Why Collagen Behaves Like Liquid Droplets Inside Cells

When a postdoc in Barcelona first looked through the microscope and saw bright, spherical clusters, no one expected they were watching the body’s most abundant protein break every rule cell biologists had lived with for decades.

Collagen has long been pictured as a rigid molecular cable — a stiff rod that bundles into the fibres giving skin, bone and tendons their tensile strength. That image is correct outside the cell. But inside, where collagen is born, new live-cell imaging from the Center for Genomic Regulation (CRG) in Barcelona reveals a different reality: collagen assembles as soft, liquid-like droplets that drift, merge and split like oil in water.

The discovery is not decorative. It solves an old size problem. Purified collagen can stretch hundreds of nanometres long, while the vesicles that shuttle proteins through the secretory pathway are only a few dozen nanometres across. How did huge rods squeeze through tiny carriers? Maybe they never do. Instead, the precursor procollagen—especially type I, which makes up the lion’s share of the body’s collagen—appears to form condensates inside the endoplasmic reticulum (ER), concentrated liquid compartments that behave like droplets.

Human liver cells showing collagen droplets inside the cell (green clusters), held in place by TANGO1 (magenta), with extracellular collagen fibres visible as the surrounding network. Cell nuclei are stained blue. 

Watch these droplets and they display classic condensate behavior: they flow, they fuse, they exchange molecules with their environment. They recruit specific chaperones that recognize correctly folded collagen, not the hallmarks of misfolded junk destined for the cell’s disposal. In other words, these are functional packing units, not trash heaps.

That neat observation forced the team to rethink how collagen gets exported. If collagen is a droplet rather than a rigid rod, it needn’t be shoved into a tiny vesicle. Instead, the researchers propose a "liquid extrusion" model: imagine a viscous bead of collagen perched against an exit site on the ER. Physical forces—capillary action or a squeeze-like transfer—could let the liquid phase flow through the opening and into the next compartment. Simple physics, applied to biology.

TANGO1, a protein discovered in the same lab two decades ago, steps into the story as a mooring line. When TANGO1 is removed, the droplets still form, but they no longer sit where cargo departs the ER, and secretion drops. That suggests TANGO1 does not act as a classical receptor that latches onto a single molecule; it seems to anchor the whole droplet at the exit site so the condensate can spread and be handed off.

The image under the microscope that led to the discovery, showing liver cells producing collagen at high levels, with droplet-like structures visible in green. 

Inside the cell, collagen is pliable — a liquid condensate rather than a pre-formed fibre.

There is elegance in this picture. A liquid state protects the cell. If collagen polymerized into rigid fibres inside the ER it would clog the factory and kill the cell. By keeping procollagen fluid and sequestered, the cell can safely organize very large structural cargo until it reaches the extracellular stage, where assembly into stiff fibres is desirable and life-giving.

The implications reach beyond basic curiosity. Fibrosis — whether in liver, lung or skin — and certain cancers depend on excessive collagen deposition. Tumors build dense collagen matrices that shield cancer cells from chemotherapy and immune attack. If collagen export depends on droplets and their interaction with TANGO1, then new therapeutic levers appear: destabilize the condensates, interfere with their mooring, or otherwise block the extrusion process. All speculative now, but compelling.

The Barcelona team is cautious. The liquid extrusion idea is a hypothesis grounded in imaging and biophysics; it still needs direct visualization of the transfer process and validation in whole tissues. The next steps are clear: stretch the experiments into mouse models, watch collagen export in living tissue, and probe whether interfering with droplet formation alters fibrosis or tumor stiffness.

Science sometimes advances by changing the way we picture things. For sixty years, the mental image of collagen has been a rope. Today, that rope looks more like a droplet — a soft, organized package that lets cells do their work without self-sabotage. If this view holds, it opens a new angle on diseases that hinge on collagen’s excess, and it asks a simple, modern question: can we un-stick biology by targeting physics?