Sugar-induced de novo cytokinin biosynthesis contributes to Arabidopsis growth under elevated CO2

Carbon availability is a major regulatory factor in plant growth and development. Cytokinins, plant hormones that play important roles in various aspects of growth and development, have been implicated in the carbon-dependent regulation of plant growth; however, the details of their involvement remain to be elucidated. Here, we report that sugar-induced cytokinin biosynthesis plays a role in growth enhancement under elevated CO2 in Arabidopsis thaliana. Growing Arabidopsis seedlings under elevated CO2 resulted in an accumulation of cytokinin precursors that preceded growth enhancement. In roots, elevated CO2 induced two genes involved in de novo cytokinin biosynthesis: an adenosine phosphate-isopentenyltransferase gene, AtIPT3, and a cytochrome P450 monooxygenase gene, CYP735A2. The expression of these genes was inhibited by a photosynthesis inhibitor, DCMU, under elevated CO2, and was enhanced by sugar supplements, indicating that photosynthetically generated sugars are responsible for the induction. Consistently, cytokinin precursor accumulation was enhanced by sugar supplements. Cytokinin biosynthetic mutants were impaired in growth enhancement under elevated CO2, demonstrating the involvement of de novo cytokinin biosynthesis for a robust growth response. We propose that plants employ a system to regulate growth in response to elevated CO2 in which photosynthetically generated sugars induce de novo cytokinin biosynthesis for growth regulation.


Results elevated Co 2 increases cytokinin precursor concentrations in shoots and roots.
To examine the effects of elevated CO 2 on growth and CK levels, plants were grown under low [280 parts per million by volume (ppmv)] and high CO 2 (780 ppmv) on soil. Two hundred and eighty ppmv is the pre-industrial atmospheric concentration and 780 ppmv is a value close to the median of values predicted at the end of this century 40 . When wild-type Arabidopsis Col-0 were germinated and grown under low or high CO 2 with a 12 h light/12 h dark photoperiod for four weeks, high CO 2 -grown plants deposited more biomass and developed more leaf area and rosette leaves than low CO 2 -grown plants did, as described previously ( Supplementary Fig. S1 32,41,42 ). Using the same growth conditions, we analysed changes in the CK concentration following exposure to high CO 2 . Sixteenday-old Col-0 plants grown in low CO 2 were transferred to low or high CO 2 , and CK concentrations in the whole shoot were followed for four days. Under these conditions, significant differences in shoot fresh weight between high and low CO 2 -treated plants became evident from day 4 onward (Fig. 1a). The levels of iP-type CK precursors (iPR and iPRPs) and tZ-type CK precursors (tZR and tZRPs) in high CO 2 -treated shoots increased after one day and stayed high until day 4 compared with those of low CO 2 -treated plants (Fig. 1b,c). On the other hand, concentrations of other CK metabolites including inactivated iP-type CKs (iP7G and iP9G), and tZ-type CKs (tZ7G, tZ9G, tZOG, tZROG, and tZRPsOG) did not change consistently during the period of observation (Fig. 1b,c; Supplementary Table S1). Furthermore, the high CO 2 -treatment did not significantly affect the levels of other plant hormones, including a gibberellin precursor (GA 24 Table S1). These results showed that iP-type and tZ-type CK precursors accumulate in the shoot prior to growth enhancement at high CO 2 under our experimental conditions. Next, we employed a growth system in which Col-0 seedlings were germinated and grown on half-strength MS (1/2 MS) agar plates placed vertically to allow the analysis of both shoots and roots. Twelve-day-old wild-type seedlings grown under continuous light in low CO 2 were transferred to low or high CO 2 , and the CK concentrations in shoots and roots were measured after 6 h and 24 h. The basal level of tZ, tZRPs and tZ-N-conjugates in this measurement (Supplementary Table S2) was very different from that in soil-grown plants (Supplementary  Table S1). This is possibly due to differences in growth conditions and plant ages, as a similar trend has been observed previously 17 . Accumulation of tZ, and iP-type and tZ-type precursors became evident in shoots and roots as early as 6 h after commencing the high CO 2 -treatment and continued until 24 h, whereas the levels of other CK metabolites did not consistently change (Fig. 2a,b; Supplementary Table S2).
It is known that the accumulation of CK precursors generally results in increased CK activity 7,43 . To verify that CK signalling is activated in parallel with precursor accumulation, the expression of immediate-early CK responsive type-A ARR genes was analysed in whole seedlings treated as in Fig. 2a. As expected, ARR4, ARR6, and ARR15 were induced, with the timing of induction similar to that of CK precursor accumulation (Fig. 2c). Since recent studies on plant membrane binding and crystal structure analysis showed that precursors do not bind to Arabidopsis CK receptors 44,45 , one would expect that active CKs (iP and tZ) are accumulated in response to an increase in CK precursor levels. However, active CKs were not always increased significantly in our experiments (for example, Supplementary Tables S1, S2). This lack of significant change in active CK levels has been reported previously 46,47 and we assume that it is because only a fraction of active CKs exists in a compartment where they can be perceived by CK receptors. Taken together, these results indicated that elevated CO 2 resulted in increased CK activity, which is triggered by CK precursor accumulation in shoots and roots.
Cytokinin biosynthetic genes, AtIPT3 and CYP735A2, are up-regulated in roots under elevated Co 2 . Generally the accumulation of CK precursors reflects increased de novo biosynthesis 7,43 . The initial step of de novo CK biosynthesis is catalysed by IPT 13,14 , and the key step of tZ-type de novo CK biosynthesis requires CYP735A 17,18 . We examined the expression levels of seven IPT (AtIPT1, AtIPT3, AtIPT4, AtIPT5, AtIPT6, AtIPT7, AtIPT8) and two CYP735A (CYP735A1 and CYP735A2) genes in Arabidopsis shoots and roots of Col-0 seedlings incubated at low or high CO 2 from 3 h, an earlier time point than when CK precursor accumulation was observed (up to 9 h). AtIPT4, AtIPT6, and AtIPT8 were not detected in shoots nor roots in our experimental conditions. In shoots, none of the genes examined were affected by high CO 2 except for AtIPT5 that was down-regulated at 9 h (Fig. 3a). In roots, the transcript level of CYP735A2 increased after 3 h and stayed high till 9 h and that of AtIPT3 steadily accumulated after the onset of high CO 2 treatment (Fig. 3b). On the other hand, the levels of the www.nature.com/scientificreports www.nature.com/scientificreports/ other transcripts remained unchanged or showed transient fluctuations (Fig. 3b). Down-regulation of AtIPT5 in both shoot and root might be caused by accumulated CK because AtIPT5 has been reported to be repressed by CK 48 . Since CK levels are determined by the balance between de novo biosynthesis and degradation, we also analysed the expression of genes encoding CK-degrading enzymes, CKX. Among seven CKXs in Arabidopsis, the expression of six genes was detected but none of these genes were down-regulated in shoots and roots under high CO 2 treatment ( Supplementary Fig. S2). Rather, the expression of CKX1, CKX4, CKX6, and CKX7 was transiently enhanced, possibly in response to CK accumulation ( Supplementary Fig. S2). Similar CK precursor accumulation and induction of AtIPT3 and CYP735A2 were observed when seedlings were grown and treated under 12-h-light/12-h-dark cycles (Supplementary Fig. S3; Supplementary Table S3). These results suggested that the induction of AtIPT3 and CYP735A2 in roots plays a role in iP-and tZ-type CK precursor accumulation under elevated CO 2 .
photosynthetically generated sugars induce AtIPT3 and CYP735A2 in roots. Next, we tested the involvement of photosynthesis in the induction of AtIPT3 and CYP735A2 by incubating wild-type seedlings in the dark or by applying the photosynthesis inhibitor DCMU, which blocks electron flow from photosystem II. When seedlings were incubated in the dark or in the light with DCMU at 280 ppmv for 6 h, expression levels of AtIPT3 and CYP735A2 were reduced compared to the control (Fig. 4a,b). The induction of AtIPT3 and . Asterisks indicate statistically significant differences between 280 ppmv CO 2and 780 ppmv CO 2 -treated samples at the same exposure time (*p < 0.05; **p < 0.01; Student's t-test). FW, fresh weight; tZ, trans-zeatin; iP, N 6 -(∆2-isopentenyl)adenine; iP-type CK precursor, sum of iPR and iPRPs; inactivated iP-type CK, sum of iP7G and iP9G; tZ-type CK precursor, sum of tZR and tZRPs; inactivated tZtype CK, sum of tZ7G, tZZ9G, tZOG, tZROG, and tZRPsOG. The concentrations of all quantified hormones are shown in Supplementary Table S1. www.nature.com/scientificreports www.nature.com/scientificreports/ CYP735A2 in response to elevated CO 2 was completely abolished by these treatments (Fig. 4a,b), indicating that photosynthetic activity is required for the maintenance and induction of AtIPT3 and CYP735A2 expression.
Elevated CO 2 -treatment reportedly increases endogenous sugar concentrations (e.g. fructose, glucose, and sucrose), whereas DCMU treatment reduces sugar levels 31,37,49 . To examine whether the DCMU-triggered attenuation of AtIPT3 and CYP735A2 induction were caused by lowered levels of sugars, we supplemented DCMU-treated seedlings with sucrose. Sucrose reversed the effect of DCMU on AtIPT3 and CYP735A2 expression (Fig. 4c,d). We also tested the effects of other sugars on AtIPT3 and CYP735A expression. Seedlings were transferred to agar plates containing metabolizable sugars (sucrose and glucose) or non-metabolizable sugars (sorbitol and mannitol) and were incubated at 280 ppmv CO 2 in the dark. Metabolizable sugars were able to induce AtIPT3 and CYP735A expression (Fig. 4e,f). On the other hand, non-metabolizable sugars were ineffective in inducing AtIPT3 expression (Fig. 4e). CYP735A2 expression was induced by non-metabolizable sugars but to a much lower extent compared with metabolizable sugars (Fig. 4f). Since CYP735A2 seems to be moderately activated by osmotic stress 50 , the induction by non-metabolizable sugars is probably due to osmotic effects.
CYP735A2 is known to be a CK-inducible gene 18,21 . Thus we tested whether CYP735A2 induction by elevated CO 2 and sugars is the result of accumulated CKs by employing ipt3 ipt5 ipt7 (ipt357) and the cytokinin receptor mutants ahk2 ahk3 and ahk3 ahk4 [51][52][53] that are defective in CK biosynthesis and signalling, respectively. Elevated CO 2 and sugars induced CYP735A2 expression in the mutants at a level comparable to Col-0 ( Fig. 5a,b), indicating that sugars induce this gene independently of CK.
These results suggested that AtIPT3 and CYP735A2 are induced in roots under elevated CO 2 by sugars generated in shoots by photosynthesis. Consistent with this, sucrose treatment resulted in an accumulation of CK precursors in shoots and roots (Fig. 4g,h; Supplementary Table S4).

Figure 2.
Cytokinin levels and activity in seedlings exposed to high CO 2 . (a,b) Cytokinin (CK) levels in shoots and roots of Col-0 seedlings exposed to low and high CO 2 . iP-type CK precursor levels (a) and tZtype CK precursor levels (b) in shoots and roots are presented. (c) Expression levels of type-A ARR genes in Col-0 seedlings exposed to low and high CO 2 . Transcript levels of ARR4, ARR6, and ARR15 were analysed by quantitative RT-PCR. Expression levels were normalized using At4g34270 as an internal control. Twelve-dayold seedlings grown on 1/2 MS agar plates at 280 ppmv were exposed to 280 ppmv (280) or 780 ppmv (780) CO 2 for the indicated periods. Error bars represent standard deviations of three biological replicates. Asterisks indicate statistically significant differences between 280 ppmv CO 2 -and 780 ppmv CO 2 -treated samples at the same exposure time (*p < 0.05; **p < 0.01; Student's t-test). FW, fresh weight; tZ, trans-zeatin; iP, N 6 -(∆ 2isopentenyl)adenine. The concentrations of cytokinin molecular species are shown in Supplementary Table S2. www.nature.com/scientificreports www.nature.com/scientificreports/ Photosynthetically generated sugars induce cytokinin precursor accumulation irrespective of the nitrate status. It is known that de novo CK biosynthesis is regulated by nitrate in Arabidopsis 54 . Since carbon availability is reported to influence nitrate transporter gene expression and nitrate uptake 55,56 , it is possible that carbon availability affects CK biosynthesis indirectly through nitrate-related pathways. To test this possibility, we measured CK levels in wild-type seedlings treated with high CO 2 or sucrose, and with and without nitrate. Twelve-day-old seedlings were treated with high CO 2 or sucrose on agar plates containing 10 mM KNO 3 , 10 mM NH 4 Cl, or no nitrogen source, and the CK concentrations in the whole seedling were measured after 24 h. The levels of CK precursors increased in all nitrogen conditions tested in response to high CO 2 or sucrose treatment ( Fig. 6; Supplementary Tables S5 and S6), suggesting that sugars induce CK precursor accumulation independent of nitrate-related pathways. the ipt3 cyp735a2 mutant still accumulates cytokinin precursors in response to elevated CO 2 .
Having shown the relevance of AtIPT3 and CYP735A2 in the elevated CO 2 -enhanced de novo CK biosynthesis, we investigated whether these processes contribute to growth enhancement under elevated CO 2 by generating an ipt3 cyp735a2 double mutant. To this end, 12-day-old seedlings grown on agar plates were incubated at low or high CO 2 for seven days. Fresh weight (FW) was measured before and after the treatment, and relative growth rate (RGR) was calculated. Growth differences of Col-0 between low CO 2 -and high CO 2 -incubated seedlings were clearly observed; the FW and RGR of both the shoot and the root were significantly increased by the high CO 2 treatment (Supplementary Fig. S4a-c). However, no significant difference in the FW and RGR was observed between the double mutant and WT ( Supplementary Fig. S4a-c). To understand this lack of growth phenotype, we analysed changes in iP-and tZ-type precursor CK levels in shoots and roots of the double mutant following exposure to high CO 2 . Under low CO 2 , the double mutant contained significantly reduced levels of iP-type CK precursors in shoots (Supplementary Fig. S4d; Supplementary Table S7). However, it accumulated both CK precursors in both organs in response to high CO 2 -treatment, though the levels of accumulation were generally lower compared with WT ( Supplementary Fig. S4d,e; Supplementary Table S7), indicating that AtIPT3 and CYP735A2 are not the only factors mediating the elevated CO 2 -induced CK precursor accumulation. Together, these results suggest that the double mutant lacks a growth phenotype because it still is able to accumulate enough CKs for elevated CO 2 -triggered growth enhancement. Expression of genes involved in cytokinin biosynthesis in shoots and roots upon exposure to high CO 2 . Transcript levels of AtIPT1, AtIPT3, AtIPT4, AtIPT5, AtIPT6, AtIPT7, AtIPT8, CYP735A1, and CYP735A2 were analysed in shoots (a) and roots (b) of Col-0 seedlings by quantitative RT-PCR. Expression levels of AtIPT4, AtIPT6, and AtIPT8 were below the detection limit in shoots and roots. Expression levels were normalized using At4g34270 as an internal control. Twelve-day-old seedlings grown on 1/2 MS agar plates at 280 ppmv were exposed to 280 ppmv (280) or 780 ppmv (780) CO 2 for the indicated periods. Error bars represent standard deviations of three biological replicates. Asterisks indicate statistically significant differences between 280 ppmv CO 2 -and 780 ppmv CO 2 -treated samples at the same exposure time (**p < 0.01; *p < 0.05; Student's t-test). www.nature.com/scientificreports www.nature.com/scientificreports/ the ipt3 ipt5 ipt7 and cyp735a1 cyp735a2 mutants are impaired in elevated CO 2 -triggered growth enhancement. Since the ipt3 cyp735a2 double mutant still accumulated CKs in response to high CO 2 ( Supplementary Fig. S4d,e), we employed higher order CK-biosynthetic mutants, ipt357 and cyp735a1 cyp735a2 (cypDM). The ipt357 mutant lacks three major IPT genes 57 and, thus, has a dramatically reduced ability to de novo synthesize both iP-and tZ-type CKs. The cypDM mutant lacks all CYP735A genes 17 and, thus, is expected to accumulate iP-type CKs but not tZ-type CKs under elevated CO 2 . To verify that elevated CO 2 -induced de novo CK biosynthesis is attenuated in ipt357 and cypDM, the CK concentrations in shoots and roots were measured. Seedlings were grown and treated as in Fig. 2 (24 h high CO 2 treatment). In the ipt357 mutant, the accumulation level of all CKs was relatively low compared with the wild type. The iP-type CK precursor concentrations were unaffected in shoots and roots, but the levels of tZ-type precursor CKs increased slightly in shoots with a high CO 2 treatment (Fig. 7a,b; Supplementary Table S8). In the cypDM mutant, iP-type precursor CKs accumulated but tZ-type precursor CKs levels were consistently low in shoots and roots under elevated CO 2 (Fig. 7a,b; Supplementary Table S8). These observations confirmed the inability of the mutants to accumulate CKs of the expected types under elevated CO 2 . These results showed that de novo CK biosynthesis, most likely mediated by IPT3, IPT5, IPT7, CYP735A1 and CYP735A2, plays an important role in CK accumulation in response to high CO 2 .
We then investigated whether these mutants are impaired in growth enhancement under elevated CO 2 . Seedlings were grown and treated on agar plates as in Supplementary Fig. S4, and the FW was measured before and after treatment, and the RGRs were calculated. In Col-0, the FW and RGR of both the shoot and the root were dramatically increased in response to high CO 2 (Fig. 7c,d; Supplementary Fig. S5). The ipt357 mutant also gained more FW both in shoots and roots in high CO 2 compared with the low CO 2 treatment, but the extent of the increase was smaller compared with that of Col-0 (Fig. 7c). RGR analysis revealed that shoots and roots of ipt357 grew faster in high CO 2 than in low CO 2 but at a lower rate compared with those of Col-0, whereas the RGR in low CO 2 was similar among all genotypes in this growth system (Fig. 7d). Interestingly, the cypDM mutant displayed essentially the same growth response defects to elevated CO 2 as the ipt357 mutant (Fig. 7c,d), showing that accumulation of tZ-type CKs is critical for the response.
We also analysed the growth response of soil-grown plants. The mutants were germinated and grown on soil together with Col-0 under low or high CO 2 , and shoot growth was analysed at 17 and 31 days after germination (DAG) by measuring dry weight (DW). Note that it was not possible to evaluate root growth in this system. Although Col-0 plants grown in high CO 2 had significantly higher shoot biomass compared with those grown in low CO 2 at the beginning of analysis (17 DAG), they gained more biomass by further growth in high CO 2 (31 DAG, Supplementary Fig. S6). RGRs between 17 and 31 DAG were significantly higher in high CO 2 -grown Col-0 plants than in the mutants (Fig. 7e). The number of rosette leaves counted on 31 DAG also significantly increased (Fig. 7f). Although the cypDM mutant gained more biomass under high CO 2 ( Supplementary Fig. S6), no significant change in RGR in response to high CO 2 treatment was observed (Fig. 7e). The RGR of the ipt357 mutant was slightly enhanced by high CO 2 treatment (Fig. 7e). Rosette leaf numbers did not change in the cypDM mutant and were only marginally increased in the ipt357 mutant (5.4 more leaves in the wild type compared with 1.9 in ipt357) in response to high CO 2 (Fig. 7f). These results show that the ipt357 and cypDM mutants are impaired in the acceleration of shoot growth and development under elevated CO 2 during the growth period examined and that the cypDM mutant, which cannot accumulate tZ-type CKs, is severely compromised.
Together, these growth analyses suggest that CK accumulation, especially of the tZ-type, through de novo biosynthesis contributes to robust growth enhancement under elevated CO 2 .
photosynthetically generated sugars induce ABCG14 in roots. It has been reported that tZ-type CKs are translocated from root to shoot by the ABCG14 protein to act as shoot growth signals 17,20,21,58 . To get insight into whether root-to-shoot translocation of CKs is relevant to the observed CK accumulation in shoots, ABCG14 expression in roots was investigated (Fig. 8). Interestingly, ABCG14 expression responded to high CO 2 and sugars in a similar manner to that of AtIPT3 and CYP735A2 (Fig. 8). Since ABCG14 has been reported to be CK-inducible 21 , we tested whether the CKs that accumulate in response to elevated CO 2 and sugars are relevant to ABCG14 induction. The ipt357 and the cytokinin receptor mutants ahk2 ahk3 and ahk3 ahk4 51-53 were analysed as in Fig. 5a,b. ABCG14 induction in response to elevated CO 2 and sugars was maintained in these mutants (Fig. 5c,d), indicating that sugars induce this gene independent of CK. Together, these results suggest that root-to-shoot translocation of CKs via ABCG14 might be involved in robust growth enhancement under elevated CO 2 by mediating tZ-type CK accumulation in the shoot.

Discussion
The availability of macronutrients such as nitrogen 7,43,48,59,60 , phosphate 9,61-63 and sulphate 9,64 affects IPT expression as well as CK levels. Therefore, macronutrient availability has been proposed to regulate CK levels through de novo biosynthesis to control plant growth and development 9,27,65 . Our investigation has revealed another pathway in which photosynthesis-derived sugars regulate de novo CK biosynthesis to control plant growth and development.
differences (*p < 0.05; Student's t-test). Error bars represent standard deviations of four biological replicates. Asterisks indicate statistically significant differences (*p < 0.05; Student's t-test). Different lower-case letters indicate statistically significant differences as indicated by Tukey's HSD test (p < 0.05). Expression levels were analysed by quantitative RT-PCR and normalized using At4g34270 as an internal control. (2019) 9:7765 | https://doi.org/10.1038/s41598-019-44185-4 www.nature.com/scientificreports www.nature.com/scientificreports/ Our study suggests that de novo CK biosynthesis is triggered by photosynthetically generated sugars (Figs 1-6). There are several other reports indicating that sugars induce the expression of genes involved in the de novo synthesis of CKs. Transcriptome analyses show that glucose 66 and sucrose 67,68 treatments up-regulate AtIPT3 and CYP735A2. However, how sugars are perceived (as signalling molecules, energy sources or building blocks) to induce the expression of these genes is still not understood. Thus, it is possible that sugars act indirectly through the signalling pathways of macronutrients because the metabolism of carbon and macronutrients are tightly intertwined. Although our data suggests that sugars induce CK precursor accumulation independent of nitrate-related pathways (Fig. 6), Kamada-Nobusada et al. 7 reported a pathway in which the internal nitrogen status regulates CK biosynthesis. Since the internal nitrogen status can also be modulated by carbon availability, we cannot rule out the possibility that sugars affect CK biosynthesis through this pathway. Further studies on sugar and internal nitrogen sensing and signalling mechanisms are required to resolve this problem. In any case, we propose that sugars generated by photosynthesis in shoots directly or indirectly promotes de novo CK biosynthesis.
Under our experimental conditions, AtIPT3 was the only gene of the AtIPT family induced by elevated CO 2 and sugars (Figs 3, 4). AtIPT3 expression is also regulated by various environmental signals to control CK levels; increases in nitrogen, phosphate, and sulphate availability induce AtIPT3 expression 9,43,48,63 , whereas drought and Figure 5. Expression of CYP735A2 and ABCG14 in cytokinin biosynthetic and signaling mutants treated with high CO 2 or sugars. (a,c) Wild-type (Col-0), ipt3 ipt5 ipt7 (ipt357), ahk2 ahk3 (ahk23) and ahk3 ahk4 (ahk34) seedlings grown on 1/2MS agar plates at 280 ppmv CO 2 for 12 days were exposed to 280 ppmv or 780 ppmv for six hours and then roots were harvested. (b,d) ipt357 seedlings were transferred to new plates containing 90 mM of sorbitol (Sorb), mannitol (Man), sucrose (Suc) or glucose (Glc), or without any sugar (-sugar). Roots were harvested after six hours. Expression levels were analysed by quantitative real-time PCR and normalized using At4g34270 as an internal control. Error bars represent standard deviation of three biological replicates. Different lower-case letters indicate statistically significant differences as indicated by Tukey's HSD test (p < 0.01). www.nature.com/scientificreports www.nature.com/scientificreports/ salt stress repress AtIPT3 expression 69 . Although it remains unclear -with the exception of nitrate -whether these signals regulate AtIPT3 expression directly 70,71 , the available evidence suggests that AtIPT3 functions to integrate and translate various signals in the root into de novo CK biosynthetic activity. We also found that CYP735A2 and ABCG14 are high CO 2 -and sugar-inducible (Figs 3, 4, 5, 8). Although CYP735A2 and ABCG14 are known to be CK-inducible genes 18,21 , we showed that sugars induce these genes independent of CK (Fig. 5). These results indicate that CYP735A2 and ABCG14 are controlled by two independent signals: shoot-derived signals (sugars) and root internal cues (root-synthesized CKs) in the response to elevated CO 2 . Thus, CYP735A2 and ABCG14 might act to integrate signals from shoots and roots and translate these signals into tZ-type CKs translocated from root to shoot. However, it remains to be determined whether ABCG14-mediated root-to-shoot translocation activity is regulated at the level of expression or by some other means.
In our expression analysis, AtIPT3 and CYP735A2 were the only genes induced under elevated CO 2 among the de novo CK biosynthetic genes (Fig. 3). However, the ipt3 cyp735a2 double mutant still accumulated CKs, although at a lower level than WT (Supplementary Fig. S4d,e). The ipt357 and cypDM mutants were unable to accumulate iP-type and tZ-type CKs, respectively, in response to high CO 2, suggesting that not only AtIPT3 and CYP735A2 but also AtIPT5, AtIPT7, and CYP735A1 are involved in the accumulation. Since these genes were not found to be regulated at the level of transcript accumulation, post-transcriptional regulation might be involved. Figure 6. Cytokinin levels in wild-type seedlings exposed to high CO 2 or treated with sucrose under different nitrogen nutrient conditions. (a) Cytokinin (CK) levels in wild-type (Col-0) whole seedlings exposed to low or high CO 2 under different nitrogen nutrient conditions. Seedlings were exposed to 280 ppmv (280) or 780 ppmv (780) CO 2 for 24 h on modified 1/2 MS agar plates containing 10 mM KNO 3  www.nature.com/scientificreports www.nature.com/scientificreports/ Figure 7. Cytokinin levels and growth of wild-type, ipt3 ipt5 ipt7 and cyp735a1 cyp735a2 seedlings exposed to high CO 2 . (a,b) The concentration of iP-type cytokinin (CK) precursors (a) and tZ-type CK precursors (b) in shoots and roots of wild-type (Col-0), ipt3 ipt5 ipt7 (ipt357) and cyp735a1 cyp735a2 (cypDM) plants exposed to 280 ppmv (280) or 780 ppmv (780) CO 2 for 24 h. Asterisks indicate statistically significant differences (**p < 0.01; *p < 0.05; Student's t-test). The concentrations of cytokinin molecular species are shown in Supplementary Table S7. (c,d) Fresh-weight (c) and relative growth rate (RGR) (d) of 19-day-old wild type (Col-0), ipt3 ipt5 ipt7 (ipt357), and cyp735a1 cyp735a2 (cypDM) seedlings treated under 280 ppmv (280) or 780 ppmv (780) CO 2 for seven days. (d) RGR was calculated using the fresh weight (FW) data obtained previously (Supplementary Fig. S5) and after (c) low or high CO 2 treatment. (e,f) Shoot growth of soil-grown wild-type, ipt3 ipt5 ipt7 and cyp735a1 cyp735a2 plants under low or high CO 2 . (e) Relative growth rates (RGR) of shoots of Col-0, ipt357, and cypDM grown under 280 or 780 on soil. Dry weights of shoots shown in Supplementary  Fig. 6b were used to calculate the RGR. Asterisks indicate statistically significant differences (**p < 0.001; www.nature.com/scientificreports www.nature.com/scientificreports/ Consistently, it has been reported that AtIPT3 farnesylation modulates this protein's subcellular localization and enzymatic properties 72 . It should be noted that we cannot exclude that other genes involved in CK biosynthesis, modification, and/or degradation, and/or post-transcriptional regulation might be relevant to the accumulation of CKs. In this study, the role of CKs in growth enhancement under elevated CO 2 was evaluated by analysing the growth of ipt357, a mutant deficient in iP-and tZ-type CKs, and cypDM, a mutant deficient in tZ-type CKs. Both mutants displayed similar growth response defects (Fig. 7), indicating that tZ-type CKs are required for robust growth enhancement of shoots and roots under elevated CO 2 . A reduction in shoot growth acceleration in these mutants is consistent with previous reports that tZ-type CKs and their root-to-shoot translocation act to promote shoot growth 17,20,21 . However, a reduction in root growth acceleration cannot be explained by CK action because CKs generally act to repress root growth 73,74 . This result suggests that CK is not the major determinant of root growth rate. It is plausible that slowed root growth is a consequence of reduced photosynthesis (as sources of energy and building blocks) by smaller shoots, but it is also possible that complex crosstalk might exist between CK and sugars.
Here, we revealed that sugar-induced de novo biosynthesis of CKs plays a role in the robust growth enhancement under elevated CO 2 . This finding provides some insight into the mechanisms that plants employ to optimise growth in a fluctuating environment. Taking into account that AtIPT3, CYP735A2, and ABCG14 are induced in the root by photosynthetically generated sugars (Figs 3,4,5,8), it is tempting to speculate that there is a systemic growth regulatory mechanism in which photosynthetically generated sugars induce de novo tZ-type CK biosynthesis in the root and root-to-shoot translocation of the CK via ABCG14 for growth regulation of the shoot.

Materials and Methods
Plant material and growth conditions. Arabidopsis thaliana ecotype Columbia (Col-0) was used as the wild type. The cytokinin biosynthetic triple mutants ipt3 ipt5 ipt7 57 , the cytokinin receptor double mutants ahk2 ahk3 and ahk3 ank4 53 , and the cyp735a1-2 cyp735a2-2 double mutant 17 were characterized previously. The ipt3 cyp735a2-1 and ipt3 cyp735a2-2 double mutants were generated by crosses between the ipt3 ipt5 ipt7 and the cyp735a1-2 cyp735a2-1 mutant, and cyp735a1-2 cyp735a2-2 mutants. For studies on soil-grown plants, stratified seeds were sown directly on nutrient-rich soil (Supermix A, Sakata, Japan), and grown in a CO 2  Quantification of plant hormones. Cytokinin level was determined using an ultra-performance liquid chromatograph coupled with a tandem quadrupole mass spectrometer equipped with an electrospray interface as described previously 75 . IAA and ABA levels were determined using an ultra-high-performance liquid chromatography (UHPLC)-electrospray interface (ESI) and a quadrupole-orbitrap mass spectrometer (UHPLC/Q-Exactive; Thermo Scientific) as described previously 76 . In the results reported, the category iP-type CK precursors comprise iPR and iPRPs; inactivated iP-type CK comprise iP7G and iP9G; tZ-type CK precursors comprise tZR and tZRPs; and inactivated tZ-type CK comprise tZ7G, tZ9G, tZOG, tZROG, and tZRPsOG.

Gene expression analysis. Total RNA was extracted from root and shoot samples using the RNeasy Plant
Mini kit (QIAGEN) in combination with the RNase-Free DNase set (QIAGEN). Total RNA was used for first strand cDNA synthesis by the SuperScript III First-Strand Synthesis System (Life Technologies) with oligo(dT) 20 primers. Quantitative reverse transcription-PCR (RT-PCR) was performed on a StepOnePlus Real-Time PCR system (Applied Biosystems) with the KAPA SYBR Fast qPCR kit (KAPA Biosystems). At4g34270 was used as an internal control because this gene has been shown to be one of the most stably expressed genes in Arabidopsis 77,78 . Similar results were obtained using other internal control genes (At1g13320 and At2g28390) as described by Czechowski et al. 78 . Primer sets are listed in Supplementary Table S9. DCMU and sugar treatment. Col-0 seedlings grown on 1/2 MS agar plates (1% agar) placed vertically under continuous fluorescent light (120 µmol m −2 s −1 ) at 22 °C in a CO 2 -controlled growth chamber at 280 ppmv CO 2 were sprayed with 40 µM DCMU or mock solution (0.05% ethanol) and exposed to 280 ppmv or 780 ppmv CO 2 under 120 µmol m −2 s −1 light or in the dark. The DCMU stock solution was 40 mM in 50% ethanol. For DCMU and sucrose co-treatment, seedlings were treated with 40 µM DCMU or mock solution (0.05% ethanol) and then transferred to 1/2 MS agar plates (1% agar) containing 90 mM sucrose. For sugar treatment, seedlings were transferred to 1/2 MS agar plates (1% agar) containing 90 mM of sorbitol, mannitol, sucrose, glucose, or 45 mM sucrose. *p < 0.01; not significant (ns), p > 0.01; two-way ANOVA). (f) Rosette leaf number of Col-0, ipt357, and cypDM counted at 31 DAG. Error bars represent standard deviations (a, n = 3; b, n = 3; c, n = 9; f, n = 10) and standard error (d, n = 9; e, n = 9). Lower-case letters denote statistically significant classes (Tukey's HSD test, p < 0.05).
www.nature.com/scientificreports www.nature.com/scientificreports/ High Co 2 and sugar treatment under different nitrogen conditions. Wild-type seedlings were pre-grown for 11 days on modified 1/2 MS agar plates (1% agar) containing 10 mM KNO 3 , 10 mM NH 4 Cl or 5 mM NH 4 NO 3 as the sole nitrogen source in the CO 2 -controlled growth chamber at 280 ppmv. Seedlings grown with 10 mM KNO 3 , 10 mM NH 4 Cl or 5 mM NH 4 NO 3 were then transferred to new 1/2 MS agar plates (1% agar) containing 10 mM KNO 3 , 10 mM NH 4 Cl or no nitrogen source, respectively. After 24 h incubation at 280 ppmv, seedlings were subjected to high CO 2 and sugar treatments under the same nitrogen conditions. Growth analysis under low or high CO 2 . For growth analysis of soil-grown plants, stratified seeds were sown directly on nutrient-rich soil (Supermix A, Sakata, Japan), and grown in a CO 2 -controlled growth chamber at 280 ppmv or 780 ppmv CO 2 under 150 µmol m −2 s −1 fluorescent light (12 h light/12 h dark) at 22 °C and 60% relative humidity. Shoots were harvested at 17 and 31 days after germination (DAG) and their dry weights were determined after drying them in an oven set at 80 °C for three days. Rosette leaf number was counted on 31 DAG.
For seedling growth analysis, surface sterilized seeds were sown on 1/2 MS agar plates (1% agar) containing 1% sucrose. After stratification, plates were placed vertically in a CO 2 -controlled growth chamber (120 µmol m −2 s −1 continuous fluorescent light, 22 °C) at 280 ppmv. Five-day-old seedlings were transferred to 1/2 MS agar plates (1% agar without sucrose) and grown vertically for another 7 days at 280 ppmv. Then, the 12-day-old seedlings were exposed to 280 ppmv or 780 ppmv CO 2 for seven days. The shoots and roots were separated and their fresh Figure 8. Effects of high CO 2 , photosynthesis and sugars on the expression of ABCG14 in roots. (a) Expression levels of ABCG14 in roots of Col-0 seedlings exposed to 280 ppmv (280) or 780 ppmv (780) CO 2 for the indicated periods. Treatment was conducted as in Fig. 3. (b) Effects of dark and DCMU on ABCG14 in Col-0 roots. Treatments were conducted as in Fig. 4. (c,d) Effects of sugars on ABCG14 in Col-0 roots. Treatments were conducted as in Fig. 4. Expression levels were analysed by quantitative RT-PCR and normalized using At4g34270 as an internal control. Error bars represent standard deviations of four biological replicates. Asterisks indicate statistically significant differences (*p < 0.05; Student's t-test). Different lower-case letters indicate statistically significant differences as indicated by Tukey's HSD test (p < 0.05).
To avoid any chamber effects, we used two growth chambers simultaneously with different CO 2 concentrations and repeated each experiment at least twice with different chamber and CO 2 concentration combinations. Although the data presented are from one representative experiment, similar results were obtained from different chamber and CO 2 concentration combinations.

Data Availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.