Versatile effect of cytokinin on detached senescing leaves of Arabidopsis in the light

The dark-detached leaves are widely used as a model to investigate senescence-delaying mechanisms of cytokinin (CК). The role of CK during senescence of detached Arabidopsis leaves in the light is less examined, although they provide unique opportunities for the study of the interaction of CK and the photosynthetic apparatus. Within the first four days of incubation of the leaves on water photosynthesis activity, chloroplast gene expression and chlorophyll content declined in all genotypes with a functioning CK receptor AHK3, while anthocyanin level and expression of its regulatory and biosynthesis genes increased. Application of exogenous CK visibly retarded the senescence symptoms but decreased photosystem II function, anthocyanin accumulation and promoted ROS production. The hormone stimulated starch accumulation in the chloroplasts and altered the expression of the gene for the extracellular cell wall invertase CWINV1. Although the AHK3 receptor played the most important role in the CK action, we also observed specific effects in loss-of function mutants for phosphortransmitters and type B response regulators. The results suggest that CK action in detached Arabidopsis leaves in the light is superimposed on modified carbon fluxes and starch turnover which results in an intricate combination of stress-related and senescence inhibiting effects.


Introduction
Leaf senescence is a genetically programmed developmental process essential for mobilization and recycling the nutrients either to the developing seeds or to other parts of the plants to ensure survival in the following season (Buchanan-Wollaston 1997;Buchanan-Wollaston et al. 2003). Senescence can also be induced prematurely by environmental stresses that shorten the lifespan of individual leaves or the whole plant. The onset of leaf senescence triggers degradation of chloroplasts as well as thylakoid membranes and decreases the level of photosynthetic pigments. (Dhami and Cazzonelli 2020). The changes in global gene expression patterns during leaf senescence are accompanied by differential accumulation of proteins and constant down-regulation of chloroplast proteins involved the biosynthesis of chlorophyll (Xu et al. 2015). This process, however, can be retarded or even reversed by various Communicated by Carlos Garcia-Mata. treatments among which light and cytokinin (CK) are considered key players (Janečková et al. 2018) The well-known inhibition of leaf senescence by CK is based on the canonical two-component signaling (TCS) pathway involving three histidine kinase receptors (AHK4/ CRE1/WOL, AHK2, and AHK3), histidine phosphotranfer proteins (AHPs) and two classes of the response regulators acting as transcription factors (type B ARRs) or negative regulators of CK signaling (type A ARRs). The senescence-specific responses are mediated by the CK receptor AHK3 and the type-B response regulator ARR2 (Kim et al. 2006), in addition to the AP2 /ERF type transcription factors CRFs which act as a side branch of TCS. CRF6 functions as a negative regulator of senescence in detached leaves in the dark (Zwack and Rashotte 2013). However, disruption of multiple CRFs resulted in delayed senescence in intact Arabidopsis leaves indicating their positive regulatory role in natural leaf senescence (Raines et al. 2016).
The downstream effects of CK during leaf senescence are thought at least in part to be linked to source-sink relationships and especially to sucrose metabolism and transport. They, in turn, depend on the activity of sucrose-phosphate synthase, sucrose synthase and three types of invertases: a C-invertase in the cytosol, a V-invertase localized in vacuoles and a cell wall-invertase (CWINV; Wang et al. 2016). The activity of the extracellular CWINV is of particular interest for leaf senescence. It was shown that increased CWINV activity in response to CK treatment is both necessary and sufficient for the inhibition of senescence (Zwack and Rashotte 2013).
In addition, the mechanisms of CK effects on leaf senescence are mediated by a crosstalk between light and CK as indicated by CK effects on the abundance of proteins associated with photosynthesis (Kusnetsov et al. 1994). This may also be related to photooxidative stress tolerance since photosynthetic disfunction in aging leaves leads to the production of reactive oxygen species (ROS; Zwack and Rashotte 2015). Although CK is generally considered to play a negative role in plant adaptation to stresses, it also has a protective function in specific contexts. For instance, CK signaling was shown to play a positive role in the response to high-light stress (Cortleven et al. 2014), while under mild photooxidative conditions CK receptors negatively regulated stress tolerance (Danilova et al. 2014).
The modes by which CK enhances leaf longevity during senescence may differ for different species and require an understanding of the contribution of CK signaling for the primary metabolism during aging. Here we used reverse genetic approaches to examine the roles of CK and CK signaling components during senescence of excised Arabidopsis leaves in the light. Our study demonstrates that in this model system CK action is superimposed on modified carbon fluxes and starch turnover which results in an intricate combination of stress-related and senescence inhibiting effects.

Plant materials, growth conditions and experimental design
The Columbia-0 ecotype of Arabidopsis thaliana plants was used as the wild type (WT). The T-DNA insertion mutants ahk2-2, ahk3-3, cre1-12 (hereafter referred to as ahk2, ahk3 and ahk4) and double CK receptor mutants ahk2 ahk3, ahk2 ahk4 and ahk3 ahk4 were gifts of Prof. Higuchi (Higuchi et al. 2004). The T-DNA insertion mutants ahp2/3/5 (CS860161), arr2 (CS6975), and arr1/10/12 (CS39992), were obtained from the NASC collection. Seeds were put on water-soaked filter papers and were kept at 4 °C for 3 d in the dark before germination. All plants were grown in a greenhouse on perlite/soil mixture under a 16 h/8 h light/ dark photoperiod (lamp OSRAM L58W/640) at a temperature of 21 °C. The light intensity was 90 µmol photons m −2 s −1 . After 4 weeks, the mature 5th and 6th rosette leaves were excised from the plants and placed in glass Petri dishes with distilled water (mock) or an aqueous solution of 5 µM trans-zeatin (tZ). The detached leaves were incubated under the same light conditions (16 h/8 h light/dark) for four days. These samples were compared with leaves collected immediately after leaf detachment which were used as a control.

Measurements of chlorophyll and anthocyanin contents
The chlorophyll content of leaves was measured as described by Lichtenthaler (1987). For anthocyanin extraction, samples (20 mg) were crushed into fine powder in liquid nitrogen and homogenized in 1 ml of acidified (1% [v/v] HCl) ethanol. After that, one volume of chloroform was added to remove the chlorophylls and the samples were centrifuged for 15 min at 12,000 rpm. The anthocyanin containing supernatant was recovered, and the absorbance of the supernatant was measured with a Genesys 10 UV spectrophotometer (Thermo Electron Corporation, USA) at 535 nm.

Determination of malondialdehyde (MDA) content
To evaluate oxidative stress, we determined the level of secondary products of lipid peroxidation of membranes reacting with thiobarbituric acid (thiobarbituric acid-reactive substances, TBARS). The content of TBARS was measured according to the method proposed by Heath and Packer (1968) using a Genesys 10 UV spectrophotometer (Thermo Electron Corporation, USA) at 532 and 600 nm.

Determination of ROS in leaves
Accumulation of ROS was accessed by precipitation in leaves as described by Ma et al. (2013). In brief, excised leaves were vacuum infiltrated for 10 min with a solution of diaminobenzidine (DAB, 1 mg ml −1 ) for H 2 O 2 or nitroblue tetrazolium (NBT 0.5 mg ml −1 ) for superoxide radicals in 50 mM phosphate buffer, pH 7.0, and left overnight in the absence of light. Chlorophyll was then bleached by boiling in a solution of ethanol/glycerol/acetic acid (4/4/1) until the appearance of brown spots (reaction product of DAB with H 2 O 2 ) or blue formazan precipitates produced upon the reduction of NBT by superoxide.

Analysis of transcript levels by quantitative PCR (qPCR)
Total RNA was extracted from the leaves by the TRIzol (Thermo Fisher Scientific) method. Equal amounts of RNA (2 µg) were used in the 20 µl reverse transcriptase reaction (Thermo Fisher Scientific). cDNA synthesis of the first strand was carried out using a mixture of oligo (dT) primers and random hexamers. Quantitative real-time PCRs were performed in a LightCycler 96 (Roche, Switzerland) with the hot start SYBR Green I technology and gene-specific primers (Table S1). The qPCR program consisted of the following steps: 95 °C for 10 min; 40 cycles of 95 °C for 15 s, 58 °C for 15 s, and 72 °C for 20 s; followed by melting curve analysis. All data were normalized relative to the level of transcripts of the UBQ10 nuclear gene, which was used as an internal control. The relative transcript abundance of each gene was calculated based on the 2 −ΔΔCt method.

Transmission electron microscopy
The detached leaves of A. thaliana were fixed overnight at 4 °C in 2.5% glutaraldehyde dissolved in 0.1 M phosphate buffer (pH 7.2) at 4 °C. Then, they were washed three times with 0.1 M phosphate buffer containing sucrose (6.8%), and post-fixed by incubation in 1% (w/v) osmium tetroxide solution in the same buffer for 4 h at 4 °C. After dehydration in a subsequent series of ethanol, acetone and propylene oxide, the samples were successively immersed in an Epon resin (Fluka, Buchs, Switzerland) which contained propylene oxide added in the proportions (v/v) 1:2, 1:1 and 2:1, with 12 h incubation at each stage. The samples were then embedded in a pure Epon resin. Ultrathin sections were prepared using a glass knife with ultramicrotome LKB 8800 (Lab Instruments, Stockholm, Sweden), mounted on copper grids and stained with 2% aqueous solution of uranyl acetate (w/v) for 20 min and Reynolds lead citrate for 7 min. The sections were examined using a transmission electron microscope (JEM-1200 EX; Jeol, Tokyo, Japan) operating at an accelerating voltage of 80 kV.

Statistical analysis
Standard errors were calculated from the data of at least three separate experiments performed under the same experimental conditions. The reliability of differences was checked by the Student's criterion (t-test) with a significance level of 0.05.

CK exerts opposite effects on leaf senescence and photosynthesis
To test the effect of CK in our system with 4 week-old detached Arabidopsis leaves, we first examined the transcript levels of the type A ARR gene ARR5 in WT and CK mutants. Consistent with previous studies, ARR5 is rapidly induced in response to exogenously applied CK in detached leaves of WT and CK mutant plants with the exception of the ahk3, ahk2/3, ahk3/4, ahp2/3/5, and arr1/10/12 mutants in which CK signaling is impaired (Fig. S1a). Next, we examined the expression level of the senescence marker SEN1 encoding rhodanese/cell cycle control phosphatase superfamily protein. SEN1 transcripts were undetectable in intact leaves, but they were revealed in the water-treated detached leaves, with the highest levels in the ahk3 and ahk3/4 mutants (Fig. S1b). Since senescence-specific CK responses are mediated by the AHK3 receptor (Kim et al. 2006), it appears that the endogenous CK participates in the restriction of SEN1 expression. In CK-treated detached leaves, the SEN1 transcript abundance was strongly decreased as compared to water controls and the inhibition was less pronounced in the ahp2/3/5, arr2 and arr1/10/12 mutants (Fig. S1b). These results suggest that CK retards senescence in detached Arabidopsis leaves via the tested CK signaling compounds.
This was also supported by the TBARS content, an indicator of free-radical-induced injury to biological membranes caused by stresses or aging (Garsia et al. 2014). The TBARS levels were significantly lower in the CK-treated leaves of plants with a functional CK signaling pathway ( Table 1). The TBARS content was much less reduced in the CKexposed arr1/10/12 mutant and in plants which lack the AHR3 receptor. It should be noted, however, that according to some researchers (Pilz et al. 2000;Janečková et al. 2018), spectrophotometric estimation of the MDA-TBA2 adduct (MDA adducted to two molecules of thiobarbituric acid), corresponded more to general oxidative damage of proteins and other compounds than to lipid peroxidation.
To further investigate the effect of CK, we measured the chlorophyll content, photosynthesis parameters and the expression of photosynthesis-related genes in the lightexposed detached leaves from WT and mutant plants. Incubation of the leaves on water induced a rapid degradation of chlorophyll; the strongest pigment loss was detectable in the mutant with the disrupted ARR2 gene (Fig. 1a). Surprisingly, the chlorophyll loss was the same for the water-and CK-treated samples (Fig. 1a). Moreover, the photosynthetic parameter Fv/Fm which represents the efficiency of the photosynthetic electron flow through the photosystem II, declined much faster in the CK-incubated leaves compared to the leaves kept on water (Fig. 1b). The faster decline in the Fv/Fm values in the CK-treated samples was less in the CKinsensitive ahk3, ahk2/3 and ahk3/4 mutants. This indicates that the exogenously applied hormone promotes the decline in the photosynthetic activity which occurs already in the water-exposed detached leaves (Fig. 1b). In contrast, the mRNA levels for two chloroplast-encoded genes, psbD for the D2 protein of the photosystem II and rbcL for the large subunit of the ribulose-1.5-bisphosphate carboxylase, were much higher in the CK-treated samples compared to those exposed to water (Fig. 2a, b). Also the nuclear-encoded LHCB2.4 and RBCS transcript levels for the plastid-localized light-harvesting antenna protein of the photosystem II and the small subunit of ribulose-1.5-bisphosphate carboxylase, respectively, were higher in the CK-treated samples (Fig. 2c, d). Apparently, CK counteracts senescence-induced reduction in photosynthetic gene expression, but promotes the decline in the photosynthetic activity in our senescence model system.

Exogenous CK promotes stress-related symptoms
The reduction of photosynthetic activity after leaf detachment was accompanied by the development pink-violet phenotypes in the water-treated detached leaves (Fig. 3). The coloration is, at least partially, caused by an increase in the anthocyanin levels (Fig. 4). Exposure of leaves to 5 µM tZ significantly blocked the accumulation of anthocyanins in all samples except in the lines in which the AHK3 receptor is inactivated, and to a lesser extent in the arr1/10/12 mutant. In order to gain a better understanding of the molecular basis of anthocyanin production we analyzed the expression patterns of the key biosynthesis and regulatory anthocyanin genes for phenylalanine ammonia-lyase (PAL), chalcone synthase (CHS), anthocyanidin synthase (ANS) and for production of anthocyanin pigment1 (PAP1). Compared to intact leaves, all genes were strongly up-regulated in the detached leaves kept on water, and much less in the CKtreated samples which is in good accordance with pigment levels (Fig. 5). Interestingly, the CHS and ANS transcript levels were much higher in the leaves of the ahk2/3 mutant which corresponds to anthocyanin data (Fig. 5b, c).
Although the antioxidative capacity of anthocyanins has been demonstrated in many species (Zeng et al. 2010), these results can be interpreted in two different ways: the reduced anthocyanin levels in the presence of CK contributes to stress that occurs in the detached leaves, or the CK-treated leaves are less stressed due to the retardation of senescence.
In order to get more insight into this scenario, we monitored ROS accumulation. DAB was used to detect hydrogen peroxide (H 2 O 2 ) and NBT for superoxide radicals (O 2− ) levels in the WT and mutant leaves with and without CK treatment. The leaves of Col-0, ahk2/3, ahk2/4, ahk3/4 and arr1/10/12 accumulated pale brown deposits in their central veins after DAB application, but there was no significant difference between CK-and water-treated leaves (data not shown). By contrast, superoxide radical detected by NBT formed more intense blue formazan deposits in CK-treated leaves of Col-0, ahk2/4 and arr1/10/12 compared to leaves which were treated with water (Fig. S2). The difference was not detectable in the leaves of the ahk2/3 and ahk3/4 mutants (Fig. S2). These data indicate that CK-treated leaves, which contain less anthocyanin, are more sensitive to oxidative stress. Higher accumulation of O 2− levels during senescence under moderate illumination in the presence of CK is in accordance with dysfunctions in the photosynthetic electron transport and consistent with the chlorophyll fluorescence data. Better stress performance of mutants impaired in AHK3 function confirms the central role of CK perception in this senescence model. Miyazawa et al. (1999) have shown that CK accelerated starch accumulation in a tobacco cell culture. A higher starch level and fewer plastoglobuli were observed in wheat leaf segments exposed to CK (meta-topolin) in the light (Vlčkova et al. 2006). CK also accelerated starch accumulation and amyloplast development in cultured Bright Yellow-2 tobacco cells (Miyazawa et al. 1999). To test whether CK influences the sink/source regulation during senescence, we investigated the influence of the hormone on the chloroplast ultrastructure 4 days after the excision of the leaves. We focused on the ahk3, ahk2/3, and ahk2/4 plants since they displayed contrasting CK responses for all previously studied parameters.

CK stimulates starch formation in the chloroplasts and CWINV mRNA accumulation in detached leaves
In our senescence assay with detached leaves kept in the light, the shape of the chloroplasts changed from ellipsoidal to distinctly more spherical. Closer inspection of the organelles uncovered that the chloroplasts in the leaves of in the presence of exogenous CK (5 μM, tZ) and without hormone (mock). Transcript levels of the intact WT leaves were set to 1. UBQ10 was used as an internal control. Error bars indicate means ± SE of data from three replicates. Symbols indicate significant differences (P < 0.05; t test) between mock and tZ treated plants within each genotype (x) and relative to the respective WT (o) water-treated WT (Fig. 6a) and ahk2/4 mutant (Fig. 6g) looked slightly swollen and contained larger starch grains, compared to those found in the chloroplasts of the ahk/3 (Fig. 6c) and ahk2/3 (Fig. 6e) mutants in which the organelle had a lens-shaped form with smaller starch grains. We observed no alterations in ultrastructure of thylakoids or in grana stacking.
CK significantly reduced the size and number of the plastoglobuli (Fig. 6b, h). Plastoglobuli are lipoprotein bodies that are engaged in metabolic pathways involving lipids, carotenoids, and tocopherols and in protecting thylakoid membranes from oxidative stress. Their size and/ or number were shown to increase under stress conditions (Bréhélin and Kessler 2008). Thus, a decrease in the size and number of the plastoglobuli upon CK treatment in our experiments points towards a stabilizing effect of CK on the cell membranes. Furthermore, CK-exposed WT chloroplasts (Fig. 6b), as well as those of the ahk2/4 mutant (Fig. 6h), exhibited a round-shaped form, with even larger and a higher number of starch grains compared to the watertreated samples. The amount and size of starch grains in the chloroplasts of the ahk3 mutant (Fig. 6d) incubated in a CK solution were reduced compared to those in WT or ahk2/4. These results indicate that starch accumulation during lightdependent senescence was significantly promoted by tZ in detached leaves with functional AHK3 receptors. However, the CK-treated ahk2/3 mutant showed enlarged starch grains. This probably suggests the involvement of the AHK4 receptor able to provide cytokinin responsiveness or some other compensatory mechanisms. The activation of compensatory mechanisms during deetiolation of the triple mutant arr1ar-r10arr12 resulting in increased chlorophyll synthesis was reported by Cortleven et al. (2014). The authors explained this effect by the expression of transcription factors belonging to alternative systems of regulation of chloroplast formation. Recently, it was shown that noncanonical and nontranscription-mediated signaling by the phytohormone protein network is more common than hitherto appreciated, and that a large majority of signaling proteins function pleiotropically in several pathways contributing to phytohormone signal integration (Altman et al. 2020).
Increased CWINV activity is a characteristic downstream response of CK required to inhibit senescence (Zwack and Rashotte 2013;Wang et al. 2016). The CWINV1 transcript level was strongly down-regulated in water-treated detached leaves of WT and CK mutants. In the presence of CK, the CWINV1 mRNA level was not down-regulated in all genotypes, except in the ahk3, ahk2/3, ahk3/4, and arr1/10/12 mutants (Fig. 7).

Discussion
The dark-detached leaves are widely used as a model to investigate senescence-delaying mechanisms of CK. Dark treatment of detached Arabidopsis leaves revealed a major contribution of AHK3 in mediating cytokinin-dependent chlorophyll retention and the ability to retard leaf senescence (Riefler et al. 2006). However, AHK2 or CRE1/AHK4 together also ensured full cytokinin responsiveness at higher cytokinin concentrations (Riefler et al. 2006). The peculiarities of the senescence of detached leaves in the light are less examined although they provide unique opportunities for the study of the interaction of CK and the photosynthetic apparatus in the late stages of leaf ontogenesis.  b, d, f, h). G, granum; P, plastoglobule; S, starch grain. Scale bars shown at the bottom of the photographs are equivalent to 1 µm Excision of Arabidopsis leaves followed by four days of incubation in water under 16/8 light regime caused obvious senescence-related symptoms which are manifested in a decrease in the total chlorophyll content (Fig. 1a), a reduction of the Fv/Fm values (Fig. 1b) and an acceleration of the plastid psbD and rbcL mRNA decay (Fig. 2a, b). Exogenously applied CK (5 µM) diminished the expression of the senescence indicator SEN1 and TBARS content, as well as the losses in transcript abundance of chloroplast genes of both nuclear and plastid coding. However, CK-treated leaves accumulated lesser amounts of photoprotective anthocyanins and higher levels of superoxide radicals, and these effects require functional AHK3 and AHK2 receptors. Harmful effects of CK during senescence of detached wheat leaves in continuous light were also previously demonstrated by Vlčkova et al. (2006). The authors proposed that protective function of CK (meta-topolin) in darkness becomes damaging in light due to overexcitation of photosynthetic apparatus resulting in oxidative stress. This was also confirmed for detached Arabidopsis leaves and interpreted by an interplay between CKs and light during senescence (Janečková et al. 2018). The absence of a functional AHK3 receptor prevented the CK-mediated feedback inhibition of photosynthesis in the light which results a higher and thus damaging PSII activity, and a faster decline of the Fv/Fm values.
We observed a direct correlation between the activity of photosystem II and the accumulation of anthocyanins in the leaves of WT plants and CK-defective mutants. Accumulation of anthocyanins is considered to be one of the symptoms of leaf aging and stress (Gepstein et al. 2003), in particular, under excessive radiation. In our experiments, anthocyanin accumulation depended on the AHK3 and AHK2 receptors, and this is best visible in vascular tissues (Fig. 3), where AHK3, ARR2, and CRF6 are highly expressed (Higuchi et al. 2004;Zwack and Rashotte 2013). This effect was not visible in the ahp2/3/5, arr2, and arr1/10/12 mutants which indicates a high degree of functional redundancy among the downstream components of the CK signaling pathway and/ or the involvement of other phosphortransmitters or response regulators.
It should be noted that our data are not consistent with the generally accepted view that CK promotes anthocyanin accumulation and the expression of its biosynthesis genes (Deikman and Hammer 1995;Hönig et al. 2018). On the contrary, CK-treated leaves contained less anthocyanins than the water controls ( Figs. 3 and 4). A possible explanation could be that the senescing leaves utilize alternative routes for relocation of their photoassimilates. There is increasing evidence that in addition to photoprotection foliar anthocyanins are synthesized and accumulate to prevent temporary sugar excess in source organs (Lo Piccolo et al. 2018). Within the senescing leaves, anthocyanins can be a carbon sink for absorbing excess photosynthetic carbon. When a detached leaf is incubated on water in the light, the photosynthetic products generated in the mesophyll cells (source) accumulate in the middle vein (acceptor). Excess sugar is temporarily metabolized into anthocyanin, and the efflux of the remaining assimilates to the phloem allows the maintenance of photosynthesis. CK converts the mesophyll cells to a carbon acceptor due to CK attracting effect (Mothes et al. 1961), and the middle vein becomes a donor and functions as source.
As a result of the sink to source transitions, the sucrose is not exported, but polymerized to starch in the chloroplasts of the mesophyll cells. Under this condition, anthocyanin synthesis is inhibited due to reduced sucrose supply in the presence of CK, and photosynthesis is suppressed by an excess of accumulating photoassimilates. Thus, the AHK3dependent mitigation in the accumulation of anthocyanin could be the result of a change in the balance between storage and transport forms of photoassimilates in light-treated detached leaves.
The ability of CK to alter the transport fluxes of photoassimilates in hormone-treated leaves and to create new source-sink relationships was first described by Mothes (Mothes et al. 1961;Mothes and Engelbrecht 1963). This was further substantiated by later findings which endowed CK with the ability to up-regulate the expression of Transcript levels of the intact WT leaves were set to 1. UBQ10 was used as an internal control. Error bars indicate means ± SE of data from three replicates. Symbols indicate significant differences (P < 0.05; t test) between mock and tZ treated plants within each genotype (x) and relative to the respective WT (o) extracellular CWINV and hexose transporters. CWINV was shown to play a crucial role in sink/source regulations and is believed to be an essential component of the mechanisms underlying the delay of senescence (Balibrea Lara et al. 2004;Zwack and Rashotte 2013;Wang et al. 2016).
In our experiments, the induction of the CWINV gene by CK is consistent with this concept and the inhibition of senescence in samples with functional AHK3 receptor and ARR B response regulators. The expected increase in CWINV activity may disrupt apoplastic phloem loading through the hydrolysis of effluxed sucrose and the accumulating hexose monomers could then be transported back into parenchyma cells as proposed by Zwack and Rashotte (2013). This resulted in a poorly productive leaf with an artificially strong sink identity and the feedback inhibition of photosynthesis despite elevated levels in the expression of chloroplast genes. The fact that the activation of sink-specific enzymes including the CWINV is coupled to feed-back inhibition of photosynthetic gene expression was previously demonstrated by Ehness et al. (1997). As a consequence, reduced chloroplast photosynthetic activity contributes to less sugar production which limits the metabolic rates and retards senescence (Breeze et al. 2008).

Conclusion
We propose that the delayed senescence in CK-treated detached leaves under light exposure can be the result of a sophisticated interplay between primary photosynthetic metabolism and rearrangement of carbon fluxes induced by blocking outflow of assimilates. Source to sink transition in the hormone treated leaves promoted the conversion of excess sugar to starch in the chloroplasts of mesophyll cells and caused a decline of the photosynthetic activity despite inhibition of senescence. Thus, the lack of sucrose outflow to the central vein inhibited anthocyanin synthesis and contributed to photooxidative stress. AHK3 is crucial for this CK effects, while the investigated phosphortransmitters and type B response regulators appear to control more specific responses in this scenario. It remains to be shown whether these results have practical values for the delay of post-harvest senescence of green leafy vegetables.

Supplementary Information
The online version contains supplementary material available at https:// doi. org/ 10. 1007/ s10725-022-00909-7. Funding This work was supported by the grant from the Russian Science Foundation  and partially by the state assignment of Ministry of Science and Higher Education of the Russian Federation (theme No. 121040800153-1). The Transmission Electron Microscopy was carried out on the equipment of the CSF -SAC FRC KSC RAS.