lower concentration in comparison with that of tZ (Fig. 3B). Our data also suggest that AHK4 plays a dominant role in mediating the cytokinin-induced organ disorganisation and respecification. Furthermore, AHK4 seems to be negative regulator of WOX5 in the root, whereas both AHK2 and AHK3 might have weaker effect on WOX5 expression (Fig. 6). Accordingly, although AHK4 receptor recognises both tZ and iP with comparable affinity (Romanov et al., 2006; Spíchal et al., 2004), AHK3 recognises iP with ∼100 times lower efficiency than tZ (Romanov et al., 2006). In line with our observations suggesting kinetin as being the least efficient ‘shooting’ cytokinin when compared with iP and tZ, kinetin was found to have significantly lower activity than iP or tZ in an AHK4-mediated β-galactosidase assay (Spíchal et al., 2004).
The factors determining the differential role of AHK2 and AHK3 on the one hand and AHK4 on the other in the regulation of WOX5 in the root quiescent centre remain to be identified. Specificity of the AHK intracellular domains to AHPs, the downstream members of the multistep phosphorelay pathway, as shown in the case of another sensor histidine kinase, CKI1 (Pekárová et al., 2011), together with differential expression and cytokinin responsiveness during de novo shoot formation (Fig. 5) might be one of the possible mechanisms. Moreover, in contrast to AHK2 and AHK3, AHK4 has been shown to have phosphatase activity in the absence of cytokinins (Mähönen et al., 2006), which might also contribute to the specific type of regulation mediated by AHK4 in the control of de novo shoot formation.
A model for the cytokinin-induced change in organ identity in the single-step protocol
Based on the aforementioned findings, we propose that, in the one-step approach, auxin induces formation of root primordia, possibly activating root-specific developmental pathways, as previously demonstrated in the two-step protocols. The process of primordia initiation is under the negative control of cytokinins (both endo- and exogenous). The process of organ primordia formation is associated with endogenous cytokinin production that seems to have an autoregulatory role in inhibiting further primordia formation, as previously demonstrated in the root (Bielach et al., 2012); this negative effect seems to be dominantly mediated by AHK4 (Fig. 7). Moreover, the expression of individual cytokinin receptors seems to be under differential cytokinin control, possibly also contributing to the final signalling output. The root primordia reaching the transition to stage V become susceptible to exogenous cytokinins that, via AHK4, preferentially mediate the loss of the root identity, as evidenced by the downregulation of root reporters. Primordia respecification, which is possibly accompanied by downregulation of CYP735A2 activity, is associated with upregulation of endogenous iP-type cytokinins that further contribute to the upregulation of AHK4-mediated cytokinin signalling and to the loss of root primordia architecture. Activated cytokinin signalling, directly or indirectly and in a cooperation with the yet unknown factors acting in a tissue- and developmental-specific context, mediate induction of WUS specifically in the disorganised primordia, leading to the acquisition of shoot identity and the activation of the downstream developmental cascade, which allows de novo shoot formation.
MATERIALS AND METHODS
Unless otherwise mentioned, all plant material used was from Arabidopsis thaliana, ecotype Col (NASC N60000). The transgenic or mutant lines have been described previously: Pro35S:AtCKX3 (Pernisová et al., 2009), pDR5rev::3XVENUS-N7 (Wabnik et al., 2013), TCSn::GFP (Zurcher et al., 2013), ProARR5:GUS (Che et al., 2002), ProWOX5:GFP (Blilou et al., 2005), ProWOL::4xYFP (Marquès-Bueno et al., 2016), ProAHK3:AHK3-uidA (Dello Ioio et al., 2007), ProSCR:GFP-SCR (Gallagher et al., 2004), ProPLT2:PLT2-YFP and ProPLT3:PLT3-YFP (Galinha et al., 2007), ahk4-1 (Ueguchi et al., 2001), cre1-2 (Inoue et al., 2001), and ahk2-5, ahk3-7, ahk2-5 ahk3-7, ahk2-5 cre1-2, ahk3-7 cre1-2 and ahk2-5 ahk3-7 cre1-2 (Riefler et al., 2006).
pDR5rev::3XVENUS-N7 and TCSn::GFP, ProWOX5:GFP and ProWUS:tdTomato, TCSn::GFP and ProWUS:tdTomato were crossed and double homozygous lines were analysed. Mutant lines ahk2-5, ahk3-7, cre1-2, ahk2-5 ahk3-7, ahk2-5 cre1-2 and ahk3-7 cre1-2 were crossed with reporters TCSn::GFP, ProARR5:GUS and ProWOX5:GFP, and homozygous lines were used in experiments.
The growth media used comprised 0.5 Murashige and Skoog medium (Duchefa) with 1% sucrose (Lach-Ner) and 0.8% Plant agar (Duchefa), pH 5.7 adjusted by KOH. Plants were cultivated in growth chambers (CLF Plant Climatics) under long-day conditions (16 h light/8 h dark) at 21°C in Petri dishes or in soil, with a light intensity of 150 μM m−2 s−1 and 40% relative humidity.
Primers for genotyping of ahk2-5 and ahk3-7 have been published previously (Riefler et al., 2006). New cre1-2 specific primers were designed as follows: CRE1-2 for (5′-CTCTTTTGTTCTTGAATTCGC-3′); CRE1-2 rev (5′-ATCCTGCAACATTCTAGCTC-3′); and cre1-2 in (5′-ATAACGCTGCGGACATCTAC-3′).
Three primers for each gene were optimised in one PCR reaction. For all three genes, the following conditions were used: 94°C for 1 min 30 s; 40× (94°C for 15 s, 56°C for 30 s and 72°C for 40 s); and 72°C for 7 min.
Hypocotyl explants assay
Plants were cultivated for 1 day in the light and 5 days in the dark in Petri dishes with Murashige and Skoog medium including Gamborg B5 vitamins (Duchefa), with 1% sucrose (Lach-Ner) and 0.3% Phytagel (Sigma), pH 5.7 adjusted by KOH. Etiolated hypocotyls were isolated by removing cotyledons and roots, and were placed on Petri dishes. Hypocotyl explant cultivation medium contained Murashige and Skoog medium including Gamborg B5 vitamins (Duchefa), with 1% sucrose, 0.3% Phytagel (Sigma) and 1 mg/l biotin (Duchefa), pH 5.7 adjusted by KOH, and enriched with the phytohormones 1-naphthalene acetic acid (NAA; Sigma), kinetin (Sigma), trans-zeatin (tZ; OlChemIm) or isopentenyladenine (iP; OlChemIm) at the appropriate concentrations. Hypocotyl explants were cultivated under continuous light conditions at 21°C, with a light intensity of 150 μM m−2 s−1 and 40% relative humidity.
DIC microscopy was performed on an Olympus BX61 microscope equipped with ×10, ×20 and ×40 air objectives and a DP70 CCD camera. Confocal microscopy was carried out on two microscopes: an inverted Zeiss Observer.Z1 equipped with a LSM780 confocal unit and ×20 air objective and ×40 water immersion objective; and an upright LEICA DM 2500 with TCS SPE confocal unit and ×40 air objective. Excitation and detection of fluorophores were configured as follows: GFP was excited at 488 nm and detected at 490-530 nm; YFP and VENUS were excited at 514 nm and detected at 520-560 nm; tdTomato was excited at 561 nm and detected at 570-630 nm. For microscopic analyses, 10-12 hypocotyl explants were inspected.
Analysis of primordium developmental stages
For morphological analyses of primordia developmental stages, 10-12 hypocotyl explants were inspected. Developmental stages were determined and evaluated based on lateral root developmental stages (Malamy and Benfey, 1997).
Signal intensity measurements were carried out using the software accompanying the confocal microscopes: ZEN (Carl Zeiss MicroImaging), LAS AF lite (Leica Microsystems CMS) and analySIS^D (Olympus Soft Imaging Solutions). ProWOX5:GFP signal intensities were measured as an average grey scale values, ranging from 0 to 4096.
AHK2-GUS, AHK3-GUS and pARR5:GUS hypocotyl explants or seedlings were stained in 0.1 M sodium phosphate buffer (pH 7.0) containing 0.1% X-GlcA sodium salt (Duchefa), 1 mM K3[Fe(CN)6], 1 mM K4[Fe(CN)6] and 0.05% Triton X-100 for 1 h (ProARR5:GUS) or 2 h (AHK2-GUS, AHK3-GUS) at 37°C and were incubated overnight in 80% (vol/vol) ethanol at room temperature. Tissue clearing was conducted as previously described (Malamy and Benfey, 1997).
Measurements of endogenous cytokinins
The AGI codes (www.arabidopsis.org) of loci used for genotyping and reporter line preparation are as follows: AHK2 (AT5G35750), AHK3 (AT1G27320), AHK4 (AT2G01830), ARR5 (AT3G48100), WOX5 (AT3G11260), WUS (AT2G17950), CKX3 (AT5G56970), SCR (AT3G54220), PLT1 (AT3G20840) and PLT3 (AT5G10510).
Statistical analysis was performed with Statistica, version 13 (Dell) by employing non-parametric Mann–Whitney U-test, Kruskal–Wallis and Fisher's exact tests. For details, see the supplementary Materials and Methods.
ProWUS:tdTomato line preparation
The ProWUS::tdTomato-N7 transgene was constructed by gateway-mediated cloning. For details, see the supplementary Materials and Methods.
The work was greatly supported, both materially and intellectually, by Paul T. Tarr and Elliot M. Meyerowitz, Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, CA, USA. We thank Bruno Müller for the TCSn:GFP, Jiří Friml for the pDR5rev::3XVENUS-N7, Philip N. Benfey for the ProSCR:GFP-SCR, and Ben Scheres for the ProWOX5:GFP, ProPLT2:PLT2-YFP and ProPLT3:PLT3-YFP lines. The Plant Sciences Core Facility of CEITEC Masaryk University is acknowledged for plant and explant cultivation.
Conceptualization: M.P., J.H.; Methodology: M.P., D.H., T. Kakimoto, M.G.H., O.N.; Formal analysis: M.P., M.G., T. Konecny, L.P., D.H., T. Kakimoto, M.G.H., O.N.; Investigation: M.P., J.H.; Resources: M.P., T. Kakimoto, J.H.; Data curation: M.P., M.G., T. Konecny, L.P., D.H., O.N.; Writing - original draft: M.P., J.H.; Writing - review & editing: M.P., J.H.; Visualization: M.P.; Supervision: M.P., O.N., J.H.; Project administration: M.P.; Funding acquisition: M.P., O.N., J.H., T. Kakimoto.
This work was supported by the Grantová Agentura České Republiky [GP14-30004P to M.P. and T. Konecny, and 13-25280S to J.H.], the Ministerstvo Školství, Mládeže a Tělovýchovy [CEITEC 2020 (LQ1601) to M.P., M.M. and J.H.; LH14104 to M.P. and J.H.; National Program for Sustainability I, LO1204 to L.P. and O.N.; LM2015062 Czech-BioImaging], the EU Seventh Framework Programme [286154 – SYLICA to M.P.] and a Japan Society for the Promotion of Science grant [25113006 to T. Kakimoto].
The authors declare no competing or financial interests.