Gibberellins and Stem Elongation

Gibberellins are a family of tetracyclic diterpenoid acids, all built on the ent-gibberellane carbon skeleton, first isolated from the rice pathogenic fungus Gibberella fujikuroi (the sexual stage of Fusarium fujikuroi) in the 1930s–1950s. Rice infected by this fungus grows abnormally tall and spindly — "bakanae," or "foolish seedling," disease — because the fungus secretes gibberellins that hijack the plant's own growth-control system. More than 130 distinct gibberellins (numbered GA1, GA2, GA3 … GA136) have since been identified across plants, fungi, and even some bacteria, but the vast majority of these are biosynthetic intermediates or inactivation products. Only a small subset — principally GA1 and GA4, and in some species GA3 (commercially known as gibberellic acid) — are bioactive, meaning they can bind the plant's GA receptor and trigger downstream signaling. The central physiological output of gibberellin signaling that this article focuses on is internode elongation: the growth in length of the stem segments between leaf-bearing nodes. GA promotes elongation by stimulating both cell division in the sub-apical intercalary meristem and, especially, cell elongation, in part by loosening cell walls and reorienting cortical microtubules (and thus cellulose microfibril deposition) so that cells expand longitudinally rather than radially. The same signaling module also drives bolting — the rapid elongation of a normally short flowering stem (as in rosette plants like Arabidopsis and many biennials such as cabbage and spinach) that converts a compact vegetative rosette into a tall inflorescence axis, often gated by day length or vernalization. The molecular logic that makes all of this work is derepression. Rather than gibberellin directly switching on elongation genes, it works by destroying a class of nuclear proteins called DELLA proteins, named for a conserved Asp-Glu-Leu-Leu-Ala motif near their N-terminus. DELLAs are constitutive brakes on growth: in the absence of GA they accumulate and bind (and inhibit) transcription factors needed for cell elongation, most notably members of the PIF (Phytochrome-Interacting Factor) bHLH family. When GA levels rise, GA is bound by its receptor and the GA–receptor complex recruits DELLA for ubiquitin-mediated proteasomal destruction. With DELLA gone, PIFs and other growth-promoting transcription factors are freed to activate expansin and other cell-wall-loosening genes, and the cell elongates. This receptor-triggered-destruction mechanism is shared conceptually with auxin signaling, where TIR1 similarly recruits Aux/IAA repressors for destruction — a recurring theme of "hormone-triggered proteolysis of a repressor" in plant signal transduction.

Step 1 — Biosynthesis: from GGPP to bioactive GA. Gibberellin biosynthesis begins in plastids with the universal terpenoid precursor geranylgeranyl diphosphate (GGPP), a 20-carbon isoprenoid. Two plastid-localized terpene cyclases convert GGPP to ent-kaurene: 1. GGPP → ent-copalyl diphosphate, catalyzed by ent-copalyl diphosphate synthase (CPS) 2. ent-copalyl diphosphate → ent-kaurene, catalyzed by ent-kaurene synthase (KS) ent-Kaurene then moves to the endoplasmic reticulum, where two cytochrome P450 monooxygenases oxidize it in successive steps — ent-kaurene oxidase (KO) converts it to ent-kaurenoic acid, and ent-kaurenoic acid oxidase (KAO) further oxidizes it to GA12, the first true gibberellin and the branch point of the pathway. From GA12, cytosolic 2-oxoglutarate-dependent dioxygenases carry out the final activation steps:

• GA 20-oxidase (GA20ox) performs sequential oxidations converting GA12 (and GA53, its 13-hydroxylated derivative) toward GA9 and GA20 — this is widely considered the key rate-limiting, regulatory step of the pathway.
• GA 3-oxidase (GA3ox) then hydroxylates GA9 and GA20 at the C-3 position to yield the bioactive hormones GA4 and GA1, respectively.

Deactivation is equally important for maintaining the right hormone concentration: GA 2-oxidase (GA2ox) hydroxylates active GAs (and precursors) at C-2, converting them into inactive forms that can no longer bind the receptor. This synthesis–deactivation balance, tuned by developmental and environmental signals (light, temperature, feedback from GA signaling itself), sets the local concentration of bioactive GA in the elongating stem.

Step 2 — Perception: the GID1 receptor. Bioactive GA is perceived by a soluble nuclear/cytoplasmic receptor called GID1 (GIBBERELLIN INSENSITIVE DWARF1), a protein related to carboxylesterases but catalytically inactive as an esterase. GA binds inside a hydrophobic pocket of GID1, causing a "lid" loop to close over the ligand. This GA-induced conformational change creates a new protein-protein interaction surface on GID1's exterior — critically, one that is only present when GA is bound.

Step 3 — DELLA recruitment and destruction. The GA-GID1 complex, through its newly exposed surface, binds the DELLA domain at the N-terminus of DELLA proteins (in Arabidopsis: GAI, RGA, RGL1, RGL2, RGL3; in rice, the single ortholog SLR1). This ternary complex (GA–GID1–DELLA) is recognized by an F-box protein — SLY1 in Arabidopsis, GID2 in rice — that is part of an SCF (SKP1–CUL1–F-box) E3 ubiquitin ligase complex. The F-box protein binds DELLA and the SCF complex polyubiquitinates it. Polyubiquitinated DELLA is then degraded by the 26S proteasome.

Step 4 — Release of growth-promoting transcription factors. DELLA proteins have no DNA-binding domain of their own; they act by physically sequestering and inhibiting other transcription factors, chiefly PIFs (phytochrome-interacting factors, bHLH transcription factors) and components of brassinosteroid signaling (BZR1). With DELLA destroyed, PIF4 and related factors are free to bind E-box promoter elements of target genes, including genes encoding expansins and other cell-wall-modifying enzymes, and genes controlling cell-cycle progression in the sub-apical meristem.

Step 5 — Cellular growth response. Expansins loosen the hydrogen bonds between cellulose microfibrils and the hemicellulose matrix, allowing turgor-driven wall extension. GA signaling also reorients cortical microtubules from transverse to more longitudinal/oblique arrays in some systems, biasing the direction of cellulose deposition and hence the direction of cell expansion toward the long axis of the stem — producing the elongated, narrow internodes characteristic of high-GA growth.

Step 6 — Negative feedback. GA signaling feeds back on its own biosynthesis: high GA (via DELLA destruction) represses transcription of GA20ox and GA3ox genes and can induce GA2ox, while low GA (DELLA accumulation) does the opposite. This homeostatic loop keeps bioactive GA concentration within a physiologically appropriate range rather than spiraling unchecked.

Step 7 — Environmental and developmental gating. Light quality (via phytochrome), photoperiod, temperature, and vernalization state all modulate GA20ox/*GA3ox* expression, which is why bolting can be triggered by long days or cold exposure in many rosette species — GA biosynthesis genes are transcriptionally induced as part of the floral transition pathway, converting a compact vegetative rosette into an elongated flowering stalk.

Discovery via a plant disease (Kurosawa, 1926; Yabuta, 1935–1938). Japanese plant pathologist Eiichi Kurosawa showed in 1926 that a cell-free filtrate from cultures of the fungus Gibberella fujikuroi — the causal agent of "bakanae," or foolish-seedling disease, in rice — could reproduce the abnormal elongation and etiolation of infected rice seedlings even without the fungus itself being present. This proved the fungus secretes a diffusible chemical growth stimulant rather than causing elongation through direct tissue invasion. Teijiro Yabuta and colleagues later purified the active substance and named it "gibberellin" (1935; crystalline gibberellin A isolated 1938). Western scientists largely learned of this work only after WWII; British (ICI) and American (USDA) groups independently isolated gibberellic acid (GA3) from the same fungus in the mid-1950s, confirming Kurosawa and Yabuta's findings and beginning the modern era of gibberellin chemistry.

Application-and-rescue experiments defining GA's role as an endogenous plant hormone. A pivotal question was whether GA was merely a fungal weapon or whether plants make and use it themselves as a bona fide hormone. Peter Brian and Hazel Hemming (Imperial Chemical Industries, UK, early 1950s) applied purified gibberellic acid to intact, non-infected dwarf and tall pea and bean varieties. Dwarf peas treated with GA3 grew to heights statistically indistinguishable from untreated tall varieties, while GA had comparatively little further effect on already-tall genotypes. This dose-dependent rescue of a specifically dwarf phenotype — not a general non-specific growth stimulant effect — was strong evidence that dwarfism in these varieties results from a deficiency somewhere in the endogenous GA pathway, and that GA is normally synthesized and required by the plant itself, not merely as a curiosity of fungal pathology.

Biochemical mapping of the dwarfing mutations (1980s–1990s, using GC-MS quantification of endogenous GAs). Once gas chromatography–mass spectrometry made it possible to measure femtomole quantities of specific GA species in plant tissue, researchers led by groups including Bernard Phinney, Jan Zeevaart, and Peter Hedden directly measured which GAs accumulated or were missing in classic dwarf mutants:

This table itself is the key evidentiary structure: mutants that are dwarf because they cannot make GA (e.g., pea le, rice d1) are rescued by spraying on GA, because the receptor/signaling machinery downstream is intact and simply needs ligand. Mutants that are dwarf because they cannot perceive or transduce the GA signal (wheat Rht, rice gid1, stabilized DELLA alleles) are not rescued by GA, because the defect lies downstream of hormone production. This genetic dissection — biosynthesis mutants rescuable, perception/signaling mutants not rescuable — is the classic logic used throughout plant hormone biology to order a pathway, and it is precisely how the GID1–DELLA module was placed downstream of GA synthesis with confidence.

Molecular proof of the degradation mechanism (2003–2008). Direct biochemical and cell-biological experiments confirmed the destruction model. Ueguchi-Tanaka and colleagues (2005, Nature) showed that rice GID1 protein binds bioactive GA with high affinity in vitro and that this GA-binding is required for GID1 to physically interact with the DELLA protein SLR1, using yeast two-hybrid and pull-down assays. Fluorescently tagged DELLA-GFP fusion proteins, imaged in living root or stem cells, visibly disappear from the nucleus within minutes of GA application and this disappearance is blocked by proteasome inhibitors (e.g., MG132) — directly demonstrating GA-triggered, proteasome-dependent DELLA destruction in real time, and confirming that sly1 and gid2 F-box mutants fail to degrade DELLA even when GA is supplied, placing SLY1/GID2 as the necessary ubiquitin-ligase component of the switch.

Example 1 — Diagnosing a dwarf mutant by GA-application phenotyping. A plant breeder isolates a new short-statured pea mutant, call it dw-x, from a mutagenized population. Two candidate explanations exist: (a) dw-x blocks GA biosynthesis, or (b) dw-x disrupts GA perception/signaling (e.g., a stabilized DELLA). The breeder sprays dw-x seedlings with GA3 (10⁻⁴ M) weekly for three weeks and measures internode length against untreated dw-x and untreated tall wild-type controls.

• If GA3-treated dw-x internodes elongate to within ~90–100% of wild-type tall length → the lesion is most likely in biosynthesis (like Mendel's le gene, a defective GA3-oxidase); exogenous hormone substitutes for the missing endogenous supply.
• If GA3-treated dw-x shows little or no elongation response (say, <15% of the wild-type–untreated difference) → the lesion is most likely in perception or signaling downstream of GID1 (e.g., a stabilized, non-degradable DELLA, analogous to wheat Rht-B1b), because no amount of added hormone can overcome a defective receptor or an indestructible repressor.

Suppose the actual data show untreated dw-x at 8 cm, GA-treated dw-x at 9 cm, and wild-type at 42 cm. Because the GA-treated mutant recovered only (9−8)/(42−8) ≈ 3% of the wild-type/mutant height gap, the breeder correctly infers a signaling-pathway lesion, not a biosynthetic block — and further work should sequence the DELLA ortholog rather than the GA20-oxidase/GA3-oxidase genes.

Example 2 — A dihybrid cross combining a GA-biosynthesis mutation and a GA-signaling mutation. Consider a cross designed to test epistasis between a biosynthesis gene and a signaling gene in a diploid model species. Gene A encodes GA3-oxidase (biosynthesis); the recessive loss-of-function allele a blocks GA1 production, producing dwarfism (as in Mendel's pea). Gene B encodes a DELLA protein; a rare dominant gain-of-function allele Bᵈ produces a DELLA that resists degradation even when GA is present (analogous to wheat Rht-B1b); the recessive wild-type allele b is normally degraded upon GA-GID1 binding. Cross a plant heterozygous at both loci, genotype AaBᵈb (tall, because one functional A allele makes enough GA1, and one wild-type b allele allows normal DELLA turnover in that copy — assume Bᵈ is fully dominant when present) × a double-heterozygote of the same genotype, AaBᵈb × AaBᵈb. Standard Mendelian dihybrid segregation with independent assortment gives:

Collecting phenotypes: dwarf = 9/16 + 3/16 + 1/16 = 13/16; tall = 3/16. This 13:3 ratio is a classic epistatic ratio (a modified 9:3:3:1) arising because the dominant, gain-of-function Bᵈ allele is epistatic to genotype at the A locus — a plant can only be tall if it both makes GA (has at least one A allele) and can degrade DELLA (is bb, homozygous wild-type at the signaling locus). This mirrors the real biology of the pathway: signaling defects downstream of hormone synthesis override the effect of hormone availability, exactly as seen when GA-insensitive wheat Rht alleles are combined with otherwise normal GA-biosynthetic backgrounds.

Eiichi Kurosawa was a Japanese plant pathologist working in Taiwan (then under Japanese colonial administration) in the 1920s, studying "bakanae" disease of rice — a condition known to farmers for centuries in which infected seedlings grow abnormally tall, pale, and spindly before collapsing and dying. Kurosawa's decisive experiment, published in 1926, was to filter culture broth from Gibberella fujikuroi through a fine filter that excluded the fungal cells themselves, then apply only the cell-free filtrate to healthy rice seedlings. The seedlings still elongated abnormally. This was the critical control: it proved the fungus was secreting a soluble chemical agent responsible for the elongation phenotype, rather than the elongation resulting from physical invasion, nutrient theft, or a hypersensitive plant immune response to the pathogen's physical presence. Kurosawa's work initially circulated mainly in Japanese-language phytopathology journals and was not widely known outside Japan for decades; Teijiro Yabuta and Yusuke Sumiki subsequently purified and crystallized the active compound in the 1930s, coining the name "gibberellin." Because of wartime isolation, Western agricultural chemists (British and American groups) independently rediscovered and characterized gibberellic acid only in the mid-1950s, initially astonished that a fungal secretion could so precisely mimic — and rescue — a plant's own growth-hormone deficiency.

The Green Revolution and global food security. Semi-dwarf wheat varieties bred by Norman Borlaug's program carry the Rht-B1b/*Rht-D1b GA-insensitivity alleles; semi-dwarf "IR8" rice and its descendants carry reduced-function sd1* (GA20-oxidase) alleles. Both produce short, stiff stems that resist lodging (falling over) under the weight of heavy, fertilizer-boosted grain heads, allowing nitrogen fertilizer use without catastrophic yield loss — a change credited with substantially increasing global cereal yields in the mid-to-late 20th century and averting famine across parts of Asia and Latin America.

Commercial horticulture and viticulture. GA3 is sprayed on seedless grape varieties (e.g., Thompson Seedless) to enlarge berry size and elongate clusters, compensating for the reduced cell-division stimulus normally provided by developing seeds. GA is also used to promote stem elongation and flowering in ornamental bulbs, to break dormancy and stimulate more uniform malting in barley for the brewing industry (endogenous GA induces α-amylase in the aleurone layer during germination, connecting directly to Seed Dormancy and Germination), and to speed bolting/flowering in biennial vegetable seed production, avoiding a full two-year field cycle.

Turf and ornamental management. GA-biosynthesis inhibitors (triazoles such as paclobutrazol, and related compounds) are applied agriculturally and horticulturally to deliberately shorten stems, reduce lodging risk, and produce compact ornamental plants — a direct practical exploitation of the biosynthetic pathway described above.

Plant breeding and biotechnology. Understanding the GID1–DELLA switch lets breeders rationally choose whether to target biosynthesis genes (GA20ox, GA3ox) or signaling genes (DELLA, GID1) when engineering stature, disease response (DELLA also modulates defense signaling crosstalk with jasmonate and salicylic acid pathways), or stress tolerance, since DELLA accumulation is also part of how plants slow growth under adverse conditions (a growth-restraint strategy sometimes framed as a "growth-survival trade-off").

Connections forward. The same GA–DELLA logic reappears in seed biology, where GA synthesized in the embryo overcomes ABA-imposed dormancy and induces hydrolytic enzymes for reserve mobilization (see Seed Dormancy and Germination), and in the transition to flowering, where cold-induced Vernalization pathways interact with GA biosynthesis to trigger bolting in biennial and rosette species.

Ask the tutor: How does the GA-DELLA switch interact with the vernalization and photoperiod pathways to control bolting time in biennial plants?