Dialing in Phenolics by Blending POF- and POF+ Strains

If you’ve ever had the pleasure of drinking a German Hefeweizen or a Belgian Tripel, you may recall the interplay of fruit aromas like banana, orange, or maybe even plum, coupled with fragrant spices like pepper or clove. The latter group belongs to a family of compounds called phenols. Phenols can run the range from pleasantly spicy (the pepper and clove from before) to perhaps a more acquired taste (smoky, barnyard) to potentially undrinkable (plastic, band-aids). While some brewers embrace these flavors, many brewers avoid them by taking advantage of yeast strains that lack the ability to produce them.

In order to understand phenolics in beer (specifically 4‑vinyl guaiacol, or 4VG), we need to talk about ferulic acid. Ferulic acid is abundant in nearly all brewing grains, including barley, corn, oats, rice, wheat, and rye. Ferulic acid can be found in three forms in grain: free, soluble-esterified, and insoluble-bound. When brewers talk about the benefit of a ferulic acid rest (holding the mash between 104 – 122°F/40 – 50°C), the aim is to free these bound ferulic acid molecules in the hopes of providing more precursors for desirable flavor and aroma compounds. 

In order to create 4VG, ferulic acid must lose a carboxyl group either through thermal or enzymatic activity. Ferulic acid decarboxylase, the enzyme specifically produced by the FDC1 gene, is responsible for the bulk of this reaction in beer fermentation. Although many wild strains of S. cerevisiae have an active version of the FDC1 gene, domestication and industrialization of brewing strains led to the selection of strains that had lost this functionality. You can see this absence in nearly all British, American, and Norwegian ale strains and all lager strains, but the phenolic character is still preserved in many Belgian and a few German ale strains. 

This experiment was designed to get a common request that we have from brewers: how do I make a banana-forward hefeweizen with less clove? We know that banana esters can be dialed in with temperature, pitch rate, and oxygenation, but phenolics are pretty consistent regardless of how you handle your yeast. So with these new strains, we thought it would be really helpful if blending the non-phenolic and phenolic versions could finally give brewers a way to dial in the phenolic character of their Hefeweizen.

We already had some evidence for this because through sensory testing we can distinguish phenolic off-flavor (POF) strains that are intensely phenolic to just barely recognizable as phenolic. This suggests that ferulic acid metabolism by yeast is rate limiting, and potentially can be adjusted by blending. We demonstrated this concept with flask fermentations that were spiked with 100 mg/L of ferulic acid. This assay was designed to be very similar to the published work of the Verstrepen lab. 

We reserved a small amount of the unfermented wort spiked with ferulic acid as the ​“Ferulic Acid Control” and another unfermented wort sample as a blank sample for the UV-vis spectrophotometer. We inoculated seven 300ml flasks (autoclaved dried malt extract, OG of 12°P with ferulic acid addition to 100 mg/L) with 10 million cells/ml total yeast but with blends of different percentages between Bananza (OYL-400) and Hefeweizen Ale I (OYL – 021). 

After fermentations were complete (7 days), we measured the absorbance of the resulting beer with our Genesys UV-vis spectrophotometer. We used the scanning mode to measure the absorbance from 270nm to 400nm where the ferulic acid peak absorbance is ~325 nm. To our delight, we were able to confirm the direct relationship between the amount of the phenolic strain (OYL – 021) and the depletion of ferulic acid, and by the use of sensory testing alone, could very clearly order the resulting beers from least to most phenolic.

Success! Blending OYL – 021 and OYL – 400 is a straightforward strategy for adjusting the amount of clove character in your hefeweizen. Our personal favorite was a 10:90 blend of the OYL – 021 and OYL – 400, respectively, which provided just the right amount of clove to contrast to the heavy banana esters.

For our first sensory panel, we asked our staff to blindly evaluate each sample, on a scale from 0 to 5, based on three general descriptors: phenolic, estery, and hoppy. Finally, as an additional validation, we asked our tasters to rank the samples from most to least phenolic. From our staff, we received 17 responses.

While our tasters were reliably able to identify the beers made with 100% OYL – 021 & the 50:50 blend as the most phenolic of the group, we were rather thrown off by the distribution of rankings for OYL – 400. The average rating of perceived ester quality seemed more in line with expectation! If any of the blends had produced a sub-threshold phenol content, we would’ve expected the 90:10 blend and pure OYL – 400 to have a similar spread of ratings. Considering the order of samples (50:50, 90:10, 100% 400, 100% 021), miscalibration wasn’t likely the culprit. The yeast cakes and beer were confirmed to match expectations for colony composition, so infection was out of the question. 

The easiest assumption we could make is that our tasters did not all have a consistent sense of what 4VG should taste and smell like and especially whether something was more or less phenolic than the other. Tasters, even trained, can have different understandings of and sensitivities to flavor and aroma compounds. As much as we can prime our tasters with questions like, ​“how peppery or clovelike is this,” provide controls for reference, or go deeper into training, developing and maintaining a large and effective tasting panel requires a lot of dedication.

It made sense to see what the impact of taking vocabulary out of the picture would do. To this end, we ran a second sensory panel as a double blind triangle test to differentiate the beer fermented solely with OYL – 400 from the beer fermented solely with OYL – 021. In this panel, we received responses from 12 tasters of which 10 could distinguish between the phenolic & non-phenolic sample (Exact p‑value of 0.001).

Some descriptors of non-phenolic sample by those who identified the outlier:

Some descriptors of phenolic sample by those who correctly identified the outlier:

One final curiosity we wanted to explore is if our tasters could distinguish an even mildly phenolic beer from a completely non-phenolic beer. We ran a third sensory panel as a double blind triangle test to differentiate the beer fermented solely with Bananza (OYL – 400) from the beer fermented with a blend of 90% OYL – 400 and 10% OYL – 021. In this panel, we received responses from 13 tasters, of whom 9 could distinguish between the phenolic & non-phenolic sample (Exact p‑value of 0.009).

Some descriptors of non-phenolic sample by those who identified the outlier:

Some descriptors of phenolic sample by those who correctly identified the outlier:

There are more trials we wanted to run on this batch but, given the 2.5 gallon volume of our fermentations, any additional sensory evaluation was not feasible. We could likely assume, from the results of the first sensory panel, that our tasters could distinguish a 90:10 POF-:POF+ blend fermentation from a 100% POF+ fermentation. 

Beyond that, there are still more questions we want to explore — how would altering the grist, the fermentation temperature, or perhaps even the amount and type of hops used affect our tasters’ perceptions? We’re still chewing on the idea of blending a bit of non-phenolic OYL – 400 with a haze-favoring strain (like British V/OYL – 011 or even its Thiolized counterpart, Cosmic Punch/OYL – 402) for an interesting fruity addition to what the hops and thiols are already contributing in a classic NEIPA. More to come!