Live cell imaging at a Barcelona research center shows that the precursor to collagen, the molecule that makes up roughly a third of human protein mass, assembles inside cells as liquid like droplets rather than rigid rods.
In Soumya Bhattacharyya's microscopy images, collagen's building blocks appear as something textbooks rarely depict: green, droplet-like blobs pooling inside living liver cells. The finding, reported by Genetic Engineering and Biotechnology News, comes from a team at the Centre for Genomic Regulation in Barcelona and inverts the standard picture of how the body's most abundant structural protein gets organized at its source.
Collagen accounts for roughly a third of total protein mass in the human body and provides the scaffolding for skin, bone, tendon, and most connective tissue. The traditional view treats it as a long, rigid rod assembled in the endoplasmic reticulum (the cell's internal protein-folding compartment) and then secreted outward. The CRG team, led by Bhattacharyya, used high-resolution live-cell imaging to watch the precursor form, procollagen 1, behave instead like a liquid: pooling into membraneless droplets inside liver cells.
That behavior is consistent with a class of cellular structures that biologists have come to call biomolecular condensates. The term, developed over the last decade, describes how proteins and other molecules can organize themselves into concentrated, gel-like or liquid-like compartments without a surrounding membrane. The familiar analogy is oil separating from water. The same physics, in cells, lets proteins cluster where they need to act and stay out of the way when they don't.
Whether procollagen 1 truly behaves this way is still the team's interpretation, made on the basis of imaging rather than a full mechanistic dissection. The condensate and phase-separation literature is active and sometimes contested, with researchers still debating which observations count as definitive evidence of a liquid-like state. The CRG work is a single observation, not a settled mechanism.
The scope also matters. The droplets are procollagen 1, the immature precursor made inside the endoplasmic reticulum, not the mature collagen that later assembles into fibrils outside the cell. Mature collagen's rod-like character in the extracellular matrix is not in dispute. What is new is the suggestion that the cell organizes the precursor's production through the same phase-separation logic that has been documented for signaling proteins, stress-granule components, and a growing list of other molecules.
That reframing matters because collagen production is, by volume, the cell's most demanding structural task. If the cell uses condensates to manage that volume, then diseases that involve too much or too little collagen production, including fibrosis, scarring, and certain connective tissue disorders, may be readable as failures of condensate formation, dissolution, or quality control. The hypothesis is still speculative, and no therapeutic strategy has yet grown out of it, but it gives researchers a new place to look.
What to watch next is whether other groups reproduce the droplet-like behavior in additional cell types, whether the CRG team or others can perturb candidate condensate-driving features (specific disordered regions of the procollagen 1 sequence, for example) and see production change, and whether the lens survives contact with the protein's transition from precursor to mature fibril outside the cell.