Autolysis to the Autolysis!

Q. I am a beginner and have been doing a lot of reading, everyone else seems to be saying autolyse helps to develop gluten, so it's not true? So it's not a necessary step? Please elaborate if you have the time.

You’ve already correctly answered your own question. (I’m not sure who “everyone else” is, but, as above, “If someone’s teaching you about the ‘science of bread,’ assume that whatever they say is wrong.”)

Presume what “they” claim is true, that “autolyse helps to develop gluten.” You simply need not knead, and, given enough time, you’ll have a (fully?) developed dough.

On its face, it’s a ridiculous assertion. (As an aside, Jim Lahey’s no-knead bread is a fantastic way to make bread at home. Minimal effort for maximal joy. It’s the bread my wife, Pip, bakes for our daughters, li’l polysaccharide gremlins capable of decimating an entire 1,200-gram loaf in a day. I’m pretty sure my middle daughter, Aster, 4, can do this on her own. To be fair, she leaves the crust, or “toast bones” as I call ‘em. As Pip says, “Why don’t more home bakers make this bread? Why do they have to over-complicate things?”) Why?

Let’s begin with the historical context. The term and method we know today as “autolysis” or as an “autolyse step” was created by Raymond Calvel, a French bakery teacher. After a bad harvest, France had to import strong, hard North American wheat, which Calvel (and I) believed inappropriate for French bread applications. The flour made for a tight, tough product with diminished volume. His goal, as a researcher, was to find a way to weaken the flour so as to attain a soft dough.

His solution was “autolysis,” allowing a barely-mixed combination of only flour and water to sit at ambient temperature for a specified period of time. (He tested durations up to 24 hours.) He chose the term “autolysis” because he believed the processing step degraded a significant amount of gluten. (It does not.)

Why did he believe this? Because he observed a weakening effect, resulting in, according to him, increased dough extensibility. (It does not.) For the rest of his life he became a vocal proponent of autolysis. Nowhere did he claim it increases gluten development. (It does not.)

Now, the theoretical background.

“Gluten” is a thermodynamic classification, an arbitrary grouping of plant proteins based upon solubility characteristics. Researchers create ways of sorting large, heterogeneous data sets, most often according to how their measuring device(s) interacts with the data set. This makes complicated things easier to study.

“Gluten,” therefore, doesn’t really exist in nature. It’s something we made up, a human construct to describe the storage proteins found in the seeds of Triticeae grasses based upon their solvent interactions.

“Gluten” are those proteins that do not want to attractively interact with an aqueous solvent or neutral salt solutions. They are further subdivided into gliadins (soluble in ethanol) and glutenins (partially soluble in acid or alkali). Both are polydisperse, polypeptide chains of unfolded conformation that want to minimise their energetic interactions with water. Both subcategories, like their parent category “gluten,” are manmade, a biomolecular mixed bag.

When describing solubility of biopolymers, like proteins or polysaccharides, we’re referring to a global property. In reality, macromolecules have intricate three-dimensional structures built from diverse molecular subunits varying in hydrophilicity. “Gluten” proteins contain many side-chains or residues, mostly in their interior, capable of forming hydrogen bonds with water molecules.

As water’s added to flour, biopolymers hydrate. In wheat flour, the principal functional biopolymers of concern to bakers are “gluten” proteins, starch, and pentosans. Hydration occurs via water molecules diffusing into the porous interior of the biopolymers.

This inflow of water, due to pressure differentials, causes biopolymer swelling. They expand in volume, which further increases their internal surface area, allowing even more water to be ad/absorbed within the biopolymer interior.

The rate of molecular diffusion in bulk flow surrounding biopolymers is greater than the rate inside the biopolymer interior.

Research estimating the water-retention capacity of wheat flour constituents is based upon the fractionation and purification of each individual component. Thus, the amount of water that starch, gluten, and pentosans can absorb as stated in research literature is based upon measurements of each done in isolation. When these values are summed together, the total water is much, much higher than what is found in real dough conditions.

The preceding means that hydration is under kinetic control. The equilibrium distribution of water in pastes, doughs and batters thus depends upon mixing conditions all else being equal. This equilibrium (where and how water goes) can be shifted, knowingly or not, by the baker, which is exactly what happens during autolysis.

Calvel wasn’t wrong. He also wasn’t necessarily right, either.

The use of an autolyse step does weaken a dough. But it’s not because “gluten” proteins are degraded, nor because extensibility values are increased. Rather, peak elasticity values are decreased, in a time-dependent manner, due to distribution kinetics of water between flour constituents.

How does this occur? 

When an autolyse step is used, the rate-limiting step is diffusion of water molecules throughout flour, a heterogeneous, porous material.

The solvent becomes poorer for “gluten” proteins at a rate faster than the hydrophilic sites on the largest of these proteins can favourably interact with a “good” solvent. Put another way, the aqueous solution becomes a non-wetting medium for a dough’s network-forming components faster than all network-forming components can be wetted.

Why is this important?

Milling aside, “strength” in wheat-dough systems directly corresponds to the molecular weight and total concentration of the largest-sized “glutenins” in a flour. These are dense, spherical particles 3 to 5 μm in diameter.

Want elasticity? Here’s where to look. (Interestingly, their spatial distribution in the wheat endosperm follows a concentration gradient, with density increasing toward the kernel centre. Want to make a flour stronger? Import spring-sown, lower-yielding, harder grains from hotter, drier new-world locales for your grist. Roller-mill, hitting the reduction rolls hard. Focus on the bull’s-eye of your concentration gradient, discarding as much outside your target as possible. You’re getting the good stuff, the purest, finest Manitoba White. It’s all inert, glassy material, free of pesky biological “doers,” like enzymes, that may interfere in your product getting high. Congrats, you’ve made panettone flour. For the purest, not the purist. Let your product get high!)

During a continuous-mixing operation, constant mechanical shear flow is applied to a dough system. The macromolecular “gluten” proteins are fluidised, oriented in the same direction as the bulk solvent. The larger, compacted spherical glutenin particles unravel into filaments, like pulling on a ball of yarn.

This laminar flow breaks the liquid threads of proteins into smaller droplets, increasing protein surface area, and allowing more thorough contact with the solvent. Imagine the dashes of Morse code becoming dots, and these dots becoming more evenly distributed throughout the dough system.

The percentage of water molecules that are able to favourably interact with “gluten” proteins will stay with the proteins after dough formation. Wheresoever the proteins “go,” so too that percentage of water follows. (The interaction length of hydrogen bonds is exceedingly miniscule. Here, we’re referring to a statistical macrostate property of the system.)

At our scale, a homogeneous, strong dough forms from the mixing process, but, at the mesoscale, it’s the result of a de-mixing process resulting from phase separation. Wheat doughs made from refined flour have four distinct phases differing in density (from lowest to highest, aqueous, pentosan, gluten, and starch phases). These phases are in thermodynamic equilibrium, and that equilibrium is based upon mixing parameters.

During mixing, changes in solvent quality occur due to solubilisation of smaller-molecular weight flour constituents, including sugars, salts, minerals, amino acids, water-soluble proteins (albumins, globulins), soluble polysaccharides (pentosans, damaged starch), lipids, ethanol, organic acids, and dissolved gasses.

As a result, the quality of the solvent (water plus soluble flour constituents) for “gluten” proteins decreases. It becomes a non-wetting medium that causes “gluten” proteins and other biopolymers to aggregate into separate phases through self-association. That is, each biopolymer class (“gluten” proteins, pentosans, and starch) seeks out their “own” kind because it’s more energetically favourable than interactions with a poor solvent or other biopolymer classes.

“Gluten” proteins represent the continuous phase of the dough system, with the other three acting as the dispersed, discontinuous phases. This expansive, dough-wide protein network cannot form unless enough mechanical energy is input, in both total quantity and in rate, into the dough system.

Upon removal of shear flow, the protein phase forms a cross-linked gel network (“gluten”) stabilised by non-specific, physical interactions. This gelation is reversibly removed if shear flow is reintroduced.

Continuous mixing results in a smoother distribution of water between the macromolecular (starch, “gluten,” and pentosans) and aqueous phases, increasing hydration of the former. An autolyse step, by stopping the mechanical shear flow, allows for the aqueous medium to change unfavourably before all the unexposed hydrophilic sites on “gluten” proteins can undergo favourable interactions.

Autolysis, rather than “destroying gluten,” prevents the unfolding, hydration and physical stabilisation of some portion of the largest-sized glutenins by a good solvent, thereby decreasing the total amount of these glutenins able to participate in gel-network formation. This is experimentally confirmed by research into passive hydration of “zero-developed” doughs quantifying the amount of “glutenin macropolymer” present in such dough systems at different time points and under different mixing regimes.

This experimental approach is a good stand-in for autolysis. The technique allows a predetermined quantity of shaved ice of specific particle size to melt overnight and hydrate a predetermined quantity of different flour types, with hydration based upon optimal absorption. Fundamental and empirical rheological measurements used in such experiments against a control further confirm the changes in strength character of doughs that have undergone an autolysis step.

Lastly, let’s end with practical considerations. Why use an autolyse step, if at all?

In the home environment, or when working in small batches to be hand-mixed, the answer is it’s unnecessary. Whatever you think it does, it doesn’t do, and, what it does do, needn’t be done, especially when working in discontinuous, one-off batches where “optimal” dough development is rarely achievable.

In such cases it has more negatives than positives. E.g., adding more time and unneeded work, or, in colder climates, decreasing your ability to nail your final dough temperature (FDT). (Remember, temperature’s everything in sourdough baking. Another way of saying this is, you can do all else perfect, but, if you don’t nail your temp, you can still end up making really shitty bread.)

For the professional baker, my answer is again the same. I’ve never encountered a scenario in any professional bakery (either my own or the ones I’ve consulted for around the world) where I thought it’d be beneficial.

In fact, I cannot even conceive of a hypothetical situation where I’d ever want to use autolysis. Remember, all it does, if anything, is to decrease peak elasticity values by one-quarter to one-third. If I was working with too tenacious a flour, I’d simply swap it for one I like more. There are also other, more practicable, effective solutions I’ve used in real-world situations while consulting to mitigate too-strong flours when their use was unavoidable, all so the baker doesn’t have to sacrifice nailing her FDT.

The unfortunate fact is that the inclusion of an autolyse step has become dogma in many smaller, independent bakeries, in both France and throughout the new world. (It’s not realistic in larger production volumes.) This practice has filtered into non-professional settings, where, in my opinion, it adds needless complication.

My advice? Make no-knead bread. It’s easier.