Plant material and growth conditions.

All experiments were conducted with cauliflower cultivar Somerset (Seedway, Hall, NY) and broccoli cultivar Barbados (Bejo Seeds, Oceano, CA), both early-maturing cultivars that grow well in the greenhouse and have little or no vernalization requirement. Seeds of cauliflower and broccoli were sown into modular trays of 50 cells (4.5 × 4.5 cm) containing Cornell Mix A (Boodley and Sheldrake, 1977) and placed in a greenhouse at day/night temperatures of 22 to 24/16 to 18 °C. Seedlings were transplanted into 2-L pots containing Cornell Mix A and 5 g 14N–6.1P–11.6K controlled-release fertilizer (Osmocote 14-14-14; Scotts, Marysville, OH), a slow-release fertilizer, and subsequently fertilized as required with 15N–2.2P–12.4K water-soluble fertilizer (EXCEL Cal-Mag 15-5-15; Scotts). Treatments were randomly assigned to pots on the greenhouse benches. Supplemental lighting provided 450 μmol·m⁻¹·s⁻² of photosynthetically active radiation for 14 h·d^–1. ‘Somerset’ and ‘Barbados’ plants were kept in the greenhouse during the whole growing period.

Determining reproductive stage.

GA applications had to occur at precisely known stages of meristem development, but that stage cannot be determined non-destructively. The exact date a particular stage is reached can vary by a few days with small variation in growing conditions. Therefore, an indirect approach was used to associate the external appearance with apical meristem maturation (Björkman and Pearson, 1998). The developmental stage of the broccoli apex is associated with distinctive changes in leaf shape and stem width. Plants with a meristem composed only of leaf primordia were classified as vegetative. They were classified as reproductive when the petioles of the two largest leaves were straight and the apical meristem had a central dome with a flat area in the axillary region of the leaf primordia. Plants were in a more advanced reproductive stage when they had evident paraclades and some floral primordia.

The reproductive transition and associated shoot tip changes can be detected non-destructively by a widening across the youngest exposed petioles. The width was ≈4 mm at the beginning of the adult vegetative stage in both the cauliflower and broccoli cultivars and ≈6 mm at the end of the adult vegetative period [≈40 d after sowing (DAS)] (Fig. 1). In addition, sacrificial plants of the same cohort were dissected at regular intervals. Five plants per cultivar were dissected under a stereomicrocope (SZ-40; Olympus, Tokyo, Japan) to determine progress and 10 plants to confirm the developmental state before applying GA to the population. The shoot apical meristem diameter increased from 0.1 to 0.25 mm during the adult vegetative stage (Fig. 1). When the apical meristem underwent the transition to the inflorescence meristem, the petioles were slightly straightened; the tip width increased to 9 mm in ‘Somerset’ cauliflower and 8 mm in ‘Barbados’ broccoli. When a small curd began to form, the petioles bowed and the tip width increased to 13 mm in cauliflower and 11 mm in broccoli.

View Figure

• Download as Powerpoint
• Download Figure

Shoot growth was measured two ways: as stem elongation (in centimeters) and node production rate (the number of nodes on which the leaves were longer than 5 cm). Nodes appeared at a consistent time interval in both cultivars. Leaf primordia were initiated before leaves were large enough to count, but the time could be inferred. In cauliflower, young leaves grew from initiation to 5 cm in ≈2 d at the adult vegetative stage, slowing to 3.5 d at the reproductive meristem stage (≈40 DAS). In broccoli, the corresponding growth rate was constant.

Determining endogenous GAs through GA extraction and purification by high-performance liquid chromatography (HPLC) and gas chromatography–mass spectrometry (GC-MS).

Six bioactive GAs (GA₁, GA₃, GA₄, GA₅, GA₆, and GA₇) are produced through two different biosynthetic pathways: the early-13-hydroxylation pathway and the non-13-hydroxylation pathway (Sponsela and Hedden, 2004). Previous studies have demonstrated that GA₄₊₇, but not GA₃, affects floral induction of cauliflower plants (Booij, 1989, 1990; Guo et al., 2004), suggesting that only GAs of the non-13 hydroxylation pathway are active in the developing meristem of the plant. When studying the effects of exogenous GAs on a plant, it is important to know which of these GAs are present in the developing meristem of the plant. Booij (1989, 1990) and Guo et al. (2004) showed that GA₄₊₇, but not GA₃, has an effect in floral induction of cauliflower plants, suggesting that, at this stage of the development, only GAs of the non-13 hydroxylation pathway are active. To determine whether GAs from the early-13 hydroxylation pathway are present in inflorescence meristem and floral bud stages, GAs were extracted from curd of cauliflower and head of broccoli and analyzed as methyl ester derivatives by GC-MS after fractionation using HPLC (Rosin et al., 2003). Internal standards were deuterated GAs (L. Mander, Canberra, Australia) of the following type and amount [in nanograms per gram fresh weight (FW)]: GA₁ = 1, GA₈ = 10, GA₁₉ = 10, GA₂₀ = 20, GA₄₄ = 5, GA₂₉ = 50, and GA₅₃ = 5. GAs for which no standards were available were identified from their mass spectra by matching their Kovats’ retention indices and ion masses obtained with the GC-MS with reference spectra from Gaskin and MacMillan (1991).

Exogenous GA effect on the vegetative-to-reproductive transition.

We conducted two experiments on ‘Somerset’ cauliflower and ‘Barbados’ broccoli to investigate the effect of exogenous GA on heading and growth indices, indicative of floral induction. We used a well-established method that assures penetration into the meristem tissue (Mandel et al., 1992; Moore, 1967). The first experiment examined only the effect of GA₃ (MC Biomedicals, Aurora, OH). One microliter of a solution containing 5 μg of GA₃, 10% (v/v) acetone, and 0.1% wetting agent (Tween 20; Sigma-Aldrich, St. Louis, MO) or 1 μL of the solution without the GA, as a control, were applied to the apical dome of each plant with a micropipette once daily for 2 d (consecutive) at the initiation of the adult vegetative stage, ≈30 DAS. At this stage ‘Barbados’ and ‘Somerset’ plants had five fully extended leaves and the diameter of the apical meristem was ≈0.2 mm. Each treatment was applied to 12 to 15 plants.

The second experiment compared the effects of both GA and GA₄₊₇ (PhytoTechnology Laboratories, Shawnee Mission, KS) on plants at a slightly more advanced phase of the adult vegetative stage, ≈38 DAS. Again, the application time was determined by inspecting the meristem, not by the calendar. The GAs were applied as indicated above using 2 μL of a solution containing 2 μg·μL⁻¹ of GA₃ or GA₄₊₇, 25% (v/v) ethanol, and 0.1% wetting agent. The larger apex allowed a larger volume to be used than in the first experiment. Again, the same volume of solution, without any GA, was used as a control. Applications were made once daily for 3 d (consecutive). This method of application provided a consistent dose of GA at the meristem, avoiding runoff, application to other plant parts, and any need for translocation within the plant. At this time, ‘Somerset’ cauliflower plants had six to seven fully expanded leaves and the diameter of the dissected apical meristem was 0.25 mm. ‘Barbados’ broccoli plants had between seven and eight fully extended leaves and an apical meristem diameter of 0.25 mm. Each treatment was applied to 12 plants.

In each experiment, the effect of the treatment on advancing heading and stem elongation was determined by measuring both temporal and developmental parameters: 1) curding time: the number of DAS when the curd was first visible (≈1 cm across); 2) node production: the number of nodes formed after GA application based on the increase in the number of leaves longer than 5 cm between GA application and curd visibility; and 3) stem elongation: the increase in the stem height (measured from the cotyledonary attachment zone to the apical meristem of the plant) between GA application and curd visibility.

Exogenous GA effect on the morphology of the curd.

A solution containing 0 or 2.1 μg·μL⁻¹ of GA₃ or GA₄₊₇ in a 25% (v/v) aqueous ethyl alcohol solution and 0.1% wetting agent was applied to the apex of cauliflower and broccoli plants in a volume of 2 μL once daily for 3 d (consecutive). Applications were made at two different stages of the reproductive development. One group of 10 plants from each cultivar was treated when the apical meristem became reproductive, as indicated by the petioles straightening, 43 d from sowing; a second group of 10 plants of both cultivars was treated later during inflorescence meristem proliferation but before floral bud development: 55 DAS in ‘Barbados’ and 50 DAS in ‘Somerset’. Measurements included those described in the previous section as well as a morphological examination to determine stage of arrest. The arrest stages were categorized into one of the following classes: cauliflower, composed of inflorescence meristem; intermediate curd, composed of small floral buds and floral primordia; and broccoli, composed of fully developed floral buds.

Exogenous GA effect on bolting.

A solution containing 0 or 2.1 μg·μL⁻¹ of GA₃ in a 25% (v/v) aqueous ethyl alcohol solution and 0.1% Tween 20 (wetting agent) was applied to a marked location on the curd of broccoli plants once daily for 3 d (consecutive) beginning when flower buds had just developed (heads ≈2 cm). The plants were photographed when the stem at the marked location elongated.

Statistical analysis of apex development.

In every experiment, cultivars were analyzed independently with the treatments assigned in a completely randomized design with individual plants as the experimental unit. Tests for normality and equal variances were run for each variable under study to determine if assumptions of normal distribution and equal variances were met for analysis of variance (ANOVA). When the data obtained did not support the normal distribution assumption, and no transformation was possible, a non-parametric test was performed: Wilcoxon (two samples) and Kruskal-Wallis (more than two samples). Given that in some experiments observations were unequal among treatments, ANOVA was performed using PROC MIXED for its ability to estimate Type III error that corrects for unbalanced design. Multiple comparisons of the means were done using Tukey’s Studentized range (honestly significantly different). Statistical analyses were performed using SAS (Version 9.1; SAS Institute, Cary, NC).

Gene expression of integrator and meristem-identity genes in cauliflower cv. Somerset.

Four genes that are indicative of the floral transition were monitored. The change in expression of the meristem-identity genes BoAP1-a and Bo-AP1-c, the meristem-identity gene/integrator gene BoLFY, and the integrator gene SOC1 were measured in cauliflower cv. Somerset plants. Twelve uniform curds arrested at the inflorescence meristem stage and with a diameter of ≈5 cm were selected. Four milliliters of a solution containing 0 or 100 μM of GA₃ in a 25% (v/v) aqueous ethyl alcohol solution and 0.1% wetting agent were applied to the entire surface of each curd with a brush. Tissue samples from two different areas of the curd were taken, using a fresh razor blade for each sample, at 0, 20 min, 1 h, 3 h, 6 h, and 24 h after the application. After isolation, the tissue was placed in a sterile tube, submerged in liquid nitrogen, and stored at –80 °C.

Total RNA of each sample was extracted using Trizol reagent (Invitrogen, Carlsbad, CA) following the protocol described by the manufacturer with some modifications (Duclos and Björkman, 2008). RNA reverse transcription was performed using the RETROscript kit (Ambion, Austin, TX) with random decamers in accordance with the manufacturer’s protocol. Gene expression was determined using SYBR Green quantitative real-time polymerase chain reaction (qRT-PCR) in the iCycler iQ Real-Time Detection System (Bio-Rad, Hercules, CA) as described by Duclos and Björkman (2008). Primer design and selection followed the guidelines outlined by Duclos and Björkman (2008); Table 1 lists the primers selected for this study. The gene expression measurements followed all 11 golden rules of qRT-PCR (Uvardi et al., 2008).

Table 1.Primer sequences and the length of amplicon used to assess meristem differentiation in broccoli and cauliflower.^z

View Fullscreen

Statistical analysis of differentially expressed genes was performed with ANOVA using PROC MIXED for repeated measurement analysis, because responses were taken sequentially on the same experimental unit, in this case each plant. Each plant was considered a random effect and the GA treatment and time fixed effects. Assumption of normal distribution and equal variances were tested in each case and transformation of the data used when necessary. The statistical model developed for both genes analyzed (BoLFY and SOC1) used a compound symmetry structure. Selection followed the guidelines suggested by Littell et al. (1998) and Wolfinger and Chang (1998).

To estimate the assay variation, ses were calculated assuming statistical independence using propagation of error formulas that allow the inclusion of experimentally determined errors for each efficiency and threshold cycle (C_t) value (Ku, 1966).

Endogenous GAs in developing inflorescence.

GAs were extracted from cauliflower curds and broccoli heads and analyzed as methyl-ester derivatives by GC-MS with deuterated internal standards. GC-MS analysis determined that methyl extracts of cauliflower contained GA₈, GA₂₉, GA₁, GA₂₀, and GA_19, and broccoli extracts contained GA₈, GA₂₉, GA₁, and GA₂₀. All of the endogenous GAs identified using standards were from the early 13-hydroxylation pathway; GA₄₄ was not detected in cauliflower or broccoli, although the GA₄₄ standard was detected. The GA₅₃ standard was never observed. GA₅₃ is a precursor of GA₁₉ in the early pathway. Additional GAs were identified in both cauliflower and broccoli heads by comparison with Kovats’ retention indices and ion masses. These were GA₄, GA₁₂, GA₁₅, and GA₂₄ (from the non-13-hydroxylation pathway) and GA₃ (from the early-13-hydroxylation pathway) (data not shown).

The most abundant GA observed in both cauliflower and broccoli was GA₂₉, an inactive metabolite of GA₂₀, with levels in cauliflower and broccoli of 167 ng GA per gram FW and 84 ng GA per gram FW, respectively (Fig. 2). In cauliflower, the physiologically active GA₁ was the second most abundant GA (5.03 ng GA per gram FW) followed by GA₁₉ (4.36 ng GA per gram FW), a GA that is intermediate in the biosynthesis pathway (Fig. 2). In broccoli, GA₁₉ was not detected; instead, the subsequent precursor, GA₂₀, was abundant (9.66 ng GA per gram FW) (Fig. 2). In the early 13-hydroxylation pathway, GA₅₃, GA₄₄, and GA₁₉ are precursors of GA₂₀; the latter is the precursor of the active gibberellin GA₁. GA₈ and GA₂₉ are inactive metabolites. These results indicate that GAs from both the early-13-hydroxylation pathway (specifically, GA₁, GA₃, precursors, and inactive metabolites) and the non-13-hydroxylation pathway (GA₄) are present in inflorescence meristem and floral buds.

View Figure

• Download as Powerpoint
• Download Figure

Effect of exogenous GA on reproductive induction.

GA was applied to developing meristems at two different stages of vegetative development. The adult-vegetative stage was observed 34 DAS and ≈16 d before expected inflorescence meristem initiation. At this stage, GA₃ advanced the reproductive transition in both cauliflower and broccoli. The earlier transition was evident as fewer vegetative nodes and earlier curding. After GA₃ application, 3.8 fewer leaves (vegetative nodes) appeared in cauliflower (P = 0.001) and 1.5 fewer leaves appeared in broccoli (P = 0.011) than in control plants (Table 2). GA₃ application caused the curd to become visible 9.6 d earlier in cauliflower (P = 0.001) and 5.0 d earlier in broccoli (P < 0.001) relative to controls (Table 2). Stem elongation increased in cauliflower; at the end of the treatment, the stem was 9.5 cm longer in cauliflower plants treated with GA₃ than in untreated controls (P ≤ 0.001). No significant change in the stem length was observed in broccoli (Table 2).

Table 2.Effect of exogenous GA₃ applied at the initiation of the adult vegetative stage, 30 d after sowing, on the reproductive transition in ‘Barbados’ broccoli and ‘Somerset’ cauliflower (n = 15).

View Fullscreen

When GA₃ and GA₄₊₇ were applied closer to the constitutive transition time (38 DAS, ≈12 d before expected inflorescence meristem initiation), the time to curding was significantly shorter in both inflorescence types. In cauliflower, GA₃ and GA₄₊₇ reduced the curding time by 2.2 d (P = 0.025) and 3.6 d (P = 0.001), respectively (Table 3). In broccoli, GA₃ and GA₄₊₇ reduced the curding time by 4.3 d (P = 0.03) and by 7 d (P < 0.001), respectively.

Table 3.Effect of exogenous GA₃ and GA₄₊₇ applied at the end of the adult vegetative stage, 38 d after sowing, on the reproductive transition in ‘Barbados’ broccoli and ‘Somerset’ cauliflower (n = 10).

View Fullscreen

GA application also resulted in fewer nodes being produced before the transition. As expected, the effect was smaller when application was closer to the constitutive time. The effect of GA₃ was smaller than that of GA_4+7. In cauliflower, GA reduced the number of leaves by 1.6 [GA₄₊₇ (P = 0.032)] or 0.4 [GA₃ (non-significant)]. In broccoli, the reduction was 2.0 [GA₄₊₇ (P = 0.007)] or 0.8 [GA₃ (non-significant)] leaves (Table 3). Stem elongation was increased by GA₄₊₇ in cauliflower, but not in broccoli. In cauliflower, GA₄₊₇ increased post-application elongation by 2.7 cm (P = 0.003), but GA₃ had no significant effect (Table 3). Overall, broccoli responded more than did cauliflower, perhaps because cauliflower was closer to the vegetative-to-reproductive transition. GA₄₊₇ had a stronger effect than GA₃ on all variables measured in both cultivars.

Effect of exogenous GA on subsequent curd morphology.

GA was applied to ‘Somerset’ cauliflower and ‘Barbados’ broccoli at two different stages of reproductive development: the onset of the reproductive meristem (43 DAS) and curd enlargement (50 and 55 DAS for ‘Somerset’ and ‘Barbados’, respectively). Whether applied at the beginning of reproductive meristem proliferation (43 DAS) or when it was well advanced (50 to 55 DAS), GA application did not affect the stage of arrest in either ‘Somerset’ cauliflower or ‘Barbados’ broccoli. Cauliflower plants were arrested at the inflorescence meristem stage and broccoli plants were arrested at the floral bud stage. However, application at the onset of reproductive development in ‘Somerset’ cauliflower plants (43 DAS) increased the development of bracts (Fig. 3). In particular, GA₄₊₇ treatment caused all the plants to have some degree of bracting with 90% exhibiting a high degree of bracting.

View Figure

• Download as Powerpoint
• Download Figure

GA did not significantly increase cauliflower stem elongation or node production when applied to the reproductive meristem (Table 4). Broccoli curded 3.4 d earlier than controls (P = 0.038) when GA₄₊₇ was applied at this stage (Table 4), but the same number of nodes was produced.

Table 4.Effect of exogenous GA₃ and GA₄₊₇ on the reproductive transition in ‘Barbados’ broccoli and ‘Somerset’ cauliflower when applied in the early inflorescence meristem stage (43 d after sowing) and in the middle of the inflorescence meristem stage (55 d after sowing) (n = 12).

View Fullscreen

Effect of exogenous GA on the transition to floral meristem in cauliflower curds.

To detect whether GA induces genes involved in floral-primordium initiation, the expression of three meristem-identity genes, BoLFY, BoAP1-a, and BoAP1-c, and the integrator gene, SOC1, was monitored after application of GA to curds of cauliflower cv. Somerset. Amplicons of BoAP1-a and BoAP1-c were not detected in any of the tissue samples taken from either the control plants or the plants treated with GA₃. BoLFY expression showed no significant differences for the main effects GA (P = 0.508) and time (P = 0.115) or for their interaction (P = 0.247) (Fig. 4). The expression of SOC1 was unaffected by the exogenous application of GA (P = 0.638), time (P = 0.119), or their interaction [GA × time (P = 0.063)]. Overall GA did not alter the expression of tested genes associated with flower development.

View Figure

• Download as Powerpoint
• Download Figure

Gibberellin advances reproductive induction in cauliflower and broccoli.

The GA pathway of reproductive induction (commonly called flowering time) varies widely in its importance among different species. In this experiment, GA substantially accelerated the transition from a vegetative to reproductive meristem. Application of GA₃ and GA₄₊₇ shortened the vegetative period in both cauliflower and broccoli. The earlier transition was evident as fewer vegetative nodes and earlier curding. This result is consistent with GA accelerating the transition from vegetative to reproductive development. In B. napus var. annua (Dahanayake and Galwey, 1999), exogenous application of GA₃ was shown to reduce the number of nodes at flowering and the time to flowering, similar to the results shown for cauliflower and broccoli in the present study.

Gibberellin does not control the transition from inflorescence meristem to floral primordium in cauliflower and broccoli.

The long duration of the inflorescence meristem stage is the defining trait of cauliflower and broccoli. In cauliflower, the inflorescence meristem stage continues until harvest without flower induction. The potential of broccoli to form so many flower buds at once is the consequence of the extended proliferation of inflorescence meristems followed by the coordinated induction of floral primordia. These distinctive phenotypes are produced because floral primordia are induced in the inflorescence meristem at different times. Applying GA₃ or GA₄₊₇ to the early reproductive meristem did not trigger this transition from inflorescence meristem to floral primordium in either ‘Somerset’ cauliflower or ‘Barbados’ broccoli plants. No change in the stage of cauliflower curds was observed; all plants were arrested at the inflorescence meristem stage. Booij (1990) similarly found no change in the quality of cauliflower curds after GA₄₊₇ treatments in summer cauliflowers. The GA content of leaves has been shown to decline during this transition (Jiang et al., 2011).

In cauliflower, GA application also did not cause detectable upregulation of the meristem-identity genes BoAP1-a and BoAP1-c, which would be indicative of reproductive induction. Although GAs are known to accelerate the transition from vegetative to reproductive meristem in arabidopsis by up-regulating LFY and SOC1 (Blazquez et al., 1998; Gocal et al., 2001; Moon et al., 2003), the homologous genes were not upregulated in cauliflower inflorescence meristems (Fig. 4). The increase in BoAP1-a and BoAP1-c transcript levels observed in ‘Green Harmony’ cauliflower just before floral-primordium initiation and the BoLFY expression observed once plants are in the reproductive stage (Duclos and Björkman, 2008) are therefore not caused by an increase in GA. Moreover, it could be caused by exactly the opposite, a decrease in the levels of the phytohormone. In this regard, cauliflower resembles arabidopsis, where GAs applied during reproductive development do not activate meristem-identity genes or A-function genes (Yu et al., 2004). In fact, in arabidopsis, the promotion of flowering occurs as a result of a reduction in GA by a GA catabolism gene that is up-regulated by LFY (Yamaguchi et al., 2014).

Gibberellin and bolting.

GA induces bolting-associated stem elongation (Hamano et al., 2002). That function of GA was confirmed for broccoli in these tests. When GA₃ was applied to small heads of ‘Barbados’ broccoli plants just after buds had developed, the area of the curd receiving the GA drop subsequently developed an elongated stalk, whereas the rest of the head remained compact (Fig. 5). Such stalk elongation normally occurs later, when the whole head elongates at the time of coordinated anthesis.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

In broccoli, applied GAs had no effect on the stage of the arrest. Flower buds were developed and remained unopened. GA also increased stem elongation to some extent during the vegetative stage when GA was applied to the apex. As part of the test of reproductive induction, cauliflower produced longer internodes after GA application early in the adult vegetative stage. No effect of the hormone was observed when its application was closer to reproductive commitment (Tables 2 to 4). The same was observed in broccoli.

Gibberellin, bracting, and vegetative reversion in cauliflower.

Although GA generally promotes reproductive development (Andrés and Coupland, 2012; Srikanth and Schmid, 2011), in one respect it promoted vegetative reversion. In ‘Somerset’ cauliflower, application of GAs to the inflorescence meristem at the initiation of reproductive development increased the incidence of bracts in mature heads (Fig. 3). The effect was particularly pronounced with GA₄₊₇, in which greater than 90% of observed curds developed many bracts. Bracting results from the release of leaf primordia subtending each inflorescence meristem. “Bracty” curds have bracteoles forming around each flower primordium, manifesting a reversion of the reproductive phase to a vegetative phase (Grevsen et al., 2003). In many perennial species, the application of exogenous GAs has caused reversion from reproductive to vegetative development (reviewed by Bergonzi and Albani, 2011). However, this result was not expected based on previous reports in cauliflower and arabidopsis. Booij (1990) found less bracting after applying GA₄₊₇ to cauliflower. In arabidopsis, floral-meristem reversion has been associated with a decrease in GA levels or GA activity (Okamuro et al., 1996), reduced levels of LFY (Ferrandiz et al., 2000), and consequently reduced upregulation of AP1 (Liljegren et al., 1999). This contradiction might be explained if the GA response is concentration-dependent in B. oleracea. In Pharbitis nil for example, a high dose of GAs suppresses floral induction and a low dose accelerates flowering (King et al., 2000). Exogenous application of GA may have caused a reversion-inducing concentration of GA immediately after the vegetative-to-reproductive transition. Because GA₄₊₇ had a stronger effect than GA₃ in bract incidence, the early and non-hydroxylation pathways may be controlling this process differentially. The inconsistent bracting response between this early-maturing cultivar and others reported in the literature may reflect the effect of variation in the bracting genes previously identified by Dickson and Lee (1980).

Genetic variation in GA response among cauliflower cultivars.

The present results differ from some previous studies regarding the effect of GA₃ and GA₄₊₇ in floral induction of cauliflower and broccoli. To some extent, that difference reflects genetic variation in the flower development among cultivars. In arabidopsis the reproductive transition is controlled by a complex network of five flowering pathways that involve response to GA as well as daylength, vernalization, carbohydrates, and an age-related autonomous pathway. In B. oleracea, the wide diversity of inflorescence timing and morphology among cultivars implies wide variation within the species in the relative influence of each of those five pathways. Indeed, several distinct quantitative trait loci have been identified for each of the two developmental transitions (Uptmoor et al., 2012). An example is the varying effect of GA among cauliflowers that vary in their vernalization requirement. Leshem and Steiner (1968) found that GA₃ potentiated but did not substitute for vernalization in reducing the number of nodes before curd formation and accelerating flowering in a winter cauliflower. However, in temperate [cv. Lawyna (Aditya and Fordham, 1995) and fall cauliflower [cv. Nautilus (Fernandez et al., 1997)], GAs did not potentiate vernalization but alone induced earlier curd formation when applied under weak curd-inducing conditions. ‘Somerset’, the summer cauliflower used in the present study, does not require vernalization to form the curd. Therefore, fewer induction pathways are involved in the vegetative-to-reproductive transition, and the relative importance of the GA pathway is likely increased. The positive response of ‘Somerset’ to both GAs supports the model that the GA response exists in all cauliflower, but it is masked in those with a strong vernalization requirement.

It has been an open question whether GAs from both hydroxylation pathways are active in cauliflower. Both pathways have been found in other brassicas [B. oleracea var. capitata (Hamano et al., 2002), Raphanus sativus (Nishijima et al., 1998), and arabidopsis (Talon et al., 1990)]. However, GAs in the early-13-hydroxylated pathway (GA₁, GA₃) were previously found not to induce flowering in summer or fall cauliflowers [‘Delira’ (Booij, 1989, 1990); ‘Nautilus’ (Fernandez et al., 1997)], raising the possibility that this pathway is absent. It is also possible that the response to GA₁ and GA₃ was not detected in earlier work if they were applied before or after the brief responsive stage or if application to leaves did not result in a sufficient dose at the meristem.

Here, ‘Somerset’ cauliflower and ‘Barbados’ broccoli responded to both GA₃ (early 13-hydroxylation pathway) and GA₄₊₇ (non-13-hydroxylation pathway); also, GAs from both pathways were present in the curds, indicating that both the early 13-hydroxylation pathway and the non-13-hydroxylation pathway are present and active in inflorescence meristem and floral buds of these cultivars.