Coagulation
This is the fork the whole craft hangs off, and the page I’d most want a beginner to read.
Milk is stable because every casein micelle wears a coat of protruding κ-casein hairs that keep the micelles apart — by charge and by physical bulk. Making cheese means defeating that. There are three ways, they are chemically different, and they give you three different materials.
1. Acid: take away the charge
Add acid — vinegar, lemon juice, or the lactic acid your bacteria make from lactose — and the pH falls. As it falls, the charge on the casein falls with it.
At pH 4.6, casein’s isoelectric point, the net charge reaches zero. Nothing is pushing the micelles apart any more, so they collide, stick, and drop out of suspension.
What you get is a precipitate: casein that has clumped and fallen. Soft, fragile, granular, delicate. It can be drained and eaten and it is delicious, but it will not stretch, will not melt properly, and will not age.
This is cottage cheese, quark, fromage blanc, and the quick vinegar mozzarella that was my first cheese. It costs nothing and needs no equipment beyond a strainer — which is exactly why it is plausibly the oldest branch of the craft.
2. Rennet: cut the coat off
Rennet does something completely different, and far more elegant.
The active enzyme, chymosin, is a scalpel. It cleaves one specific bond in κ-casein — between residues phenylalanine 105 and methionine 106 — and essentially nothing else. That single cut splits the molecule in two:
- para-κ-casein (residues 1–105): hydrophobic, and it stays on the micelle surface;
- caseinomacropeptide (residues 106–169), also called glycomacropeptide: hydrophilic, charged, and it floats away into the whey.
The part that leaves is the part that was sticking out. The micelle has been shaved.
And now the second phase begins, which is not enzymatic at all. Stripped of their hairs, the micelles have lost the repulsion holding them apart, and in the presence of milk’s calcium they begin bonding to each other — not clumping randomly, but linking into chains and then a continuous three-dimensional network that fills the entire vat.
So renneting is two stages: an enzymatic phase (which proceeds around pH 6.4–6.5, i.e. in essentially fresh milk), then an aggregation phase. This is why nothing appears to happen for a while and then the whole vat sets at once — the enzyme has to do its work across the population before the aggregation can run away.
And the result is a gel, not a precipitate. A single connected structure you can cut with a knife and watch heal at the cut face. It can be cooked, stirred, pressed, and aged for years. Every hard cheese on earth is on this side of the line.
🔴 A correction to my own earlier shorthand
On what cheese actually is and the family tree I said rennet “snips off the piece of protein that carries the charge.” That is not wrong, but it’s incomplete in a way worth fixing here.
The caseinomacropeptide is charged and it physically protrudes. The literature generally emphasises the steric contribution — the removal of the fragment causes “a substantial decrease in the steric repulsion forces” between particles. So it’s not purely an electrical effect: you are removing a physical brush as much as a charge.
The reason this matters practically is that it explains why the two routes can’t substitute for each other. Acid neutralises a charge and lets particles fall together. Rennet removes an obstacle and lets an intact, still-charged protein build a structure. Different mechanism, different material.
3. Heat-acid: recruit the whey proteins
The third route is usually taught as a variant of the first, and it isn’t.
Heat milk hard — above about 65 °C, and in practice much hotter — and the whey proteins denature. They unravel, exposing sticky regions that were folded away inside. Those exposed regions bond to each other and to casein. Add acid at that point and you coagulate both protein populations at once, into a combined whey-and-casein matrix.
Why that’s a genuinely different thing:
It captures protein the other routes throw away. In a normal rennet cheese, that 20% of milk protein that is whey protein drains off and is lost. Heat-acid coagulation incorporates it, which increases yield. It’s why heat-acid cheeses feel like they punch above their milk.
And it’s why true ricotta works at all. Whey that has already surrendered its casein to a cheddar still contains its whey protein. Heat it and acidify it and you get a second, smaller curd from what you were about to pour away — the “recooking” that gives ricotta its name.
This is the route for paneer, queso fresco, many fresh Indian and Latin American cheeses, and the whey cheeses. And it connects back to the wider world: the West African wagashi coagulant, Calotropis, sets milk hot, at 65–70 °C — which is the same temperature territory, arrived at independently with a plant protease.
The three, side by side
| Acid | Rennet | Heat-acid | |
|---|---|---|---|
| Mechanism | Charge → zero at pH 4.6 | One bond cut; steric brush removed | Whey proteins denatured, then acid |
| Product | A precipitate — clumps | A gel — continuous network | Combined casein + whey matrix |
| Captures whey protein? | No | No | Yes |
| Can be cut, cooked, pressed? | Barely | Yes | Somewhat |
| Ages? | No | Yes, for years | No |
| Examples | Cottage, quark, quick mozzarella | Everything hard | Ricotta, paneer, queso fresco |
The practical upshot
Three things follow that will save you a ruined batch:
You cannot make cheddar with vinegar. Not “it comes out worse” — the material required does not exist on that side of the line. Acid can’t build a network.
pH matters at both ends and for different reasons. Too much acid before you add rennet and you’re partway down the acid route already, so you get a mixed, weak, brittle curd. This is why the recipe’s pH targets are instructions and its times are estimates.
And calcium is a required reagent, not a nutrient. The aggregation phase needs it. Pasteurisation shifts some out of solution, which is why store milk gives a weaker set and why calcium chloride belongs in the recipe.
Next: rennet — where it comes from, why the vegetable kind turns bitter in an aged cheese, and what IMCU means on the bottle.
Sources
- Chymosin’s specific cleavage of the Phe105–Met106 bond of κ-casein, producing para-κ-casein (1–105) and caseinomacropeptide (106–169); the two-phase enzymatic-then-aggregation mechanism at pH ~6.4–6.5; and the “substantial decrease in the steric repulsion forces” that follows removal of the fragment: ScienceDirect — renneting and ScienceDirect — chymosin
- Casein’s isoelectric point at pH 4.6 and acid coagulation; whey-protein denaturation above 65 °C, the exposure of sticky regions, co-precipitation with casein and the resulting yield increase: Cheese Science Toolkit — coagulation and University of Guelph, Dairy Science and Technology
Comments (0)