Average daily temperature controls floral bud formation rate, callose deposition and flower development of Hydrangea macrophylla ‘Early Blue’

ABSTRACT Hydrangea macrophylla ‘Early Blue’ was exposed to six different day (DT) and night (NT) and average daily temperatures (ADT) during flower bud formation (DT/NT°C: 12/12, 17/12, 17/17, 22/12, 22/17, 22/22). The aim was to evaluate the effect of ADT and DT/NT on the rate of bud formation and forcing time after a period of chilling. Furthermore, the role of callose in dormancy induction and release during cold storage was investigated. ADT ≤ 17°C accelerated flower bud formation irrespective of DT and NT compared to an ADT > 17°C. Floral buds developed at ADT 17°C contained significantly lower level of callose compared to buds from the higher ADT regimes (ADT >17°C). After 6 weeks of chilling (10°C) the callose level was decreased to a similar level irrespective of the DT/NT regime during floral bud formation and/or floral stage at the start of the chilling treatment. We conclude that callose may play an important role in the dormancy release of H. macrophylla. However, floral buds do not have to be fully developed to remove the callose in response to chilling and no clear relationship was found between the levels of callose after chilling and forcing time.


Introduction
The environmental factors that synchronise flowering to the season are specifically photoperiod, irradiance, low temperature giving a vernalisation effect, and ambient growth temperature (Bernier & Perilleux, 2005).These environmental factors are also applied to regulate flowering time in the production of ornamentals in controlled environments.In addition to plants with desirable morphological characteristics and predictable flowering, energy-efficient production methods and short production time are increasingly important in the ornamental greenhouse industry.
Hydrangea macrophylla (Thunb) Ser. is a popular deciduous shrub, widely used as an ornamental plant.Greenhouse production of H. macrophylla includes four phases: (1) vegetative growth, (2) the flower bud formation, (3) chilling to release the flower buds from dormancy and (4) forcing of flowers (Figure 1) (Bailey, 1989).Temperatures between 15°C and 19°C are known to promote generative growth and temperatures above 20-23°C suppress flower bud formation and promote vegetative growth (Litlere & Strømme, 1975;Nordli et al., 2011).Plant morphology is mainly controlled by temperature and light, and to prevent excessive shoot height is an important goal in the production of H. macrophylla (Terfa & Torre, 2019).
Energy is the largest cost in greenhouse production of ornamentals in temperate climates.To supply less heating and let the climate in the greenhouses depend more on the outdoor climate is a strategy to save energy (Aaslyng et al., 2003;Blanchard & Runkle, 2011).The consequence is higher day temperatures (DT), lower night temperatures (NT), bigger differences in DT and NT, and fluctuations from year to year.Hence, knowledge on how daily temperature alternations affect flowering time is important.Studies regarding flowering and temperature responses of H. macrophylla have mainly been carried out with constant temperatures and limited information is available on the effects of day (DT) and night temperatures (NT) (Bailey, 1989).Plants discriminate between DT and NT, a process referred to as thermoperiodism (Went, 1944).Diurnal variations in temperature can cause significant changes in growth, morphology and flowering time.Typically, a higher DT in combination with a low NT results in elongated flowering pot plants, with longer internodes than constant temperatures or a higher NT than DT (Myster & Moe, 1995).It is also well known that NT is particularly important for flower induction in many species.When exposed to day/night temperature cycles, flowering time of different plant species can be accelerated or delayed relative to constant-temperature conditions depending on their photoperiodic response (Karsai et al., 2008;Thingnaes et al., 2003;Wendell et al., 2017;Yin et al., 1996).
H. macrophylla needs a period of chilling to release the flower buds from dormancy (Anderson et al., 2009;Bailey, 1989).It is common to place the plants at low temperatures (5-10°C) for 6-8 weeks in darkness to break the dormancy before forcing to flower (Bailey, 1989).Floral meristem stages of H. macrophylla are commonly categorised according to a scale from 0 (vegetative) and 7 (fully developed inflorescence) described in Galopin, et al. (Galopin et al., 2008) and Litlere and Strømme (Litlere & Strømme, 1975).Such information is used to decide when to start the chilling treatment.To fulfill the chilling, it is claimed that the floral stigmas must have reached the gynoecium or 'G' stage of development (Struckmeyer, 1950), which is stage 4 according to Galopin et al. (Galopin et al., 2008) and Litlere and Strømme (Litlere & Strømme, 1975).In spite of this, there are still problems with delayed flowering and development of vegetative shoots during forcing due to incomplete flower bud development, and/or floral abortion.In the study of Nordli et al. (Nordli et al., 2011) more vegetative shoots were observed in plants exposed to ADT > 17°C during bud formation although the floral stage was > 4 when the chilling started.It can be questioned if the depth of dormancy may vary depending on the conditions during bud formation.It is well known from vegetative buds in a number of woody species that the environment during dormancy induction has an influence on the depth of dormancy (Olsen, 2010;Tanino et al., 2010).Callose (1,3-β-D-glucan) blocks plasmodesmata in dormant buds of woody species and is degraded during dormancy release (Lee et al., 2017b;Rinne, Welling, Vahala, Ripel, Ruonala, Kangasjarvi, et al., 2011).It has been used as a marker of the depth of dormancy in monitoring dormancy release in vegetative buds in Vitis vinifera L (Aloni & Peterson, 1997;Aloni et al., 1991).It may thus be hypothesised that the callose content in flower buds of H. macrophylla depends on the floral stage and temperature regime during the flower bud formation, and that the callose content influence on the forcing time.
The main objective of this study was to investigate how the average daily temperature (ADT) and different DT and NT regimes during bud formation affect the rate of flower differentiation and development after a period of chilling.To study the depth of dormancy after exposing buds to different DT/NT regimes and to test the hypothesis that callose content is important for forcing time, we measured the callose content in the floral buds before and after chilling.

Plant material and pre-cultivation
Rooted cuttings of Hydrangea macrophylla (Thunb) Ser.'Early Blue' were delivered from Tretteteig nursery (Vestfold, Norway).The rooted cuttings, in a vegetative stage (stage 0, according to Litlere andStrømme 1975 (Litlere, 1975)), were re-planted in 13 cm pots with standard fertilised peat (Floralux, Nittedal torvindustri A/S, Norway).The plants were pinched above 3-4 leaves and pre-cultivated in a growth chamber at The Norwegian University of Life Sciences (NMBU) for five weeks before floral bud formation treatments started.The temperature during pre-cultivation was maintained at 24°C ±0.5°C to ensure vegetative growth and the relative air humidity (RH) was 70% ± 5%.During the precultivation the plants were irrigated with nonfertilised tap water.The light was provided by high pressure sodium lamps (HPS, Osram NAVT-400 W, Munich, Germany) 16 h daily at a photosynthetic photon flux density (PPFD) of 150 (±10) µmol m −2 s −1 (measured with a Li-Cor, Model L1-185, quantum sensor, Li-Cor, Lincoln, NE, USA).

Experimental set-up during bud formation
The plants had shoots with 2-3 leaf pairs before the floral bud formation treatments started.Three growth chambers were used in the experiments and the temperature set-points were 12°C, 17°C, and 22°C (±0.5°C) respectively.The plants were exposed to 16 h lighting provided by HPS lamps (Osram NAVT-400 W).During bud formation, plants were moved twice daily in order to be exposed to six different day (DT) and night (NT) temperatures and average daily temperatures (ADT).ADT was calculated as follows: DT °C × 16 h + NT°C × 8 h)/24 h) The floral bud treatments lasted for 9 weeks as described in Table 1.

Experimental set-up during chilling and forcing and recordings of plant growth and flowering
After 9 weeks of floral bud formation the plants were placed in a cold storage chamber at 10°C ±0.5°C for 6 weeks in darkness.The plants were not defoliated but the plants abscised most of their leaves during the chilling.After the chilling, plants were moved to a growth chamber with 21°C ±0.5°C and 16 hr photoperiod with a PPFD of 150 μmol m −2 s −1 provided by HPS lamps (as described above).A group of unchilled plants (n = 5) was placed directly at 21°C.Time (days) to macroscopic visible bud (evaluated by naked eye) as well as time to visible bud >1 cm were recorded daily.The plants were harvested at marketing stage, when at least two of the inflorescences were developed.At this stage plant height was measured from the base of the shoot until the middle of the highest inflorescence.The total number of internodes on the main shoots were counted and the number of flowering shoots (<10 cm) were recorded on 10 plants per treatment.

Microscopy of floral meristems
Vegetative samples of random shoot apical meristems (SAM) were collected before the DT/NT treatments started and dissected in a binocular to confirm their vegetative stage 0 according to Litlere (Litlere, 1975).Samples of terminal SAM were then collected 3, 6 and 9 weeks after start of the floral bud formation phase and after 6 weeks of chilling (see Figure 1).The buds were put in fixative containing 2% paraformaldehyde and 1.25% glutaraldehyde in 50 mmol/L L-piperazine-N-N`-bis(2-ethane sulphonic) acid buffer (pH 7.2).The samples were stored at 4°C before dissecting.The floral stage of the fixed dissected samples was determined using a scale from 0 to 7 according to (Litlere & Strømme, 1975) and Galopin, Codarin, Viemont and Morel (Galopin et al., 2008) by using a binocular loupe.For anatomical/histological studies plant material from three plants per treatment were embedded in LR White resin (London Resin Company, England) according to (Lee et al., 2017a).Briefly, the fixed plant materials were dehydrated in a graded ethanol series and then gradually infiltrated in the LR White resin.Thereafter the plant materials were embedded in the LR White resin by polymerisation at 60°C for 12 h.The materials were then longitudinally sectioned into 1 µm-thick sections using an Ultracut microtome (Leica EM UC6, Germany), stained with toluidine blue O (Sigma -Aldrich, USA) and examined using a light microscope (Leica DM6B, Germany). .Immunolabeling was done using mouse β-1,3-D-glucan antibody (Biosupplies, Australia) and anti-mouse IgG coupled to 10 nm gold particles (Sigma -Aldrich) as the primary and the secondary antibodies, respectively.The sections were incubated with 3% bovine serum albumin (BSA)/PBS, treated for 1 h at 20°C with the primary glucan antibody and then incubated for 1 h at 20°C in the secondary antibody solution containing anti-mouse IgG coupled to 10 nm gold particles (Sigma -Aldrich).The sections were then rinsed with PBS and stained with uranyl acetate and lead citrate before examining using a TEM (FEI Morgagni 268, the Netherlands).The average number of gold particles were quantified in 100 individual plasmodesmata from three different plants per treatment.

Statistics
Statistical analysis was conducted using one-way ANOVA on Minitab software package (version 19, Minitab Inc., State College, PA, USA).Means were separated using Post-hoc Tukey's test at the 5% level of significance.A Pearson correlation was used to measure the degree of relationship between callose content in the buds after 9 weeks of bud formation and average daily temperature.The growth experiment was repeated twice with similar results, but the microscopy work was only done on plant material collected from the first repeat.

Effect of ADT, DT/NT on floral bud development
After 3 weeks, no significant differences were found between the different ADTs, DTs or NTs (Table 2).However, after 6 weeks the buds developed at ADT ≤ 17°C showed accelerated floral bud development compared to buds developed at ADT > 17°C (Table 2, Figure 2).Within the two groups (ADT  ≤17°C and ADT > 17°C), no significant effect of NT was found (p > 0.05).The buds developed at DT/NT°C 17/12 and 17/17 reached floral stage ≈ 4 after 6 weeks and the DT/NT°C 12/12 reached stage ≈ 6 after 9 weeks (Table 2).The temperature treatments with an ADT > 17°C (DT/NT°C 22/12 and 22/17) reached floral stages ≈ 2 and ≈ 3 after 6 and 9 weeks, respectively (Table 2, Figure 2).Buds from more or less inductive temperature treatments (DT/NT°C 17/17, 17/12, 22/12 and 22/17) were dissected and studied further for floral stages after 6 weeks of chilling.The stages of the floral buds were not significantly affected by the chilling (p > 0.05) which indicates that no visible morphological changes occur in the floral meristems during chilling (Table 4).

Plant morphology and time to visible floral buds after a period of chilling
Total plant height and internode number were significantly higher in plants exposed to ADT > 17°C compared with ADT ≤ 17°C during floral bud formation but no significant differences were found between the different DT/NT (Table 3).Increasing ADT resulted in reduced number of lateral shoots, but only data from DT/NT°C 22/22 and 12/12 was significantly different (p < 0.05).After six weeks of chilling the plants were forced to flower at 21°C.Time to visible bud was affected by the temperature treatments during bud formation (Figure 3).The fastest time to visible bud was observed in plants treated with DT/NT 17/17°C.

Forcing time and callose content in floral buds
Plants not exposed to chilling did not develop flowers (data not shown).However, after 20 days of forcing at 21°C 100% of the plants exposed to 17/17°C showed visible flower buds (>1 cm) compared to 90% of the plants from 17/12°C and about 30% and 20% of the plants from DT/NT 22/12°C and 22/17°C respectively (Figure 3).Time to anthesis was affected similarly (  H. macrophylla 'Early Blue' requires chilling for dormancy release.When plants were placed directly to forcing without chilling they did not flower (data not shown).The callose level was evaluated in floral buds right before and after 6 weeks of chilling (Figure 1).A correlation was found between callose content and ADT during bud formation (Figure 5, Pearson correlation: 0.977), and more callose was observed with increasing temperature.Furthermore, significantly higher callose content was found at DT/NT 22/17 and 22/12 than at 17/17°C and 17/12°C, but for the temperature regimes with the same day temperature, the night temperature did not affect the callose level significantly (Figure 5).After 6 weeks of chilling, the callose content was decreased significantly in all buds irrespective of floral stage and temperature treatment during bud formation, despite the significantly different values, no clear trend was found, which could be related to temperature regime or to callose content before chilling (Figures 4 and 5).

Discussion
Basic knowledge of the flowering process, and better tools to regulate it are necessary to control flowering time of ornamentals more accurately.The present study demonstrates that floral bud formation rate (floral stage divided by numbers of weeks, Table 2) was mainly controlled by ADT.ADT ≤ 17°C is a threshold temperature to ensure fast floral bud formation in H. macrophylla (Table 2) and short production time (Tables 3-4, Figure 2-3), as shown in earlier studies (Nordli et al., 2011).Buds exposed to temperatures ≤ 17°C were fully developed after 9 weeks (stage 6) and buds exposed to temperatures > 17°C reached stage 3 after 9 weeks (Figure 2) (Galopin et al., 2008).Similar temperature responses are evident in many horticultural subtropical and tropical woody species which often require temperatures around 15-20°C to flower (Wilkie et al., 2008).The results also showed that NT does not affect the floral bud formation rate significantly (Table 3).However, a significantly accelerated forcing time was found after a period of chilling in plants exposed to increased NT (Table 4, Figure 3).Reducing the production time is very important for the production efficiency and the possible energy savings in the greenhouse production of ornamentals.Hence, the bud formation should take place at temperature regimes that accelerate flower development during forcing to reduce energy consumption without reducing the quality, and in this study DT/NT17/17°C seems to be the optimal temperature regime.
Most Hydrangea cultivars usually undergo a period of dormancy between floral initiation and anthesis and chilling is required to break dormancy (Bailey, 1989).Plants were exposed to chilling for 6 weeks before forcing.However, a few plants (n = 5) from the different DT/NT regimes were placed directly into the greenhouse for forcing at 21°C but none of the plants developed flowers (data not shown).Unchilled buds of H. macrophylla were characterised by heavy deposits of callose blocking the plasmodesmata (Figure 4).However, when the buds were exposed to chilling, a strong reduction in callose content was found in all buds, confirming breaking of dormancy according to previous studies of woody species (Rinne, Welling, Vahala, Ripel, Ruonala, Kangasjärvi, et al., 2011).The temperature during bud formation had an influence on the extent of callose deposition in the floral buds, and the lower the temperature, the less callose was observed (Figure 4).However, after 6 weeks of chilling, the callose level was very low in all buds irrespective of the floral bud stage and temperature regime during the bud formation.Even the early stage floral buds formed at temperatures > 17°C (stage 3) had very low callose level after six weeks of chilling.
If the callose content at the plasmodesmata in floral buds in H. macrophylla is related to the depth of dormancy, it may be hypothesised the callose content will influence the forcing time.If so, we should have seen a greater decline in callose after chilling in buds developed at DT/NT 17/17°C, which showed the fastest forcing time.However, this was not the case (Figure 5).We observed very low callose levels in all buds after a period of forcing, and there was even a slight trend that the highest level of callose was present in buds developed at DT/NT 17/17°C.In conclusion, the callose levels at the plasmodesmata could not explain why the forcing time was different in plants with similar floral bud stage, or DT/NT regime and is not a good indicator for forcing time.Although callose is thought to be important in regulating plasmodesmata trafficking, other factors may also be of importance, such as cytoskeleton components and remodulation of the contact sites between the membranes spanning the plasmodesmata (Nicolas et al., 2017;Tilsner et al., 2016).Furthermore, although not investigated in H. macrophylla, other barriers affecting connectivity between cells and tissues may influence the forcing time of buds through effect on the transport of water and solutes into the buds.Vegetative and floral buds of woody species such as Norway spruce (Picea abies) and peach (Prunus persica) were shown to have higher abundance of de-methyl-esterified homogalacturonan pectin and proteins in cell walls than growing shoot tips, resulting in decreased water and solute permeability due to reduced pore size in the cell wall matrix through Ca 2+ binding and crosslinking (Lee et al., 2017b;Wisniewski & Davis, 1995).The possible differential forcing time in H. macrophylla buds may be related to different extents of such dormancy-related bud-shoot barriers and thus different time to re-establish full connectivity.It has been claimed that the floral stigmas must have reached the gynoecium or 'G' stage of development (Struckmeyer, 1950), which is stage 4 according to Litlere (Litlere, 1975), to fulfill the chilling.However, our results showed that all buds irrespective of floral stages responded to chilling by callose removal (Figure 5).

Figure 1 .
Figure 1.Experimental set-up for Hydrangea macrophylla 'Early Blue', showing temperature regimes during consecutive growing phases in the different production phases (1-4), timepoints for collecting buds for callose measurements, and recordings of visible buds, flowering time and plant morphology.Plants were grown in controlled climate chambers and photosynthetic active radiation was provided by high pressure sodium lamps (150 µmol m −2 s −1 ) with a photoperiod of 16 h light/8 h darkness.

Figure 5 .
Figure 5. Callose content in the sections collected from H. macrophylla 'Early Blue' floral buds before (a) and after chilling (b).Data are expressed as number of gold particles.The buds were developed under different day (DT) and night temperatures (NT) for 9 weeks and the chilling lasted for 6 weeks at 10°C.The values are mean ± SE of 100 individual plasmodesmata images in sections from three different plants per treatment.Different letters indicate significant differences.

Table 2 .
Floral bud phenological stage (0-7) of Hydrangea macrophylla 'Early Blue', according to[23] and [15], 3, 6 and 9 weeks after start of different average daily temperatures (ADT) day (DT) and night temperatures (NT) at the start of the experiment all plants were at stage 0 (vegetative).Mean followed by different letters within columns are significantly different at p < 0.05 according to Tukey's test.N = 10.StDev in brackets.

Table 3 .
Effect of different day (DT) and nigh temperatures (NT) and average daily temperature (ADT) during bud formation on morphology and time to the stage of visible floral buds of H. macrophylla 'Early Blue'.The plants were chilled for 6 weeks at 10°C to break dormancy and forced at 21°C.Plants were grown in controlled climate chambers and photosynthetic active radiation was provided by high-pressure sodium lamps (150 µmol m −2 s −1 ) with a photoperiod of 16 h light/8 h darkness.The morphological measurements were taken at the end of the experiment when at least two inflorescences were developed.Mean followed by different letters within columns are significantly different at p < 0.05 according to Tukey's test.N = 10.StDev in brackets.

Table 4 .
Effect of different average daily temperature (ADT), day (DT) and nigh temperatures (NT) regimes during bud formation of H. macrophylla 'Early Blue' on number of days until flowering (marketing stage) and floral stage according toLitlere (1981)after 6 weeks of chilling (10°C).Plants were forced to flower at 21°C in controlled climate chambers and photosynthetic active radiation was provided by high pressure sodium lamps (150 µmol m −2 s −1 ) with a photoperiod of 16 h light/8 h darkness.(N = 10).Mean followed by different letters within a column are significantly different at p < 0.05 according to Tukey's test.n = 10.StDev in brackets.
Effect of different day (DT)/nigh temperature (NT) (°C) regimes during floral bud formation of H. macrophylla 'Early Blue' on time (days) to first macroscopically visible floral bud (>1 cm) during forcing.After the DT/NT treatments for 9 weeks, the plants were chilled for 6 weeks at 10°C to break the dormancy and forced at 21°C.Plants were grown in controlled climate chambers and photosynthetic active radiation was provided by high pressure sodium lamps (150 µmol m −2 s −1 ) with a photoperiod of 16 h light/ 8 h darkness.N = 10.