Apical Constriction: The Actomyosin Purse-Strings That Fold Epithelial Sheets

What apical constriction is and where it happens

Epithelial cells are polarized: they have an apical face (pointing to the outside or a lumen), a basal face (anchored to basement membrane), and lateral sides joined to neighbors by adherens junctions. Apical constriction is the selective shrinking of the apical face while the basal face keeps its area, converting a roughly columnar or cuboidal cell into a wedge (a truncated pyramid).

When many neighboring wedges tile together in a sheet, geometry forces the sheet to curve, the apical side becoming the concave inner surface of a fold. This is the workhorse mechanism of tissue bending across animals:

• Gastrulation — the Drosophila ventral furrow invaginates the mesoderm; sea-urchin and C. elegans gut precursors ingress the same way.
• Neural tube closure — hinge-point cells in the vertebrate neural plate constrict apically to roll the plate into a tube.
• Placodes and tubes — lens, salivary gland, and airway buds pinch inward via apically constricting cells.

It happens at the apical cortex, a thin (~100–500 nm) actin-rich shell just under the apical membrane.

The mechanism, step by step

The contractile engine is non-muscle myosin II pulling on filamentous actin (F-actin), wired to the cell edges through adherens junctions. Force generation proceeds in stages:

• Signal to constrict. A patterning input (e.g. transcription factors Twist and Snail in the fly mesoderm) turns on secreted Folded gastrulation (Fog) and the transmembrane protein T48, which recruit RhoGEF2 to the apical cortex.
• Rho activation. RhoGEF2 loads GTP onto Rho1 (RhoA). Active Rho-GTP switches on Rho-kinase (Rok/ROCK).
• Myosin activation. Rok phosphorylates the myosin regulatory light chain (Ser19/Thr18) and inhibits myosin phosphatase, so myosin II assembles into bipolar minifilaments and starts walking along actin.
• Contraction. Myosin pulls the apical actin meshwork inward. Because the network is anchored at cell–cell junctions, shortening the meshwork shrinks the apical perimeter.
• Ratcheting. Each contraction is a pulse; between pulses, junctional and cortical elements lock in the gained area so the cell can't relax back — a molecular ratchet.

Key molecules and characteristic numbers

The pulsatile, ratchet-like nature of the process was quantified by Adam Martin, Matthias Kaschube and Eric Wieschaus (2009) in the Drosophila ventral furrow. Their live imaging showed a medioapical (apicomedial) actomyosin network — a meshwork in the center of the apical surface, distinct from the junctional belt — that assembles and contracts in bursts.

• Non-muscle myosin II — the motor; heavy chain Zipper in flies, MYH9/MYH10 in humans; hexamer with two heavy chains, two regulatory and two essential light chains.
• F-actin — the track and the load-bearing cable.
• Rho1/RhoA, RhoGEF2, Rok, myosin phosphatase — the on/off control module.
• E-cadherin, α- and β-catenin, Canoe/Afadin — couple the network to junctions so force transmits between cells.

Characteristic values: apical area falls from ~25 µm² toward <5 µm²; contractile pulses recur every ~60–90 s and each shrinks the apex by a few percent; ventral-furrow invagination completes in roughly 5–10 minutes; myosin minifilaments are ~300 nm bipolar assemblies.

How it is studied, observed, and regulated

Apical constriction is a flagship system for quantitative cell biology because it is fast, geometrically simple, and genetically tractable.

• Live confocal / light-sheet imaging of GFP-tagged myosin (Sqh-GFP) and actin (Utrophin- or Lifeact-based probes) reveals the pulses and flows in real time.
• Particle-image velocimetry and PIV / kymographs quantify cortical flow and contraction rates.
• Laser ablation of the cortex measures tension: the recoil velocity of severed junctions reports on force.
• Optogenetics (light-gated recruitment of RhoGEF or the CRY2/CIBN system) can switch contraction on or off with subcellular precision, showing that local myosin activation is sufficient to constrict.
• Genetics — mutants in fog, T48, concertina (a Gα), RhoGEF2, rok, and sqh block or disorganize the furrow.

Regulation is spatial and temporal: myosin must be concentrated medioapically, contraction must be pulsatile, and gains must be ratcheted. Loss of the ratchet (e.g. reduced junctional coupling) leaves cells pulsing without net constriction.

How it differs from related processes

Several morphogenetic mechanisms use the same actomyosin toolkit but produce different outcomes, and confusing them is a common error.

• Basal constriction contracts the opposite (inner) surface, bending the sheet the other way — as in the zebrafish optic cup. Same motors, mirror geometry.
• Convergent extension uses planar-polarized junctional actomyosin to make cells slide past one another (intercalation), narrowing and lengthening tissue rather than bending it.
• Apical–basal cell elongation relies on microtubules to make cells taller; it thickens the sheet and can assist bending but is not contraction-driven.
• The cytokinetic contractile ring is a transient equatorial actomyosin ring that pinches one cell in two; apical constriction is a persistent apical network that reshapes a whole tissue.

The key distinctions are which surface contracts, whether the network is medioapical or purely junctional, and whether the goal is bending, intercalation, or division.

Significance, disease relevance, and open questions

Because apical constriction bends nearly every epithelial fold in development, its failure causes major birth defects. Incomplete neural tube closure produces spina bifida and anencephaly — among the most common human structural malformations (~1 in 1,000 pregnancies) — and mouse mutants in Shroom3 (a myosin/ROCK adaptor), RhoA, and cytoskeletal regulators show neural-tube defects. Aberrant apical constriction is also implicated in cleft palate, gut and heart-tube defects, and in cancer, where dysregulated Rho/ROCK–myosin contractility alters epithelial invasion.

Open questions remain lively:

• What exactly ratchets each pulse — junctional remodeling, actin turnover, or a mechanosensitive feedback loop?
• How do cells coordinate pulses across a tissue so contraction is supracellular rather than random?
• How is the medioapical network mechanically connected to junctions to transmit force without tearing them?
• How much of tissue bending comes from apical constriction itself versus buckling driven by surrounding tissue and the vitelline membrane?

Optogenetic and physical-modeling approaches are steadily turning these into quantitative answers.

Apical constriction versus other epithelial shape-change and contraction mechanisms

Process	What contracts / changes	Net tissue outcome	Example
Apical constriction	Medioapical + junctional actomyosin shrinks the apical surface	Cells wedge; sheet bends/invaginates	Drosophila ventral furrow; vertebrate neural plate
Convergent extension	Cells intercalate via planar-polarized junctional actomyosin	Tissue narrows and lengthens	Xenopus/zebrafish axis elongation
Basal constriction	Actomyosin contracts the basal (inner) surface	Sheet bends the opposite way	Zebrafish optic cup, Drosophila wing disc folds
Apical-basal cell elongation	Microtubule-driven columnar shape change	Sheet thickens, can drive bending	Neural plate thickening
Cytokinetic ring	Equatorial actomyosin ring at cell midzone	One cell divides into two	Every dividing animal cell