Whey proteins for reducing astringency: If it works for tea…

Important note: The following story should NOT lead you to conclude that you need to stop drinking wine if you generally avoid dairy. Please see the comments at the bottom of this post, this post on wine allergies (and why you probably don’t have them), and this post on gluten in wine.

Short: The main protein in whey, a voluminous byproduct of cheesemaking, might be a good (and cheap) fining agent for reducing astringency in red wine, but it’s going to need a lot more testing, and I’m (very) suspicious.


Many people* add milk to black tea because it softens the harsh edges of what is so often an unfortunate cup of broken orange pekoe. The combination works because a milk protein, β-lactoglobulin, binds tannins. This was news a few years back because adding milk reduces astringency, but not nutrition value (unless you’re fine-filtering your tea after adding milk, and I can’t imagine you are). Whey protein-bound tannins won’t bind to your salivary proteins to dry out your mouth, but still work as antioxidants in your body.

Adding β-lactoglobulin from whey protein might for essentially the same reasons be useful to decrease the astringency of red wine. Casein, another milk-derived protein, is routinely used for wine fining. The benefit of making the switch is cost. Casein ends up in cheese. β-lactoglobulin ends up in the whey left over after making cheese. Ergo, the cheese-eating world has more β-lactoglobulin sitting around than casein, body-builders’ post-workout protein shakes and the occasional bottled whey beverage notwithstanding.

This is one of those so obvious and sensible-sounding ideas that I’d be inclined to think: surely someone’s already tried this and found that it doesn’t work or everyone would be doing it already. Still, a quick search of the scholarly literature suggests that no one has, publicly, though this study from 2007 found that β-lactoglobulin binds well to resveratrol.**

The chemists behind this just-published study found that β-lactoglobulin reduced astringency about as well as gelatin, gram for gram, at least for the random cheap French merlot they tested. Even better, β-lactoglobulin worked about as well as a combination of β-lactoglobulin and casein.

If you’re waiting for the catch(es), here they are:

  • β-lactoglobulin shouldn’t work as well as gelatin on the basis of established knowledge about tannin-protein interactions. Tannins bind best to big, loosely folded proteins with lots of an amino acid called proline. Tannins bind less well to small, balled-up proteins. Gelatin (and some major salivary proteins, coincidentally) are loosely folded and proline rich. β-lactoglobulin is small and balled up. The authors make a gesture toward figuring out why β-lactoglobulin still works by quantifying the concentration of total protein-precipitable tannins and a few specific, important tannin molecules in the wines before and after treatment with gelatin, β-lactoglobulin, or the β-lactoglobulin-casein combination. Those tests confirmed that β-lactoglobulin is binding tannin molecules, but also that they’re binding weakly.
  • This weak binding seems to involve groups of tannin molecules strung together. β-lactoglobulin doesn’t do a good job of binding to individual tannin units (monomers) hanging out in the wine on their own.
  • These authors only tested whey protein with one red wine (the random French merlot), and only measured astringency by chemical approximations, not with real swirling and spitting tasters. So I’m suspicious. Surely, the idea of using whey protein for wine fining has occurred to at least one food processing house or winemaking supply company with the resources to test the theory out on their own, on many different wines, with their in-house sensory panels doing the tasting. If it worked, whey protein would be on the market. If it didn’t, they might well not broadcast the news. And – forgive me this, because excellent research can come from unlikely places – the authors are from the University of Reading in the UK and the National University of Rosario in Argentina, neither of which is a hot bed of groundbreaking wine research. That’s not to say the researchers aren’t good, but it is to say that they might not have ideal access to the accumulated wisdom of the field.

An interesting idea. Unlikely to be the next wine chemistry success story, but an interesting idea.


*I add milk to black tea on rainy days because the aroma evokes England and memories of a childhood I never actually had of drippy days shut up inside with endless tea and books in a British country house. I can tell you this because it has rained here every day for more than a week.

**I think that at least the general idea that β-lactoglobulin binds to wine tannins might be buried in this 2011 masters thesis conducted at Massey University in New Zealand, but I’ll be honest: I didn’t have the patience to read the whole thing carefully enough to find out.

Does adding tannin boost aromatic thiols, too? It just might.

Thiols aren’t quite like bacon, but they’re not too far off trend-wise. These aromatic sulfur-containing molecules are highly appealing in small quantities — even low concentrations lend a wine’s aroma fresh fruity notes (tropical in sauv blanc, black currant or berry in reds). Just about everyone wants them, or wants more of them. They’re at work in the expected places (thiols in sauvignon blanc are like the bacon in your pasta carbonara; bland without, and much better with), but also do a fair bit in the unexpected ones, too (thiols contribute to the aroma of Bordeaux reds and Provençal rosés, for example, and bacon, I’m told, does excellent things to cupcake frosting*).

Unlike bacon, we still don’t have an especially good idea of how thiols are formed (we figured this out for bacon a good long while ago, I believe). The amounts yeast transform from various precursors under realistic wine conditions just don’t add up to the final concentrations we find in wine, and how the rest happen remains an open question. Last year’s news was that tannins contain thiol precursors upon which yeast act during fermentation. Now, those researchers (an Italian group, with the aid of a Sauvignon blanc-oriented researcher from New Zealand) have demonstrated what I’m sure they’d hoped for when they published last year’s paper: adding tannins to wine before fermentation increases a wine’s thiol concentrations, specifically 3-mercaptohexan-1-ol (3MH). (For some context on 3MH and other sulfur compounds, Jamie Goode’s blog article on the topic is a good primer).

This study is very much a first step, and a bit of a disappointing one. Tannin was only added at one concentration: 1.6 g per 2 kg batch, compared with a no tannin-added control. Seeing a dose-dependent response — add more tannin, get more thiols — or showing that the relationship between those two variables isn’t linear, anything other than just two points, would have been much more convincing. As would using larger than 2 kg batches for those experimental wines (2 kg ~ 1 750 mL bottle), since the volumes in which experimental wines change yeast fermentation and oxygen exposure dynamics; the oxygen mightn’t be relevant here, but the fermentation parameters are. AND, each wine was only made in duplicate, not satisfying the usual experimental expectation of performing studies in triplicate. With two samples, if one is off you can’t tell which reflects the trend you’d see if you did the experiment a hundred times (and you certainly shouldn’t just average them together); if three samples all group, you can feel better about life (and your results). AND, with so little wine, the authors couldn’t conduct a proper sensory analysis, not that doing so would have been worthwhile in any case with their mini-make-do winemaking technique. In other words, this study is less than convincing on methodological grounds.

All of that said and duly noted, this study points toward some interesting possibilities. For instance, I’ve recently talked with a few winemakers who have been experimenting with tannin additions to good but confusing effect. (I seemed to come across people talking about tannin additives about as often as I did bacon-laden menu items on my most recent trip through Eastern Washington, which is to say, a lot.) They know good results when they see them, and they like what they taste. But tannin assays sometimes seem to yield results that conflict with experience, with the assay saying that the with-addition and without-addition wines contain the same amount of tannin even though the winemaker can taste a difference. All manner of possible explanations exist for that phenomenon, and I don’t want to suggest that thiols are responsible for those sensory differences. Nevertheless, this study is a good reminder that adding anything to wine is bound to have more than just one obvious, direct effect, and that adding tannins could play with wine aromas in ways we hadn’t expected.

*I’m told, because I’m one of three people on the planet who likes neither bacon nor cupcake frosting.

Measuring not just tannin concentration but tannin behavior: Kennedy’s stickiness assay

Why does it always seem that we know the least about stuff that’s the most important? Tannins garner a lot of wine researcher’s attention, and for good reason. No one needs convincing about how important tannins are to wine quality (especially not the consulting companies who’ve correlated high tannin concentration with high wine magazine ratings). The amount of noise made about tannins, though, could give someone seriously inflated ideas about how well we understand them.

Excellent wine chemists are, in fact, still thinking about really good, consistently accurate, and every-day-practical ways of measuring a wine’s tannin concentration. The well-known Harbertson-Adams assay went a long way in that direction, but isn’t the last word on the topic. But just looking how much tannin a wine has doesn’t tell us enough. “Tannin” describes a whole group of molecules, and those molecules behave in different ways.

What we really need is a way of measuring not just how much tannin a wine has, but how astringent it’s likely to feel. That’s a tall order — astringency is a complicated sensation affected by alcohol concentration, sugars, polysaccharides, the person doing the tasting, and undoubtedly other factors. Just tasting the darn thing is, without question, the most elegant and reliable way to measure wine astringency. But it would still be useful to have a way of measuring the relative astringency of different types of tannins to correlate with how different production techniques affect those tannins and make some predictions. And, just as importantly, if we’re ever going to figure out what tannins do, how they behave, and how astringency works, having more tools to look at them is important.

James Kennedy’s group at Fresno State is working on a way to go beyond traditional tannin measurements, which just tell you how much tannin you have, to develop analyses to tell you what the tannin you have does and how it’s likely to produce astringency. More particularly, they’ve developed a way to measure the stickiness of any particular type of tannin molecule. Stickiness, as defined in the article, is “the observed variation in the enthalpy of interaction between tannin and a hydrophobic surface.” Or, to put it a lot more simply, stickiness describes how strongly a tannin is inclined to attach itself to something else (without actually reacting with it). This seems pretty commonsensical — if we sense astringency when tannins glom together with our salivary proteins, then we’d like to know how glom-inclined those tannins are. They’ve shown that their stickiness measurement for a particular set of wine tannins remains constant no matter how much of the tannin you test — in other words, they can measure stickiness as a tannin quality, not tannin quantity.

It’s a trickier puzzle than it might seem. How do you measure how tightly two molecules are holding on to each other? And when you’re interested in how different tannins interact with proteins, which are themselves a very diverse group of molecules, how do you choose which protein is going to be the protein that represents all other proteins?

For Kennedy and company, the solution involved choosing something that isn’t a protein at all but polystyrene divinylbenzene, a polymeric resin that holds on to tannin in remarkably the same way as the specific amino acid (proline) that acts as the tannin-attractant in salivary proteins. The resin allows for a standardized stickiness measurement and no doubt has all sorts of advantages in terms of working with it in the lab. It won’t actually behave like real salivary proteins which, being folded up into various shapes with proline more or less accessible along their various crannies, don’t bind tannins in ways so predictable. The upshot is that this is a standardized measure of stickiness (a defined scientific parameter), not an actual measure of astringency (a subjective sensation). Nevertheless, stickiness values and astringency should be related in predictable ways. We’ll very likely see a publication verifying that relationship with human tasters before too long.

Stickiness assessment involve some fairly complex chromatography, improving on a method the lab published last year. The methodological details are less important than realizing that this isn’t something that even a well-equipped winery lab is going to be able to do on their own (unlike that Harbertson-Adams assay, which is pretty accessible for a lot of winemakers). Though some wineries may measure tannin concentrations with that Harbertson-Adams assay, which is pretty accessible for a lot of winemakers, stickiness measurements aren’t going to become the new best thing in figuring out how long your syrah needs to spend on its skins before being pressed off. Too expensive (the chromatography columns needed for this kind of work run hundreds of dollars each), too training-intensive (unless you have a chemistry grad student hanging out in your winery), and too little of an improvement over just tasting the darn thing. This research isn’t likely to change the way anyone makes wine tomorrow or even for the next year or two. But it very well may change the way scientists study and think about tannins, the kinds of questions they can answer — those tricky issues around the relative astringency of various seed and skin tannins, for example — and what they can tell winemakers about targeting specific wine styles a few years down the road. And that’s worth making some noise over.