Single-node cuttings of R. hybrida ‘Radrazz’ rose bushes were obtained as in Demotes-Mainard et al. (2013a). When the second leaf of the primary axis unfolded, the young plants were transferred to two identical growth chambers (Froids et Mesures, Beaucouzé, France). All lamps were metal halide lamps (HQI, OSRAM, München, Germany). The light spectrum in the growth chambers is presented in Supplementary Material 1, R:FR was 1.99 over the ranges (655–665 nm):(725–735 nm). Light spectrum was measured with a spectrophotometer (Avaspec-2048-6-RM, Avantes, Apeldoorn, The Netherlands). The plants were spaced 15 cm apart and surrounded by a border row. The photoperiod was 16h day/8h night, air temperature was 18°C day/17°C night, and humidity was 69%. The plants were sub-irrigated, and fertilized as in Demotes-Mainard et al. (2013a). The shoot apical meristem of the primary axis produced vegetative phytomers before differentiating into a terminal flower bud. The flower bud was first hidden within the unfolded young leaves until it became visible, and then referenced to as the FBV stage. It coincides in R. hybrida ‘Radrazz’ with the outgrowth of the first buds on the primary axis.

All plants were exposed to a to a PPFD of 529 ± 6 μmol m^-2 s^-1 starting when the plants were placed in the growth chamber until the FBV stage. At FBV stage the plants were randomly assigned to one of the two following light treatments. The first treatment, hereafter referred to as “high PPFD,” corresponded to exposure to a continuous PPFD of 529 ± 6 μmol m^-2 s^-1 until 14 days after the FBV stage. In the second treatment, hereafter referred to as “low PPFD,” the plants were exposed to 89 ± 1 μmol m^-2 s^-1 from the FBV stage to 14 days after the FBV stage. PPFD was measured with a horizontal cosine corrector quantum sensor (LI-190 Quantum Sensor, LI-COR, Lincoln, NE, United States).

The fourth internode from the apex (not counting the peduncle, see Supplementary Material 2A) was collected at the FBV stage (day 0), days 4 and 8 after the FBV stage on plants grown under high or low PPFD. Samples were collected 6 h after the beginning of the light period, immediately frozen in liquid nitrogen and stored at -80°C. Then they were lyophilized and crushed. Sucrose, glucose, fructose, and starch (enzymatic digestion step) were determined by colorimetry (Konelab), as presented in Supplementary Material 3. CKs, free ABA, and free IAA concentrations were determined as in Li-Marchetti et al. (2015). Sugar and hormone contents were determined from four replicates with five plants per light treatment each.

To check that sugars supplied by both the cotton-wick method and the rachis-feeding method were indeed delivered to the stem tissues, the plants grown under low PPFD were supplied with labeled sucrose. Labeled ¹³C-sucrose (ID 605417_Aldrich, Sigma–Aldrich, Darmstadt, Germany) was incorporated at 3% in a solution of 200 mM sucrose supplied by a cotton wick just below node 4, or incorporated at 3.4% in a solution of 600 mM sucrose supplied through the rachis of leaf 4. Water supplied either by a cotton wick or through the rachis served as a control. Exogenous supply started at the FBV stage onward. Days 4 and 8 after the FBV stage, the third, fourth, and fifth internodes from the apex (not counting the peduncle, Supplementary Material 2A) were sampled as described previously. Total carbon and the ¹³C:¹²C ratio were determined for each sample using an elemental analyzer (NA 1500 NCS; Carlo Erba, Milan, Italy) coupled with a Delta-S isotope ratio mass spectrometer (Finnigan-Mat; Thermoquest Corp., San Jose, CA, United States). δ¹³C was calculated as:

where R is the ratio of heavy-to-light isotope, and the R reference for ¹³C in the PDB standard is 0.0112372.

Three to six times a week from FBV onward, the state (dormant or growing out) of each bud along the primary axis was noted, and bud length was measured with a ruler. Buds were considered to have grown out when at least one visible leaf protruded between the two bud scales (Girault et al., 2008; Roman et al., 2016). The mean bud lengths were calculated from the sizes of outgrowing and dormant buds from a same batch. Fourteen days after the FBV stage, buds were dissected using a stereomicroscope. Bud meristem organogenic activity, which results in the formation of new organs in the bud, was assessed by counting the number of leaves in the bud, leaves being here defined as young leaves plus foliar primordia, excluding the two bud scales. Measurements were performed on 8–16 plants per experimental modality (light treatment and/or chemical exogenous supply), with three biological replicates, except otherwise stated.

The CO₂ net assimilation rate was measured as the quantity of CO₂ assimilated per leaf surface unit per second, and represents the difference between photosynthetic assimilation and respiration. To assess the effect of the light treatments, CO₂ net assimilation rates were measured both in a leaf still rapidly expanding and partially unfolded at the FBV stage (leaf 4 from the apex), and in a mature fully unfolded leaf (leaf 6 from the apex), on five plants, 3 h before and 3 h after the switch in light intensity, and then once a day for 14 days, 6 h after the beginning of the light period.

To test if the CK supply or the CK – ABA supply affected photosynthesis in plants grown under low PPFD, CO₂ net assimilation rates were measured in leaves of plants supplied according to the following conditions: CK 500 μM + NaOH 40 mM, CK 500 μM + ABA 1.25, 2.5, 5.0 mM + NaOH 40 mM, control NaOH 40 mM. CO₂ net assimilation rates were measured in the leaves positioned immediately above and below the supply point (leaves 4 and 5, respectively, Supplementary Materials 2A,B) on four plants per condition, every 2 days from FBV onward, 6 h after the beginning of the light period. Leaf positions were chosen so as to maximize the likelihood to observe an effect of the exogenous supply if it existed.

All measurements of CO₂ net assimilation rates were performed using an infrared gas analyser, Li-Cor-6400 (Li-Cor Inc., Lincoln, NE, United States), at a temperature of 18°C, a relative humidity of 64%, a cuvette air flow rate of 300 ml min^-1, and an ambient CO₂ concentration of 400 μmol mol^-1. PPFDs of 540 and 90 μmol m^-2 s^-1 were provided to the measured leaves for the plants grown under high and low PPFD, respectively, by a mixture of red and blue LEDs.

Statistical analyses were performed using R software for Windows (R Core Team, 2014). Groups were compared using one-way ANOVA (α = 0.05) followed by Tukey’s test when the ANOVA indicated significant differences at P < 0.05 and there were more than two groups.

The number of outgrown buds per plant was much lower under the low PPFD treatment than under the high PPFD treatment (Figure 1A-inset and Supplementary Figure S1). Under high PPFD buds grew out all along the stem, displaying an acrotonic pattern with a decreasing outgrowth rate from the apical to the basal part of the stem (Figures 1A,B). The time course of bud outgrowth was sequential, starting from the two uppermost apical buds 2 days after the FBV stage and proceeding downward (Figure 1C). The uppermost two apical buds did not respond to a reduction of PPFD because they are sylleptic buds whose outgrowth is independent of environmental conditions (Huché-Thélier et al., 2011; Demotes-Mainard et al., 2013b). But, the outgrowth rate of the third topmost bud was reduced as compared to high PPFD conditions (72% vs. 100%), and outgrowth of all the buds from the fourth topmost bud and below was totally inhibited (Figures 1A,B). Bud 4 displayed the most contrasted outgrowth behavior between PPFD treatments. While under high PPFD all buds 4 had grown out 8 days after the FBV stage on average, none grew out in low PPFD (Figures 1A,C). Elongation of bud 4 was rapid and continuous under high PPFD, whereas it was slow and stopped rapidly under low PPFD (Figure 1D).

To determine if sugars are involved in mediating the effect of PPFD on bud outgrowth, we analyzed the sugar content of internode 4, which is adjacent to the bud displaying the most contrasted response to PPFD treatments (Figures 1A,B). Internode 4 displayed a lower content in sucrose, glucose, fructose, and starch under low PPFD than under high PPFD days 4 and 8 after the FBV stage (Figure 2). Day 8 was the time when bud 4 grew out under high PPFD (Figure 1C), and day 4 was the time when we presume that conditions may have affected the inducement of bud 4 outgrowth since 4 days are required for bud outgrowth after apical dominance has been relieved under favorable light conditions in this cultivar (Girault et al., 2010). The correlation between sugar content in internode 4 and bud 4 outgrowth thus suggests that a low sugar level in the node may be involved in bud outgrowth inhibition under low PPFD.

To investigate this hypothesis, we tested if local sugar supply could relieve bud inhibition under reduced PPFD. Sugars were supplied to plants either through a cotton wick (50, 100, and 200 mM) or through the rachis for the higher concentrations (300, 600, and 800 mM). Using labeled ¹³C-sucrose, we demonstrated that sugars were indeed delivered into the stem tissues by these two methods (Supplementary Figures S2A,B). Sucrose supplied under low PPFD did not promote bud outgrowth, nor did it change the gradient of bud outgrowth along the stem as compared to the control, whatever the concentration (Figures 3A,A-inset,D,D-inset). Neither did it stimulate bud elongation (Figures 3B,E) or leaf formation into the bud (Figures 3C,F). Similar results were observed when glucose (100 or 800 mM) instead of sucrose was supplied (Supplementary Figures S3A–F). Altogether, these results show that although low PPFD leads to reduced sugar levels in the internode, the bud outgrowth inhibition induced by low PPFD cannot be solely relieved by local sugar supply. We therefore investigated whether CKs and ABA may limit bud outgrowth under low PPFD.

Cytokinin (zeatin riboside-5′-monophosphate, ZRMP; isopentenyl adenosine-5′-monophosphate, iPRMP; isopentenyl adenine riboside, iPR) contents were measured in internode 4 under high and low PPFD treatments 4 and 8 days after the FBV stage. ZRMP and iPRMP are intermediate forms in the CK biosynthesis pathway, and iPR is an active form. The three CK forms followed different temporal dynamics under high PPFD: the content in ZRMP increased over the 8 days following the FBV stage, whereas the content in iPRMP was stable, and the content in iPR decreased over time (Figures 4A–C). Low PPFD reduced the content in all three CK forms in internode 4, 4 and 8 days after the FBV stage (Figures 4A–C), whereas the free IAA content in internode 4 was not affected by the PPFD level (Figure 4D).

In order to assess if CKs may be players in the regulation of the bud outgrowth pattern by PPFD, we tested if exogenous CKs could relieve the bud outgrowth inhibition induced by low PPFD. Synthetic CK BAP was supplied to internode 4 of plants grown under low PPFD using cotton wicks. When 50 and 500 μM CK were supplied, the number of buds that grew out per plant increased significantly as compared to the control (Figure 5A-inset). Increased outgrowth was observed locally close to the supply point for bud 4, but also at a distance for bud 3 (Figure 5A). Bud 4, whose outgrowth was totally inhibited under low PPFD in the absence of CK, displayed rates of 91 and 100% outgrowth when the plants were supplied with 50 and 500 μM of CKs, respectively, under low PPFD. Under low PPFD again, outgrowth of bud 4 was observed 7.78 ± 0.44 and 7.49 ± 0.39 days after the FBV stage in plants supplied with CKs at 50 and 500 μM, respectively, and bud 4 elongation was increased by exogenous CKs (Figure 5B), giving a similar date of outgrowth and dynamics of elongation as observed under high PPFD (Figures 1C,D). CK supply under low PPFD also increased the number of leaves formed into the bud (Figure 5C). Taken together, these results suggest that low PPFD inhibition of bud outgrowth is mediated at least partly through a lower stem CK content.

Measurements of free ABA contents in the adjacent internode to bud 4 revealed that inhibition of bud outgrowth by low PPFD was also correlated to the maintenance of higher levels of ABA 4 and 8 days after the FBV stage (Figure 6).

To assess whether ABA could be involved in mediating PPFD control of bud outgrowth, we tested whether exogenous supply of ABA (0, 1.25, and 2.5 mM) to internode 4 could inhibit bud outgrowth under high PPFD (350 μmol m^-2 s^-1). ABA supply at both 1.25 and 2.5 mM had a strong inhibitory effect on the outgrowth and elongation of bud 4 (Figures 7B,C) but not on the number of leaves formed per bud (Figure 7D). Interestingly, outgrowth of buds located below the point of ABA supply increased (Figure 7A), so that the total number of outgrown buds per plant was not affected (Figure 7A-inset).

To test whether a high ABA stem content could inhibit bud outgrowth under low PPFD, ABA was supplied under low PPFD together with CK, a promoter of outgrowth. ABA antagonized the stimulating effect of CK in a dose-dependent manner and decreased bud 4 outgrowth (Figures 8A,A-inset). With the lowest ABA concentration, only 46% of these buds grew out and their outgrowth was delayed by 2 days, occurring only 10.2 ± 0.4 days after FBV, whereas with the highest ABA concentration only 4.7% of buds 4 grew out. In the presence of CKs, ABA also drastically limited the elongation of bud 4 (Figure 8B). With the highest ABA concentration supplied together with CK, the frequencies of bud 4 outgrowth and the kinetics of bud 4 elongation were very similar to those observed under low PPFD without any exogenous supply (Figures 1, 8). ABA had no effect on the number of leaves formed into the bud (Figure 8C). The buds below the supply point did not grow out, whatever the ABA concentrations (Figure 8A).

Altogether, these results indicate that low PPFD maintains high ABA levels in the stem and that stem ABA can inhibit bud outgrowth under low and high PPFD in rose. They suggest that the high ABA stem contents induced by low PPFD may contribute to the inhibition of bud outgrowth, notably by antagonizing the promoting effect of CKs on bud outgrowth.

The decrease in PPFD from the FBV stage onward markedly and durably reduced photosynthetic CO₂ net assimilation rates as compared to high PPFD, both in mature and in still rapidly expanding leaves (Supplementary Figure S4). The literature indicates that ABA may dampen photosynthesis (Raschke and Hedrich, 1985; Zhou et al., 2006), and that CKs alone or CKs and ABA together may affect photosynthesis in a way that varies with the species, the conditions of application and the plant growth conditions (Pospíšilová, 2003; Pospíšilová and Bat’ková, 2004; Shu-Qing et al., 2004). We studied whether the effect of CKs and ABA on bud outgrowth passed through an indirect effect on photosynthetic assimilation in our conditions. Over the 2 weeks following the FBV stage, CO₂ net assimilation rates of leaves 4 and 5 of plants grown under low PPFD were affected neither by CK supply alone nor by combined CK and ABA supply, whatever the ABA concentration (Figures 9A,B). These results indicate that the effects of CKs and ABA on the bud outgrowth pattern under low PPFD are not directly mediated by changes in photosynthetic assimilation.

The reduced PPFD applied from the FBV stage onward reduced the number of outgrowing buds, consistently with previous results in rose (Bredmose, 1995; Furet et al., 2014; Wubs et al., 2014). To address the question of how the CK, ABA, and sugar regulation pathways take part in the control of the branching pattern by light intensity, we focused on bud 4 that displayed the most contrasted fate with 100% or 0% outgrowth under high or low PPFD, respectively. Changes in PPFD led to variations in endogenous CK, ABA, and sugar contents in internode 4, consistent with bud outgrowth response to PPFD: CKs and sugars, which promote bud outgrowth, had their levels decreased under low PPFD, while ABA, an outgrowth inhibitor, had its content increased. The effect of PPFD on the CK stem content was consistent with the effect observed following exposure to light or darkness by Roman et al. (2016), but information on the effects of PPFD on the ABA stem content was lacking. The reduction in sugar stem content was consistent with the severe decrease in photosynthesis observed under low PPFD but may also have involved changes in partitioning. In addition, the variations in endogenous CK, ABA, and sugar contents in internode 4 preceded the outgrowth of the adjacent bud under high PPFD by at least 4 days; this supports the hypothesis that these chemicals could indeed contribute to the regulation of bud outgrowth. In rose, modifications in sugars and hormone (CK) levels and in expression of genes involved in bud outgrowth regulation were reported to be triggered before the occurrence of bud outgrowth (Girault et al., 2010; Henry et al., 2011; Rabot et al., 2012; Barbier et al., 2015; Roman et al., 2016).

Despite the major role of CKs in promoting bud outgrowth (Dun et al., 2012; Müller et al., 2015), studies investigating their involvement in the regulation of bud outgrowth by light are scarce. Stem endogenous CKs dropped under low PPFD concomitantly to bud inhibition and conversely CK supply to plants grown under low PPFD relieved the inhibition of bud outgrowth. CK supply stimulated bud meristem organogenic activity and bud elongation, that are the two fundamental processes of bud outgrowth. This stimulation happened in such a way that CK supply under low PPFD resulted in the same timings of bud 4 outgrowth and the same kinetics of bud 4 elongation as those observed under high PPFD. Altogether, these results are consistent with our previous findings showing that CKs are initial targets of light in the regulation of bud outgrowth in decapitated defoliated plants (Roman et al., 2016). They further bring evidence that (i) CKs allow buds to be relieved from ecodormancy due to unfavorable environmental conditions (low PPFD) and from paradormancy (apical dominance and other correlative inhibitions, Zieslin and Halevy, 1976; Domagalska and Leyser, 2011); (ii) a reduction of CK levels not only mediates repression of bud outgrowth induced by total darkness but also by low PPFD, as can be encountered under normal daylight conditions.

The stimulating effect of CKs on bud meristem organogenic activity under low PPFD is consistent with the ability of CKs to rescue organogenic activity under darkness in shoot apical meristems of tomato and rose (Yoshida et al., 2011; Roman et al., 2016).

Knowing that stem IAA can downregulate stem CK contents (Tanaka et al., 2006; Shimizu-Sato et al., 2009), we wondered if the changes in stem CKs in response to PPFD may have resulted, at least partly, from a change in stem IAA. Knowledge on the effects of PPFD on stem auxin contents is poor, and the results obtained in various species and growth conditions are conflicting (Kurepin et al., 2006, 2007, 2008; Hersch et al., 2014). Here, the absence of difference in IAA contents in internode 4 between the two PPFD treatments indicates that the changes in stem CK content was not a consequence of changes in stem IAA levels. Roman et al. (2016) also observed changes in CK stem contents in response to light in the absence of variations in IAA content, but the stem IAA levels in their plant system were extremely low due to decapitation.

Both stem-synthesized CKs and root-derived CKs can contribute to bud outgrowth regulation, with conflicting results between species (Faiss et al., 1997; Tanaka et al., 2006; Müller et al., 2015). In our study we can assume that the regulation of the CK stem content by PPFD may result at least partly from local regulation of the CK metabolism in the stem, since CK synthesis genes are downregulated and CK degradation genes are upregulated in the stem of plants grown in darkness as compared to light (Roman et al., 2016). Regulation of the CK stem content may also partly result from changes in xylem-carried CKs coming from the roots, since in tobacco plants shade reduces xylem carried CKs by reducing the transpiration rate (Boonman et al., 2007).

Our results suggest that the high stem ABA content induced by low PPFD contributed to the regulation of bud outgrowth. Exogenous ABA supply to the stem inhibited bud outgrowth in rose under high PPFD, i.e., in environmental conditions favorable to outgrowth, in accordance with previous data in Arabidopsis, pea, Ipomoea nil, and tomato (Arney and Mitchell, 1969; Chatfield et al., 2000; Cline and Oh, 2006; Yao and Finlayson, 2015). To investigate further the role of ABA in mediating bud outgrowth regulation by PPFD, we showed that ABA could also inhibit bud outgrowth under low PPFD. ABA was able to antagonize the promotive effect of exogenous CKs on bud outgrowth in a dose-dependent manner under low PPFD. The antagonistic role of ABA and CKs had been evidenced in other processes such as stomatal regulation, seed germination or seedling development (Dodd, 2005; Wang et al., 2011; Guan et al., 2014), but such a dose-dependent antagonism was evidenced in this study for the first time in the regulation of bud outgrowth. Altogether, our results suggest that the effect of PPFD on bud outgrowth may be complex and that reduced PPFD inhibits bud outgrowth both by reducing the stem content in CKs (a branching inducer) and by maintaining a high content in stem ABA (a branching repressor). The high ABA content can dampen the effects of the remaining CKs. ABA was already known to mediate the response to the R:FR ratio or the photoperiod in Arabidopsis (Reddy et al., 2013; Yao and Finlayson, 2015; González-Grandío et al., 2017; Holalu and Finlayson, 2017). Our results extend its role to the mediation of environmental regulation of bud outgrowth by another light signal and species. In addition, they also point out the contribution of stem ABA in controlling bud outgrowth, and complement the previous works that focused on bud ABA.

Interestingly, ABA inhibited bud outgrowth and impaired bud elongation, but did not affect meristem organogenic activity in the bud, irrespective of the PPFD level. In contrast, CKs stimulated bud meristem organogenic activity, which indicates that the antagonistic effect of ABA and CKs was limited to certain bud outgrowth-related processes.

Under low PPFD, exogenous supply of both sucrose and glucose close to the bud failed to promote bud outgrowth in intact plants. These results suggest that sugars are not the main limiting factor for bud outgrowth in intact plants under low PPFD, but would act downward other factors, at least CKs and ABA. These results are consistent with those obtained in the upper bud of decapitated defoliated plants grown under darkness (Roman et al., 2016).

Cytokinins have positive effects on source-sink organ activities (Cowan et al., 2005; Peleg et al., 2011). In our case the CK supply that promoted outgrowth under low PPFD is unlikely to have acted on the global sugar content per plant as it did not change photosynthetic assimilation. CK supply to the stem is known to increase the sugar sink strength of the bud (Roman et al., 2016), in agreement with CK-mediated upregulation of sugar sink strength in other tissues (Roitsch and González, 2004). Thus, we can hypothesize that one action of CK supply in the vicinity of the bud under low PPFD was to increase bud sugar sink strength and allow buds to better compete for sugars at the expense of other plant organs. Unlike CKs, ABA affects negatively sink strength by limiting the availability of nutrients and energy (Garciarrubio et al., 1997; Parish et al., 2012). It would be of great interest to investigate in future works if ABA negatively affects the sink strength of buds and how it interacts with CKs in response to PPFD. Mason et al. (2014) demonstrated that sugar supply alone can promote bud outgrowth when plants are grown in a non-limiting light environment in which the shoot tip’s high demand for sugars is the limiting factor for bud access to sugars. In our experiments the apical organs, that were in rapid expansion (Demotes-Mainard et al., 2013a), most probably intensively competed with buds for sugars too. However, sugar supply to the stem could not trigger outgrowth under low PPFD because PPFD controlled bud outgrowth through an upstream mechanism, involving CK and presumably ABA regulation. We hypothesize that at least part of this control is due to low stem CK content that restricts sugar import into the bud under low PPFD.

Exogenous CK supply had an effect on the directly treated bud (bud 4) and on the first bud located above it (bud 3). The absence of effect on the buds located below the supply point is consistent with results on isolated nodes and decapitated plants (Ohkawa, 1984; Chatfield et al., 2000).

Abscisic acid supply in the stem inhibited only the local bud. We did not expect exogenous ABA supply to inhibit lower buds since ABA inhibits outgrowth in isolated nodes only when supplied from the basal side of the stem (Chatfield et al., 2000). The absence of inhibition of bud 3, located apically, may result from the dose of ABA that reached bud 3 and/or from a lower sensitivity of this bud to ABA due to its physiological state. Bud 3 was indeed much less sensitive to PPFD than the buds below, as shown by the outgrowth patterns under low and high PPFD.

When exogenous ABA totally inhibited bud 4 outgrowth under high PPFD, the outgrowth rate of the buds just below it (buds 5 and 6) increased. This increase in outgrowth may be an indirect effect of bud 4 inhibition under high PPFD: bud 4 inhibition by ABA may have both reduced the correlative inhibition exerted on lower buds (Domagalska and Leyser, 2011) and made sugars more available for these lower buds, thereby facilitating their outgrowth under high PPFD (Mason et al., 2014). In isolated internodes of Arabidopsis, ABA supplied from the apical side of the stem reduced inhibition by IAA (Chatfield et al., 2000). Such an effect cannot be ruled out but was not observed in Ipomoea nil or in sunflower (Cline and Oh, 2006).