DISCUSSION

The CK concentration is increased in response to photoperiod stress

Here we reported on the functional relevance of root-derived CK in the response to photoperiod stress. Wild-type plants grown under short-day conditions and experiencing a PLP responded by increasing the CK concentration in their leaves (Fig.1, Table S1). As plants with a reduced CK concentration or signaling are particularly sensitive to photoperiod stress (Nitschke et al., 2016), this response may be part of a defense mechanism enabling wild-type plants to react appropriately to photoperiod stress and to cope with its consequences. Altered CK concentrations are often part of the response to abiotic stress and they may either increase or decrease. A decreased CK concentration was found after exposure to several abiotic stresses like heat, salt or drought stress (Bano, Hansen, Dörffling, & Hahn, 1994; Caers, Rüdelsheim, van Onckelen, & Horemans, 1985; Itai, Ben‐Zioni, & Ordin, 1973; Nishiyama et al., 2011). Plants with a lower CK status were more stress resistant indicating a functional relevance of the reduced CK level ( Nishiyama et al., 2011). In contrast, under high light stress CK has a protective function. It represses excessive starch grain and plastoglobuli formation and is required for a functional D1 repair cycle (Cortleven et al., 2014). In the response to biotic stress such as Pseudomonasinfection, CK is required for an effective defense regulating the oxidative burst through ARR2 (Arnaud et al., 2017; Choi et al., 2010). CK is known also from other instances to regulate the response to oxidative stress (Pavlů et al., 2018; Cortleven et al., 2019) which is a hallmark of the response to photoperiod stress (Nitschke et al., 2016). We propose that one function of the enhanced CK formation could be to properly respond to oxidative stress caused by PLP treatment (Nitschke et al., 2016).

Root-derived tZ-type CKs act as protectants against photoperiod stress

Root-derived t Z-type CKs were shown to be the most relevant to protect plants from the negative consequences of photoperiod stress. This result supports the role of CK in root-to-shoot communication and extends its function to a specific response to stress beyond its better known roles in regulating development and growth. The results show also that tZ-type CKs are particularly relevant for stress protection. The major transport form, t ZR, as well as to a minor extent its bioactive derivative t Z, are transported from the root to the shoot via the xylem flow requiring the transporter ABCG14 (Ko et al., 2014; Zhang et al., 2014). abcg14 mutants are thus deficient intZ in the shoot (Ko et al., 2014; Zhang et al., 2014) and consistently these mutants showed a very strong response to photoperiod stress. In cypDM mutants, the lower levels of t Z-type CKs in the shoot are compensated by an increased level of iP-type CK (Kiba et al., 2013). The inability of these higher levels of iP-type CKs to compensate the sensitive photoperiod stress response of cypDMmutants corroborates the functional relevance of t Z-type CKs. Consistent with a major role of t Z-type CKs is also the functional relevance of AHK3 in photoperiod stress signaling (Nitschke et al., 2016). AHK3 displays an about tenfold higher sensitivity tot Z than to iP while AHK2 and AHK4/CRE1 have similar affinities to both iP and t Z (Lomin et al., 2015; Romanov, Lomin, & Schmülling, 2006; Stolz et al., 2011). It has been proposed that the affinity profile of AHK3 is particularly set to respond to root-derived CK (Romanov et al., 2006).
Further support for a role of root-derived CK in photoperiod stress protection came from supplementation experiments. Watering ofcypDM plants with either t ZR or t Z demonstrated that both metabolites can protect plants against photoperiod stress although t Z was more effective (Fig. 3). Both t Z andt ZR supplementation rescued the decrease in type-A ARRtranscript abundance in these plants demonstrating that after application through roots they reached the shoot in a biologically effective concentration. Different roles for root-derived t Z andt ZR have been reported by Osugi et al. (2017). It might be that the ability of certain tissues to convert inactive t ZR to activet Z, as discussed in Romanov et al. (2018), might have an impact on the plant´s response to photoperiod stress.
The functional relevance of root-derived CK in the response to photoperiod stress raises the question how information about a stress perceived and acting primarily in the shoot is relayed to the root. One possibility is that the light signal is perceived and interpreted in the root directly (Sun, Yoda, & Suzuki, 2005; Sun, Yoda, Suzuki, & Suzuki, 2003). Another possibility is that an instructive chemical signal is formed in the shoot and transported to the root to induce synthesis oft Z CK. This signal could be iP-type CK as these are mainly formed in the shoot and known to be transported to the root through the phloem (Hirose et al., 2008; Kudo, Kiba, & Sakakibara, 2010). iP-type CKs could then positively regulate t Z-type CK formation as they not only serve as a precursor for t Z-formation but also induce the expression of CYP735A2 (Takei et al., 2004). Another candidate for a chemical signal is jasmonic acid which is increased in response to photoperiod stress in sensitive genotypes (Nitschke et al., 2016) and which has recently been shown to be a shoot-to-root signal (Schulze et al., 2019).

ARR2, ARR10 and ARR12 regulate the response to photoperiod stress in a complex manner

AHP2, AHP3 and AHP5 act redundantly in the response to photoperiod stress (Fig.4). The functional redundancy of these AHPs has been shown before in the context of seed, primary root and hypocotyl development (Hutchison et al., 2006). Our results integrate AHPs into the CK-dependent photoperiod stress signaling pathway that so far involved AHK3 and ARR2 as the main signaling components (Nitschke et al., 2016).
Downstream of the AHPs act several transcription factors to realize the transcriptional output of the photoperiod stress response. ARR2 has a predominant role in mediating CK activity in leaves (Hwang & Sheen, 2001) but its redundant function with ARR10 and ARR12 has not yet been described. The latter two ARRs are better known for their role in regulating most CK-related vegetative developmental processes together with ARR1 (Argyros et al., 2008; Ishida, Yamashino, Yokoyama, & Mizuno, 2008). Analysis of single and double mutants showed that loss of eitherARR10 or ARR12 rescued the stress phenotype of arr2plants while the loss of both factors enhanced the stress response ofarr2 (Fig. 5). This hints to a complex regulatory mechanism between these three transcription factors during photoperiod stress signaling. A complex relationship among these type-B ARRs has also been described for their role in regulating root elongation. arr12 andarr10,12 root elongation was less affected by CK treatment than that of arr2,12 and arr2,10,12 (Mason et al., 2005). For type-B ARR dependent gene regulation, a model has been proposed in which simultaneous binding of multiple/different type-B ARRs and unknown factors to certain promoter regions is crucial (Ramireddy, Brenner, Pfeifer, Heyl, & Schmülling, 2013). However, experimental evidence for a direct interaction between members of the type-B ARR family is rare. An interaction of ARR2 and ARR14 has been described using a two-hybrid system in yeast (Dortay, Mehnert, Bürkle, Schmülling, & Heyl, 2006). Recently, it was found that the C-termini of ARR1 and ARR12 interact to regulate auxin synthesis (Yan et al., 2017). It could also be that interactions between type-B ARRs are context-dependent as it is known for the phosphorylation-dependent homodimerization of bacterial RRs (Mack, Gao, & Stock, 2009). Similarly, ARR18 can homodimerize when both ARR18 proteins are either both phosphorylated or both not phosphorylated (Veerabagu et al., 2012).
The different phenotypic and in part molecular responses to photoperiod stress of arr mutants could be explained by a model, in which ARR2, ARR10 and ARR12 interact with a yet unknown interaction partner (X) that is essential for photoperiod stress resistance (Fig. 6). It is predicted that the affinity of ARR2 to X would be higher than the affinities of ARR10 and ARR12 to X. In addition, we propose a direct or indirect interaction of ARR10 and ARR12. In photoperiod stress-treated wild-type plants, ARR2 would interact with X resulting in photoperiod stress resistance while ARR10 and ARR12 together would have independent auxiliary functions. In arr2 plants, X would not have an interaction partner and thus would be unable to function in stress protection because ARR10 and ARR12 would not be available as interaction partners. Consequently, stress resistance would be lowered. Resistance of arr2,10 and arr2,12 plants would be caused by the loss of the ARR10-ARR12 association and the resulting interaction of X with ARR10 or ARR12. Ultimately, the enhanced stress phenotype ofarr2,10,12 plants would be caused by the complete loss of interaction partners for X.
Beside the interaction amongst ARRs, interactions between several type-B ARRs and other proteins exist. For example, ARR1, ARR2 and ARR14 interact with the DELLA proteins RGA1 and GAI to regulate root development and photomorphogenesis (Marín-de la Rosa et al., 2015; Yan et al., 2017). During the regulation of auxin synthesis, EIN3 interacts with the C-terminus of ARR1 and thereby increases ARR1 activity (Yan et al., 2017). As part of the crosstalk between CK and abscisic acid, ARR1, ARR11 and ARR12 directly interact with SUCROSE NON‐FERMENTING‐1 (SNF1)‐RELATED PROTEIN KINASE2 (SnRK2) kinases and thereby inhibit their function prior to drought stress (Huang et al., 2018). Future experiments might resolve whether and how type-B ARRs interact with each other or with other proteins during photoperiod stress.