Gibberellins in Suspensors of Phaseolus coccineus L. Seeds

Analysis of extracts from 6300 (1.2 grams fresh weight) Phaseolus coccineus suspensors by combined gas chromatography-mass spectrometry has demonstrated the presence of five C₁₉-gibberellins, GA₁, GA₄, GA₅, GA₆, GA₈, and one C₂₀-GA, GA₄₄. The major GAs present were GA₁ and GA₈. Data are discussed in relation to previous results obtained in P. coccineus seed as well as in the embryo-suspensor system.     

Gibberellins in developing fruits of Pisum sativum cv. Alaska: Studies on their role in pod growth and seed development

Gibberellins A₁, A₈, A₂₀ and A₂₉ were identified by capillary gas chromatography-mass spectrometry in the pods and seeds from 5-d-old pollinated ovaries of pea (Pisum sativum cv. Alaska). These gibberellins were also identified in 4-d-old non-developing, parthenocarpic and pollinated ovaries. The level of gibberellin A₁ within these ovary types was correlated with pod size. Gibberellin A₁, applied to emasculated ovaries cultured in vitro, was three to five times more active than gibberellin A₂₀. Using pollinated ovary explants cultured in vitro, the effects of inhibitors of gibberellin biosynthesis on pod growth and seed development were examined. The inhibitors retarded pod growth during the first 7 d after anthesis, and this inhibition was reversed by simultaneous application of gibberellin A₃. In contrast, the inhibitors, when supplied to 4-d-old pollinated ovaries for 16 d, had little effect on seed fresh weight although they reduced the levels of endogenous gibberellins A₂₀ and A₂₉ in the enlarging seeds to almost zero. Paclobutrazol, which was one of the inhibitors used, is xylem-mobile and it efficiently reduced the level of seed gibberellins without being taken up into the seed. In intact fruits the pod may therefore be a source of precursors for gibberellin biosynthesis in the seed. Overall, the results indicate that gibberellin A₁, present in parthenocarpic and pollinated fruits early in development, regulates pod growth. In contrast the high levels of gibberellins A₂₀ and A₂₉, which accumulate during seed enlargement, appear to be unnecessary for normal seed development or for subsequent germination.Abbreviations GA_(a) gibberellin A_n - GC-MS combined gas chromatography-mass spectrometry - HPLC high-performance liquid chromatography - PFK perfluorokerosene - PVP polyvinylpyrrolidone     

Metabolism of gibberellins in a cell-free system from immature seeds of Phaseolus vulgaris L.

The biosynthetic steps from gibberellin A₁₂-aldehyde (GA₁₂-aldehyde) to C₁₉-GAs were studied by means of a cell-free system from the embryos of immature Phaseolus vulgaris seeds. Stable-isotope-labeled GAs were used as substrates and the products were identified by gas chromatography-mass spectrometry. Gibberellin A₁₂-aldehyde was converted to GA₄ via non-hydroxylated intermediates and to GA₁ via 13-hydroxylated intermediates. 13-Hydroxylation took place at the beginning of the pathway by the conversion of GA₁₂-aldehyde to GA₅₃-aldehyde. The conversion of GA₂₀ to GA₅ and GA₆ was also shown but no 2-hydroxylating activity was found. Endogenous GAs from embryos and testas of 17-dold seeds were re-examined by gas chromatography-selected ion monitoring using stable-isotopelabeled GAs as internal standards. Gibberellins A₉, A₁₂, A₁₅, A₁₉, A₂₃, A₂₄, and A₅₃ were identified for the first time in P. vulgaris, in addition to GA₁, GA₄, GA₅, GA₆, GA₈, GA₁₇, GA₂₀, GA₂₉, GA₃₇, GA₃₈ and GA₄₄, which were previously known to occur in this species. The levels of all GAs, except the 2-hydroxylated ones, were greater in the embryos than in the testas. Conversely, the contents of GA₈ and GA₂₉, both 2-hydroxylated, were much higher in the testas than in the embryos.Abbreviations GA_n gibberellin A_n - GC-MS gas chromatography-mass spectrometry - GC-SIM gas chromatography-selected ion monitoring - HPLC high-performance liquid chromatography - TLC thin-layer chromatography - m/z ion of mass     

Gibberellins inCitrus sinensis: A comparison between seeded and seedless varieties

The gibberellin (GA) content of the reproductive organs ofCitrus sinensis (L.) Osb., cv. Bianca Comuna and the seedless variety, Salustiana, were examined by combined gas chromatography-mass spectrometry (GC/MS) at different stages of development. Gibberellins A₁, A₂₀, and A₂₉ were identified in the reproductive buds of both cultivars 21 days prior to anthesis and in fruits 35 days after anthesis by comparison of their mass spectra and Kovats retention indices with those of standards. In addition, three uncharacterized isomers of GA₁ were detected, one in buds and two in fruits. The presence of GA₄ in both tissues, and of GA₈ in the reproductive buds, was indicated by the occurrence of characteristic ions at the expected retention times, although their spectra were too weak for full identification. Vegetative shoots of cv. Salustiana contained gibberellins A₁, A₁₉, A₂₀, and A₂₉, and the unidentified isomer of GA₁ present in reproductive buds. The presence of trace amounts of gibberellins A₈ and A₁₇ was also indicated. Although the two varieties did not differ qualitatively in the GAs present during flower and fruit development, the seedless variety contained slightly greater amounts. The concentrations of gibberellins A₁, A₄, and A₂₀ were determined by gas chromatography-selected ion monitoring (GC/SIM) throughout ovary development and early fruit growth. In both varieties, the maximum GA₁ concentration occurred at anthesis. Highest concentrations of gibberellins A₂₀ and A₄ were found in fruit 35 days after anthesis, although the GA₁ concentration at this stage remained low.     

Conversion of [14C]gibberellin A12-aldehyde to C19- and C20-gibberellins in a cell-free system from immature seed of Phaseolus coccineus L.

A cell-free system prepared from developing seed of runner bean (Phaseolus coccineus L.) converted [¹⁴C]gibberellin A₁₂-aldehyde to several products. Thirteen of these were identified by capillary gas chromatography-mass spectrometry as gibberellin A₁ (GA₁), GA₄, GA₅, GA₆, GA₁₅, GA₁₇, GA₁₉, GA₂₀, GA₂₄, GA₃₇, GA₃₈, GA₄₄ and GA₅₃-aldehyde, all giving mass spectra with ¹⁴C-isotope peaks. GA₈ and GA₂₈ were also identified but contained no ¹⁴C. All the [¹⁴C]GA₁₂-aldehyde metabolites, except GA₁₅, GA₂₄ and GA₅₃-aldehyde, are known endogenous GAs of P. coccineus.Abbreviations GA_n gibberellin A_n - GC-MS combined gas chromatography-mass spectrometry - HPLC highperformance liquid chromatography - MVA mevalonic acid - S-2 2000-g supernatant     

Influence of an inhibitor of gibberellin biosynthesis, paclobutrazol, on apple seed gibberellin content

Tissue-culture-propagated own-rooted cv. Spartan apple trees (Malus domestica Borkh.) planted in 1979 were treated in 1983 and 1985 via a soil-line trunk drench with the plant growth retardant paclobutrazol [(2RS, 3RS)-1-(4-chlorophenyl)-4.4-dimethyl-2-(1,2, 4-triazol-1-yl) pentan-3-ol]. Seeds of immature fruits from untreated and treated trees were sampled in 1989 ca 75 days after full bloom. After seeds were freeze-dried, gibberellins (GAs) were extracted, purified and fractionated via C₁₈ reversed-phase high-performance liquid chromatography (HPLC). Gibberellins A₁, A₃, A₄, A₇, A₈, A₉, A₁₅, A₁₇, A₁₉, A₂₀, A₂₄, A₃₄, A₃₅, A₄₄, A₅₁, A₅₃, A₅₄, A₆₁, A₆₂, A₆₃ and A₆₈ were identified by using C₁₈ HPLC, gas chromatography-selected ion monitoring and Kovats retention indices. Eight of the GAs identified were also quantified by using deuterated internal standards. The paclobutrazol applications caused a 55% reduction of vegetative shoot elongation in 1989, but both treated and untreated trees had developed a biennial bearing pattern by that time (heavy bloom or “on year’in 1989). Levels of early 13-hydroxylation pathway GAs, viz. GA₅₃, GA₁₉, GA₂₀, GA₁ and also GA₃, were not altered by treatment. However, GA₄, GA₇ and GA₉ were increased 13.4, 6.5 and 3.8 times, respectively, in seeds of fruit from treated compared to untreated trees.     

Overexpression of 20-Oxidase Confers a Gibberellin-Overproduction Phenotype in Arabidopsis

In the gibberellin (GA) biosynthesis pathway, 20-oxidase catalyzes the oxidation and elimination of carbon-20 to give rise to C₁₉-GAs. All bioactive GAs are C₁₉-GAs. We have overexpressed a cDNA encoding 20-oxidase isolated from Arabidopsis seedlings in transgenic Arabidopsis plants. These transgenic plants display a phenotype that may be attributed to the overproduction of GA. The phenotype includes a longer hypocotyl, lighter-green leaves, increased stem elongation, earlier flowering, and decreased seed dormancy. However, the fertility of the transgenic plants is not affected. Increased levels of endogenous GA₁, GA₉, and GA₂₀ were detected in seedlings of the transgenic line examined. GA₄, which is thought to be the predominantly active GA in Arabidopsis, was not present at increased levels in this line. These results suggest that the overexpression of this 20-oxidase increases the levels of some endogenous GAs in transgenic seedlings, which causes the GA-overproduction phenotype.     

Gibberellins in dark- and red-light-grown shoots of dwarf and tall cultivars of Pisum sativum: The quantification,metabolism and biological activity of gibberellins in Progress no. 9 and Alaska

The stem growth in darkness or in continuous red light of two pea cultivars, Alaska (Le Le, tall) and Progress No. 9 (le le, dwarf), was measured for 13 d. The lengths of the first three internodes in dark-grown seedlings of the two cultivars were similar, substantiating previous literature reports that Progress No. 9 has a tall phenotype in the dark. The biological activity of gibberellin A₂₀ (GA₂₀), which is normally inactive in le le geno-types, was compared in darkness and in red light. Alaska seedlings, regardless of growing conditions, responded to GA₂₀. Dark-grown seedlings of Progress No. 9 also responded to GA₂₀, although red-light-grown seedlings did not. Gibberellin A₁ was active in both cultivars, in both darkness and red light. The metabolism of [¹³C³H]GA₂₀ has also been studied. In dark-grown shoots of Alaska and Progress No. 9 [¹³C³H]GA₂₀ is converted to [¹³C³H]GA₁, [¹³C³H]GA₈, [¹³C]GA₂₉, its 2-epimer, and [¹³C³H]GA₂₉-catabolite. [¹³C³H] Gibberellin A₁ was a minor product which appeared to be rapidly turned over, so that in some feeds only its metabolite, [¹³C³H]GA₈, was detected. However results do indicate that the tall growth habit of Progress No. 9 in the dark, and its ability to respond to GA₂₀ in the dark may be related to its capacity to 3-hydroxylate GA₂₀ to give GA₁. In red light the overall metabolism of [¹³C³H]GA₂₀ was reduced in both cultivars. There is some evidence that 3-hydroxylation of [¹³C³H]GA₂₀ can occur in red light-grown Alaska seedlings, but no 3-hydroxylated metabolites of [¹³C³H]GA₂₀ were observed in red light-grown Progress. Thus the dwarf habit of Progress No. 9 in red light and its inability to respond to GA₂₀ may be related, as in other dwarf genotypes, to its inability to 3-hydroxylate GA₂₀ to GA₁. However identification and quantification of native GAs in both cultivars showed that red-light-grown Progress does contain native GA₁. Thus the inability of red light-grown Progress No. 9 seedlings to respond to, and to 3-hydroxylate, applied GA₂₀ may be due to an effect of red light on uptake and compartmentation of GAs.Abbreviations AMO-1618 2-isopropyl-4-(trimethylammonium chloride)-5-methylphenyl piperidine-1-carboxylate - cv. cultivar - GC-MS gas chromatography-mass spectrometry - GA_(n) gibberellin A_(n) - HPLC high-pressure liquid chromatography     

Identification of endogenous gibberellins from oilseed rape

Oilseed rape (Brassica napus, canola variety `Westar') plants were grown in greenhouse conditions and shoots were harvested during the final stages of shoot elongation. Leaves and immature pods were removed and the remaining stem tissue was extracted and purified. The extract was chromatographed on sequential, step-eluted silica gel partition and reverse-phase C<sub>18</sub> HPLC columns, and gibberellin (GA)-like substances were detected using the `Tan-ginbozu' dwarf rice microdrop assay. Purified fractions showing GA-like activity were analyzed by capillary gas chromatography-mass spectrometry (GC-MS) and GC-selected ion monitoring (GC-SIM). Gibberellins A₁, A₃, and iso-A₃ were identified by full spectrum GC-MS with GA₁ being the most abundant GA in the stem tissue. Gibberellins A₁₉ and A₂₀ were identified by GC-SIM and are logical precursors of the GA₁.     

Internode length in pisum. Mutants lk,lka, and lkb do not accumulate gibberellins

The levels of gibberellin A₁ (GA₁), GA₈, GA₁₉, GA₂₀, GA₂₉, and GA₄₄ in the short Pisum sativum L. mutants lk, lka, and lkb, and comparable wild-type plants, were determined by gas chromatography-selected ion monitoring (GC-SIM) using ²H or ¹³C internal standards. The mutants possessed similar GA₁ levels to wild-type plants, consistent with their classification as GA-sensitivity rather than GA-synthesis mutants. However, these mutants differ from certain sensitivity mutants in other species, in which substantial accumulation of GA₁ occurs. The results suggest that if the proposed feedback model for the regulation of GA synthesis occurs in peas it is not the reduced growth per se that is the trigger for elevated levels of C₁₉ GAs. The results are also consistent with the hypothesis that in those GA-sensitivity mutants which do not accumulate C₁₉ GAs, the biochemical lesion may be well down the transduction pathway which leads from GA₁ reception to stem elongation.     

Interaction of growth retardants,daylength, and gibberellins A19, A20, and A1 on shoot elongation in birch and alder

Gibberellins A₁₉, A₂₀, and A₁ were applied to seedlings of birch (Betula pubescens Ehrh.) and alder (Alnus glutinosa (L.) Gaertn.) in order to test their ability to counteract growth inhibition induced by growth retardants (ancymidol and BX-112) or short day (SD, 12 h) photoperiod. Ancymidol inhibits early and BX-112 inhibits late steps in gibberellin biosynthesis. BX-112 inhibited stem elongation in both species while ancymidol, applied as a soil drench, was effective in alder only. Growth retardants affected stem elongation mainly by inhibiting elongation of internodes. All three gibberellins were equally active when applied to seedlings treated with ancymidol; however, only GA₁ was able to counteract the growth inhibition induced by BX-112. SD-induced cessation of elongation growth in birch was counteracted by GA₁, and to some degree, by GA₂₀, while GA₁₉ was inactive. SD treatment did not induce cessation of apical growth in alder. These results are consistent with the hypothesis that of gibberellins belonging to the early C-13 hydroxylation pathway, GA₁ is the only active gibberellin for stem elongation.     

Conversion of gibberellin A20 to gibberellins A1 and A5 in a cell-free system from Phaseolus vulgaris

The soluble fraction of a cell-free system from immature seeds of Phaseolus vulgaris L. converts gibberellin A₂₀ (GA₂₀) to GA₁ and GA₅. It does however not metabolize GA₁ and GA₂₉ to GA₅, showing that in this system GA₂₀ is converted directly to GA₅. The steps from GA₂₀ to GA₁ (3-hydroxylation) and from GA₂₀ to GA₅ (² double-bond formation) require oxygen, Fe²⁺ and -ketoglutarate, and are stimulated by ascorbate. The enzymes catalyzing these conversions bate. The enzymes catalyzing these conversions have properties similar to those of GA oxidases found in Cucurbita maxima and Pisum sativum.Abbreviations GA_n gibberellin A_n - HPLC high-performance liquid chromatography - GC-MS combined gas chromatography-mass spectrometry - TLC thin-layer chromatography - TMSi/TMSi trimethylsilyl ether/trimethylsilyl ester Graduate student, University of Tokyo     

Metabolism of tritiated gibberellin a(20) in maize

After the application of 2.36 Curies per millimole [2,3-³H]gibberellin A₂₀ (GA₂₀) to 21-day-old maize (Zea mays L., hybrid CM7 × CM49) plants, etiolated maize seedlings, or maturing maize cobs, a number of ³H-metabolites were observed. The principal acidic (pH 3.0), ethyl acetate-soluble metabolite was identified as [³H]GA₁ on the basis of co-chromatography with standard [³H]GA₁ on SiO₂ partition, high resolution isocratic elution reverse phase C₁₈ high performance liquid chromatography and gas-liquid chromatography radiocounting. Two other acidic metabolites were identified similarly as [³H]GA₈ and C/D ring-rearranged [³H]GA₂₀, although gas-liquid chromatography radiocounting was not performed on these metabolites. Numerous acidic, butanol-soluble (e.g. ethyl acetate-insoluble) metabolites were observed with retention times on C₁₈ high performance liquid chromatography radiocounting similar to those of authentic glucosyl conjugates of GA₁ and GA₈, or with retention times where conjugates of GA₂₀ would be expected to elute. Conversion to [³H]GA₁ was greatest (23% of methanol extractable radioactivity) in 21-day-old maize plants. In etiolated maize seedlings, the C/D ring-rearranged [³H]GA₂₀-like metabolite was the major acidic product, while conversion to [³H]GA₁ was low.     

Endogenous Gibberellins in the Xylem Exudate from Apple Trees

In the xylem exudate extracted from the current-year stems of apple (Malus domestica Borkh.), gibberellins A₁₅, A₁₇, A₁₈, A₁₉, A₂₃, A₄₄, and A₅₃ were identified, and 16,17-dihydro-17-hydroxy GA₁₉ was presumed from full-scan mass spectra and Kovats retention indices.     

Autoradiographic studies on the role of plasmodesmata in the transport of gibberellin

Translocation of [¹⁴C]gibberellic acid into antheridial cells of Chara vulgaris L. was investigated in relation to the presence of symplasmic connections between the antheridium and the thallus. It was found that manubria, capitular cells, and antheridial filaments were about three-fold more strongly labelled in young antheridia connected to the thallus by plasmodesmata than in older antheridia in which spontaneous symplasmic isolation had occurred. Plasmolytically induced symplasmic isolation of young antheridia severely diminished the radioactivity of all the cells, down to the level characteristic for spontaneously isolated antheridia. It is concluded that plasmodesmata are the main channel of gibberellin transport into antheridia. The change in the character of symplasmic connections during the course of morphogenesis might, among other events, constitute a signal determining a shift of cell metabolism in a new direction, in response to a rapid change in gibberellin level.Abbreviations GA_(n) gibberellin (A_n) - GA₃ gibberellic acid - IAA indole-3-acetic acid This study was supported by the Polish Academy of Sciences research project CPBP 04.01.5.05.     

Gibberellins inCitrus sinensis: A comparison between seeded and seedless varieties

The gibberellin (GA) content of the reproductive organs ofCitrus sinensis (L.) Osb., cv. Bianca Comuna and the seedless variety, Salustiana, were examined by combined gas chromatography-mass spectrometry (GC/MS) at different stages of development. Gibberellins A₁, A₂₀, and A₂₉ were identified in the reproductive buds of both cultivars 21 days prior to anthesis and in fruits 35 days after anthesis by comparison of their mass spectra and Kovats retention indices with those of standards. In addition, three uncharacterized isomers of GA₁ were detected, one in buds and two in fruits. The presence of GA₄ in both tissues, and of GA₈ in the reproductive buds, was indicated by the occurrence of characteristic ions at the expected retention times, although their spectra were too weak for full identification. Vegetative shoots of cv. Salustiana contained gibberellins A₁, A₁₉, A₂₀, and A₂₉, and the unidentified isomer of GA₁ present in reproductive buds. The presence of trace amounts of gibberellins A₈ and A₁₇ was also indicated. Although the two varieties did not differ qualitatively in the GAs present during flower and fruit development, the seedless variety contained slightly greater amounts. The concentrations of gibberellins A₁, A₄, and A₂₀ were determined by gas chromatography-selected ion monitoring (GC/SIM) throughout ovary development and early fruit growth. In both varieties, the maximum GA₁ concentration occurred at anthesis. Highest concentrations of gibberellins A₂₀ and A₄ were found in fruit 35 days after anthesis, although the GA₁ concentration at this stage remained low.     

Quantitation of gibberellins A1, A3, A4, A9 and an A9-conjugate in good- and poor-flowering clones of Sitka spruce (Picea sitchensis) during the period of flower-bud differentiation

The levels of endogenous gibberellin A₁ (GA₁), GA₃, GA₄, GA₉ and a cellulase-hydrolysable GA₉-conjugate in needles and shoot stems of Sitka spruce [Picea sitchensis (Bong.) Carr.] grafts with different coning or flowering histories were estimated by combined gas chromatography-mass spectrometry selected ion monitoring using deuterated GA₃, GA₄ and GA₉ as internal standards. The samples were taken at the approximate time of the start of flower-bud differentiation, i.e. when the shoots had elongated approx. 95% of the final length. The needles of the good-flowering clones contained 11–12 ng per g fresh weight (FW) and 15–28 ng· (g FW) ^–1 of GA₉-conjugate and GA₉, respectively. The shoot stems of the same material contained no detectable amounts of GA₉-conjugate and 11–15 ng-(g FW)^–1 of GA₉. The amounts of GA₉-conjugate and GA₉ were apparently lower in the poor-flowering clones, the needles containing 4–9 ng-(g FW)^–1 and 7–17 ng·(g FW)^–1, respectively. Also in this material the shoot stems contained no detectable amounts of GA₉-conjugate. The amounts of GA₄ were very small in both materials, ranging from 1–1.6 ng-(g FW)^–1. The good-flowering clones contained no detectable amounts of the more polar gibberellins, GA₁ and GA₃. The poor-flowering clones, on the other hand, contained high levels of GA₁₅ 17–19ng·(gFW)^–1 in the needles and 10–13 ng·(g FW) ^–1 in the shoot stems, and also smaller amounts of GA₃, 2–3 ng·(g FW)^–1 in the needles and approx. 1 ng·(g FW)^–1 in the shoot stems. The results demonstrate differences in GA-metabolism between the poor- and the good-flowering clones. The higher amounts of GA₉-conjugate and GA₉ might indicate a higher capacity for synthesizing GA₄ in the good-flowering material. This synthesis does not, however, result in a build-up of the GA₄-pool, maybe because of a high rate of turnover. Gibberellin A₄ was apparently neither hydroxylated to GA₁ nor converted to GA₃ in the goodflowering material, as was the case in the poor-flowering material. This might indicate that gibberellin metabolism in the poor-flowering material is directed towards GA₁ and GA₃, GAs preferentially used in vegetative growth.Abbreviations FW fresh weight - GA_n gibberellin A_n - HPLC high-performance liquid chromatography     

Persistence of [14C] gibberellin A3 and [3H] gibberellin A1 in senescing,ethylene treated Citrus and tomato fruit

The fate of [¹⁴C] gibberellin A₃ and [³H] gibberellin A₁ was examined in senescing fruit of Shamouti orange (Citrus sinensis L. Osbeck) and tomato (Lycopersicon esculentum Mill.). Gibberellin A₃ was highly persistent in Citrus peel (t 1/2=18 days) and to a lesser degree in tomato (t 1/2=5.5 days). Ethylene and ethephon caused a slight enhancement of gibberellin A₃ metabolism in Citrus and tomato fruit, respectively. Gibberellin A₁ was metabolized by Citrus peel at a relatively high rate (t 1/2 < 24 h) and ethylene slightly reduced this rate. It is concluded that the ethylene-induced enhancement of senescence does not involve major effects on the deactivation of applied gibberellins.Abbreviations GA₃ gibberellin A₃ - GA₁ gibberellin A₁     

The role of gibberellins A1 and A3 in fruit growth of Pisum sativum L. and the identification of gibberellins A4 and A7 in young seeds

Gibberellins A₁ and A₃ are the major physiologically active gibberellins (GAs) present in young fruit of pea (Pisum sativum L.). The relative importance of these GAs in controlling fruit growth and their biosynthetic origins were investigated in cv. Alaska. In addition, the non-13-hydroxylated active GAs, GA₄ and GA₇, were identified for the first time in young seeds harvested 4 d after anthesis, although they are minor components and are not expected to play major physiological roles. The GA₁ content is maximal in seeds and pods at 6 d after anthesis, the time of highest growth-rate of the pod (Garcia-Martinez et al. 1991, Planta 184: 53–60), whereas gibberellic acid (GA₃), which is present at high levels in seeds 4–8 d after anthesis, has very low abundance in pods. Gibberellins A₁₉, A₂₀ and A₂₉ are most concentrated in seeds at, or shortly after, anthesis and their abundance declines rapidly with development, concomitant with the sharp increase in GA₁ and GA₃ content. Application of GA₁ or GA₃ to the leaf subtending an emasculated flower stimulated parthenocarpic fruit development. Measurement of the GA content of the pods at 4 d after anthesis indicated that only 0.002–0.5% of the applied GA was transported to the fruit, depending on dose. There was a linear relationship between GA₁ content and pod weight up to about 2 ng · (g FW)⁻¹, whereas no such correlation existed for GA₃ content. The concentration of endogenous GA₁ in pods from pollinated ovaries is just sufficient to give the maximum growth response. It is concluded that GA₁, but not GA₃, controls pod growth in pea; GA₃ may be involved in early seed development. The distribution of GAs within the seeds at 4 d post anthesis was also investigated. Most of the GA₁, GA₈, GA₁₉, GA₂₀ and GA₂₉ was present in the testa, whereas GA₃ was distributed equally between testa and endosperm and GA₄ was localised mainly in the endosperm. Of the GAs analysed, only GA₃ and GA₂₀ were detected in the embryo. Metabolism experiments with intact tissues and cell-free fractions indicated compartmentation of GA biosynthesis within the seed. Using ¹⁴C-labelled GA₁₂, GA₉, 2,3-didehydroGA₉ and GA₂₀ as substrates, the testa was shown to contain 13-hydroxylase and 20-oxidase activities, the endosperm, 3β-hydroxylase and 20-oxidase activities. Both tissues also produced 16,17-dihydrodiols. However, GA₁ and GA₃ were not obtained as products and it is unlikely that they are formed via the early 13-hydroxylation pathway. [¹⁴C]gibberellin A₁₂, applied to the inside surface of pods in situ, was metabolised to GA₁₉, GA₂₀, GA₂₉, GA₂₉-catabolite, GA₈₁ and GA₉₇, but GA₁ was not detected. Gibberellin A₂₀ was metabolised by this tissue to GA₂₉ and GA₂₉-catabolite. Received: 23 July 1996 / Accepted: 2 September 1996