Tetrahydropyridine

Forms of Tetrahydropyridines (abbreviated and often referred to generically by brewers as THP), specifically 2-acetyl-3,4,5,6-tetrahydropyridine and 2-acetyl-1,4,5,6-tetrahydropyridine (abbreviated ATHP or ACTPY), 2-ethyltetrahydropyridine (abbreviated ETHP or ETPY), and 2-acetylpyrroline (abbreviated ACPY or APY) , which are classified as a ketone and a cyclic imine , are commonly attributed to the "mousy", "urine" (in high amounts) Cheerios® or Cap'n Crunch® (in low amounts), "breakfast cereal", or more generically, "cracker biscuit" flavor in sour beers. The flavor is detected towards the end of the swallow, and the aftertaste can remain for a few minutes. Not all people are able to detect the flavor of THP. The low pH of sour beer or wine makes it harder to detect the flavor and often impossible to detect the aroma. An increase in pH is needed in order to detect the off-flavor, and the mouth's salivary pH serves that purpose when tasting beer or wine with THP. For example, as the mouth's pH adjusts back up after swallowing a sip of sour beer or wine, the THP becomes detectable in the aftertaste. This effect on sensory detection by low pH might also explain why some people are better at detecting it since people have different pH's on the surface of their tongues and saliva . Diacetyl is sometimes mistakenly indicated as a potential cause of "mousy" flavors in sour beers, however, tetrahydropyridines are the accepted cause. The flavor tends to age out of sour beers after 2-6 months in the fermenter, kegs, or bottles (although aging periods as long as possibly 8-12 months have been reported on MTF ). The exact mechanism for how THP ages out of beer is not fully understood, and it is unknown whether cold or room temperature storage speeds up the breakdown process. It is more likely that room temperature storage will result in faster breakdown of THP, and anecdotes from MTF members seem to support this . Many brewers have noticed that pitching rehydrated wine yeast at packaging reduces the amount/duration of this flavor in kegs and bottles .

In food, Tetrahydropyridines are associated with the aroma of baked goods such as white bread, popcorn, and tortillas, and is formed by Maillard reactions during heating.

Traditionally, the mousy/Cheerios® flavor from THP is considered an off flavor in both wine and sour beer. There is some debate and differing opinions as to whether or not a small amount of THP flavor is allowable (or even enjoyable) in sour beers, however most consider any level to be an off flavor .

History of Scientific Research

Grbin et al. (1996) Survey of microorganisms that produce THP .

THP in wine ("mouse taint") was first described in wine by Müller-Thurgau and Osterwalder in 1913, although mention of an off-flavor in wine that 'is closely resembling the smell of a residence of mice' dates to 1894 in "A Treatise on wines" by J.L.W. Thudichum. Müller-Thurgau and Osterwalder attributed it to be produced by bacteria. They also established that the ability to detect THP varies from person to person . A 1908 text by the Germans Nessler and Windish linked THP off-flavors to wine that aged longer on lees, perhaps due to yeast autolysis . In the 1940's, some scientists proposed that THP was purely a chemical reaction, however, their evidence was inconclusive. In 1977, Tucknott et al. was able to identify that THP production was attributed to both Brettanomyces yeast and lactic acid bacteria, and that ethanol and L-lysine played a role in its production, and this was confirmed by Heresztyn et al. in 1986 .

The chemical analysis of THP has proven to be a difficult endeavor in science. In 1995, Herderich et al. out of Australia developed a method for chemically analyzing THP for the first time. It was at this time that all three forms of THP (ETHP, ATHP, and APY) could be identified consistently in contaminated wine. It was also during this time that the three forms of THP contributed to the flavor of various foods, such as tortilla chips and taco shells. For example, APY was found in the fermentation of cocoa in 1995. In 1995, Grbin et al. described Brettanomyces yeast as the yeast that produces THP, while wine strains of S. cerevisiae did not .

In 2000, the Australians Grbin and Henschke showed that some strains but not all of B. bruxellensis, B. anomalus, B. nardenensis, and B custersianus produce THP, and that THP production was influenced by the carbon source, but not dependent on it (THP was still produced in dry wines with little available nutrients, and fermentation rate was not always related to THP levels) .

In 2007, the Australian team of Grbin et al. developed a complex and unique method of analyzing forms of THP using a Finnigan TSQ 70 mass spectrometer directly coupled to a Varian 3400 gas chromatograph. The chromatograph was equipped with a 30 m J&W Carbowax 20 CAM fused silica column, 0.25 mm i.d, and 0.25 μm film thickness. They were able to confirm that higher lysine levels increased the amount of ATHP produced (but not ETHP), although with diminishing increases of THP as the lysine level was increased. They also discovered that L-ornithine also functioned as a precursor for THP production in Brettanomyces. The group proposed a biochemical pathway for the different forms of THP in Brettanomyces .

Due to the specialized GC/MS equipment needed for measuring forms of THP that most labs do not have, certain answers will be difficult to obtain. Some studies have used alkaline strips as a way to smell the aroma of THP, and while not as precise as the specialized GC/MS lab equipment, could still help further the understanding of THP. These strips were prepared in the following way: knowing that mousy off-flavour has a lingering sensory impact, the technique of alkaline paper strip assessment was adapted from Heresztyn (1986a). Paper strips (Whatman No. 1, 4–5 mm × 50 mm) were prepared by soaking in NaOH (0.1M) and drying overnight at room temperature. The alkaline paper strips were then briefly dipped into cell-free samples (centrifuged) and immediately assessed for the mouse-like odour by sniffing .

Other techniques for detected THP in food have been developed, which might be applicable to wine or beer. For example, Grimm et al. (2001) developed a a technique for detecting 2-acetyl pyrroline (APY, or 2AP) in rice. The rice samples had to be heated to 80-85°C in order to extract the volatile APY, and then APY levels in the headspace of the rice container could be detected using solid phase microextraction (SPME) with fibers that operate at the higher temperatures . It isn't known of such methods would also work for measuring THP comopunds in beer or wine, but they could provide an option for beer and wine researchers.

Forms of THP

ATHP

ATHP/ACTPY (which has two forms: 2-acetyl-3,4,5,6-tetrahydropyridine and 2-acetyl-1,4,5,6-tetrahydropyridine) has a much lower flavor threshold than ETHP (see Thresholds), and has been historically cited as the cause of mousy off-flavors detected in wine. In wine, its aroma cannot be detected due to the low pH of wine (it can be detected if the pH is raised), only the flavor. It is easier to detect in higher pH wines. ATHP is the form of THP that is the major contributor to the aroma of freshly baked bread, corn tortilla chips, and crackers. How different foods/wines/beers interact with ATHP on the palate may explain the different flavors that are detected by people, as well differing concentrations and peoples' ability to detect ATHP . In food, ATHP has a slightly different form (6-Acetyl-1,2,3,4-tetrahydropyridin). Its formation in food is due to cooking and thought to be associated with Maillard reactions, and its presence ages out quickly in food and is considered extremely volatile .

ETHP

ETHP/ETPY (2-ethyl tetrahydropyridine) was first identified in wine in 1973, but until recently further studies weren't able to confirm its presence in wine. Its odor threshold is quite high (see Thresholds), and so it was not considered a major source of mousy off-flavors in wine for some time. Consequently, research on ETHP has been limited. More recently, it was shown that Lactic Acid Bacteria (LAB) can produce above threshold levels of ETHP, making it recently important to wine researchers .

It has been speculated by scientists studying mousy off-flavors in wine that its production is the result of slow metabolism of ATHP into ETHP by Brettanomyces. ETHP was observed to form much slower than ATHP, and coincided with a decrease in ATHP. This slow production of ETHP may be another reason it has been underestimated by researchers until recently . ETHP has not been shown to break down into another byproduct.

APY

APY/ACPY (2-acetyl pyrroline) is a more volatile but more potent form of THP. It has a significantly stronger odor and lower odor threshold in wine than ATHP. It can also be found in damp pearl millet, white bread, taco shells, tortilla chips, boiled potatoes, roasted sesame seeds, pan-fired green teas, cured tobacco leaves, peanut and pumpkin seed oils, honey, several soy-based products, and more aromatic rice such as Indian Basmati. APY is primarily produced by heterofermentative LAB. In food, APY formation is due to cooking and thought to be associated with Maillard reactions, and its presence ages out quickly in food and is considered extremely volatile so much so that the food industry has created powdered forms of APY to increase the flavor stability of some foods associated with it. Some plants such as rice crops produce varying levels of APY naturally .

Unidentified "Transient" Forms

There have been anecdotal reports of other forms of mousy off-flavors. During growth of lactic acid bacteria (LAB), mousy off-flavor detection fluctuated with high levels detected early on, and lower levels detected towards the end of growth. This indicates that there may be a transient, strain-dependent form of THP that can occur during malolactic fermentation. There have also been a sensory detection of mousy off-flavors at different levels than the documented levels of ATHP, ETHP, and APY, which were not associated with LAB or Brettanomyces .

Sulfur-Containing Forms

Sulfur-containing versions of THP exist in many vegetables and processed meats and are important aspects of their aromas. These include 2-Acetyl-2-thiazoline (the sulfur containing version of APY) and 5-Acetyl-2,3-dihydro-4 H -1,4-thiazine (the sulfur containing version of 6ATHP). These are formed in meat during processing, for example in beef and chicken broths, roasted beef, etc. They are characterized as being similar to APY and ATHP, and have a roasty, popcorn-like odor and low thresholds. These are chemically formed during cooking and food processing and are not associated with biological metabolism or fermentation .

Production

Proposed pathway for THP production by Brettanomyces

All species of Brettanomyces can produce forms of tetrahydropyridine in varying amounts, although some below threshold. Additionally, Lactic Acid Bacteria (LAB) including Lactobacillus and Pediococcus can produce forms of THP. Acetic Acid Bactera (AAB) has also been demonstrated to produce forms of THP .

Brettanomyces

Although the exact pathway is not known for Brettanomyces (several are proposed), the conditions for THP production are well documented. ATHP is produced by metabolizing the amino acid L-lysine or D-lysine , along with ethanol and a glucose or fructose molecule. Iron is also needed for THP production, although its exact role in biosynthesis is not known . As with other amino acids, lysine is taken up by Saccharomyces during fermentation and then released after fermentation. Levels of lysine fluctuate slightly throughout fermentation but are generally high throughout a beer's lifetime (including after fermentation) . Wheat generally has a slightly lower amount of lysine than barley, and oats have a slightly higher amount of lysine than barley . In red wine, yeast autolysis releases many amino acids including lysine . Aging beer on trub and its effects on THP production has not been studied, but it might not be a factor in beer since lysine levels are high in beer regardless of yeast autolysis .

Oxygen plays a key role and has a stimulatory effect in ATHP and ETHP production (particularly ATHP), but its exact role is not understood. It has been speculated that since ATHP production is associated with Brettanomyces growth, and Brettanomyces grows better under aerobic conditions, that this is why more ATHP is produced under aerobic conditions . It has also been hypothesized that oxygen may have a direct effect on the THP molecules themselves . ATHP production was also shown to increase when anaerobically precultured cells were transferred to an aerobic environment, indicating that oxygen has a direct role on the production of ATHP, not just a byproduct of Brettanomyces growth . Limiting oxygen exposure during kegging/force carbonating is recommended for helping to reduce ATHP production; even very small amounts can have an effect (although the exact threshold of how much oxygen is required has not been determined). For example, the purity of the CO² supply should thus be taken into consideration when force carbonating. At 0.5% impurity (the impurity is air, 1/5 of which is oxygen) and at 2 volumes of CO², ~1,420 ppb of O² would be added to the packaged beer, which is an exceedingly high amount of oxygen. The CO² supply should ideally be 99.990% pure or better (this would introduce 46 ppb of oxygen at 2 volumes of CO²). The method that the CO² is added can also determine how much oxygen is introduced into the packaged beer. Sparging CO² (bubbling it through the beer) dissolves significantly less oxygen due to Henry's Law (see reference), while injecting (flushing) dissolves significantly more oxygen . Vessel purging methods with CO² are also less efficient than some might expect, and might still leave enough oxygen behind to stimulate THP production (see this HomebrewTalk thread). Pitching fresh Saccharomyces at bottling/kegging time and naturally carbonating the beer with sugar has reportedly reduced mousy off-flavor detection, perhaps because Saccharomyces metabolizes both the oxygen and sugar faster than Brettanomyces. Different strains of S. cerevisiae might be more efficient than others at helping reduce THP. For example, Mitch Ermatinger from Speciation Artisan Ales anecdotally observed that switching from CBC1 conditioning yeast to EC1118 reduced THP off-flavors from 1 month to two weeks (see Packaging for details on re-yeasting at packaging time). Justin Amaral and Mike Karnowski reported an anecdote that purging bottles with CO² reduced THP levels, although limiting oxygen in the bottle also had some negative effects on some conditioning yeast strains .

Interestingly, for unknown reasons Brettanomyces cells grown under aerobic conditions and then transferred to an anaerobic environment still produced significant amounts of ATHP in the anaerobic environment. It has been suggested that the aerobic conditions made the Brettanomyces cells predisposed to creating ATHP . Oxygen exposure during Brettanomyces starters could potentially stimulate ATHP production later on down the road, even if the beer is not exposed to oxygen, although anecdotal evidence shows that this may not be a concern for brewers. It is still advised to use an aerobic or semi-aerobic starter for Brettanomyces unless the brewer believes this might be the direct cause of ATHP problems in their beer because Brettanomyces requires at least a small amount of oxygen for growth. Any other oxygen pick up after the beer has finished fermentation is the more likely cause of THP production and the brewer's post-fermentation processes should be examined first.

The level of ATHP production varies widely between species and strains of Brettanomyces, with some strains producing insignificant amounts and others producing very high amounts above taste threshold . Additionally, ATHP production requires glucose or fructose, which explains why ATHP may be seen more often in stuck wine fermentations rather than wine that has finished fermenting. ATHP production by Brettanomyces was observed in wine with glucose or fructose added, along with synthetic growth media, suggesting that the type of growth substrate does not effect production .

The production of ATHP is not efficient, meaning that the amount of ATHP produced is not proportional to the amount of L-lysine consumed. Therefore, the production of ATHP appears to be a byproduct (secondary metabolite) of L-lysine catabolism . ATHP is further metabolized into ETHP by Brettanomyces, although not much is known about this metabolic process . ETHP has a significantly higher taste threshold, and is often not detected in contaminated wine .

Although Brettanomyces is capable of producing APY from L-ornithine , the amount produced is much less than that of LAB and high amounts of L-ornithine are required. In wine, there isn't enough L-ornithine present to production significant amounts of APY from L-ornithine. Therefore, the presence of APY (which is much easier to detect aromatically than ATHP) indicates a bacterial contamination in wine (it is unknown if this applies to beer) .

The presence of the "mousy off-flavor" caused by forms of THP appears to be temporary in beer. Although not much is known about the degradation or metabolic breakdown of ATHP/ETHP, it tends to age out of beer after 2-6 months. Since the odor/taste threshold for ETHP is much higher than ATHP, and ATHP appears to be metabolized into ETHP by Brettanomyces over time, this may be one of the mechanisms by which the mousy off-flavor ages out of beer. The possibility of ETHP breakdown is not mentioned in any studies that we know of. Another unknown is why does Brettanomyces produce ATHP shortly after kegging and force carbonating a beer that has reached final gravity. The most likely cause is oxygen pick up during the kegging process. Pitching fresh Saccharomyces at bottling/kegging time and naturally carbonating the beer with sugar has reportedly reduced mousy off-flavor detection, perhaps because Saccharomyces metabolizes both the oxygen and sugar faster than Brettanomyces.

Lactic Acid Bacteria

Proposed pathway of APY and ATHP by L. hilgardii by Costello and Henschke (2002).

Heterofermentative Lactobacillus spp., particularly L. hilgardii and L. brevis, as well as Leuconostoc oeni , can also produce high levels of ATHP and APY from L-lysine/L-ornithine, ethanol (must be present), and iron. Although many strains of heterofermentative lactic acid bacteria can produce THP, not all do. For example, Costello et al (2008) found that all strains tested of L. brevis (3 strains tested), L. bucherni (3 strains tested), and L.hilgardii (8 strains tested) produced THP, several heterofermentative species did not produce any detectable levels of THP in a grape-based media (one strain each of L. fermentum and L. cellobiosus). Some strains within a species produce high amounts while others produce low amounts, for example Costell et al. (2008) found that some strains of O. oeni produced very high amounts between 50-150 µg/L while others produced very little between 5-20 µg/L in a grape-based media . A strain of L. plantarum (L11a) was shown to produce relatively low amounts. L-lysine stimulates production of ATHP, and L-ornithine stimulates the production of APY . Acetaldehyde has a stimulatory effect on ATHP and APY production, but is not required. No studies have been done to show whether or not oxygen plays a role in ATHP/APY production in LAB . Most species of Pediococcus do not create forms of THP, although a few species do produce relatively small amounts. In particular, these include P. pentosaceus , and P. clausenii , although one study found no THP in two strains of P. pentocaseus and only transient/occasional THP production in one out of five strains of P. parvulus . Oenococcus oeni and Leuconostoc mesenteroides have also been associated with creating ATHP, APY, and ETHP all above threshold amounts. Since only heterofermentative species produce significant amounts of THP, it is thought that its production is linked to the heterolactic pathway, and thus the metabolism of sugars in LAB . A pathway for APY and ATHP production in Lactobacillus hilgardii was proposed by Costello and Henschke, which involves the intake of lysine or ornithine, along with ethanol (which is broken down into acetaldehyde) to produce APY and ATHP . Lactobacillus pontis has been shown to break down proteins via proteolysis, yielding free amino acids such as ornithine which could serve as precursors to THP formation, and it might be reasonable to presume that other species of Lactobacillus could also free up ornithine as a precursor to THP .

Acetic Acid Bacteria and Mould

Although research is limited, acetic acid bacteria (Gluconobacter sp. and many strains of Acetobacter aceti) have been shown to produce forms of THP . Mediterranean dried sauses covered in mould have been characterized as having APY as a flavor contributor. The source of the APY was identified with a mould that grows on the surface called Peniciilium nalgiovense .

Thresholds and Quantities Found in Mousy Wine

Editor's note: the following thresholds are from studies on wine, and may not hold true for beer. As stated above, detection is influenced by pH, and so the low pH of sour beer may have the similar effect of repressing odor (more so) and taste (less so) detection, whereas non-sour Brettanomyces beers may have a higher detection rate.

• ETHP
  • Odor threshold (wine): 150 µg/L
  • Concentration reported in wines exhibiting mousy off-flavour: 2.7-18.7 µg/L (ETHP is generally not the cause of the detected mousy off-flavor)
• ATHP
  • Odor threshold (water): 1.6 µg/L
  • Concentration reported in wines exhibiting mousy off-flavour: 4.8-106 µg/L (ATHP is generally the cause of the detected mousy off-flavor)
• APY
  • Odor threshold (water): 0.1 µg/L
  • Concentration reported in wines exhibiting mousy off-flavour: Trace quantities up to 7.8 µg/L

Since a low pH hinders the detection of THP, one detection used by wine makers is to rub some wine on one's palm and smelling for THP. A more reliable method is to dip an alkaline paper strip into the wine, and then smelling the strip to detect the aroma of THP .

Off Flavor Kits

• FlavorActIV's "Mousy" is 2-Acetyletrahydropyridin. This is not the same form that appears in beer and wine fermentations though. This form is found in jasmine rice, oolong tea and baked bread, and has a similar flavor to the other forms of THP found in fermentation but does not have the lingering aftertaste .

Discussions

What We Don't Know

1. How does oxygen affect THP production biochemically?
2. How much oxygen has an effect, and is there a maximum amount of dissolved oxygen that would be insignificant in its production?
3. Does bottle orientation during storage have an effect?
4. Does creating an acid tolerance starter for conditioning yeast affect THP production?
5. How do THP off-flavors age out of beer (hypothesis: ATHP is metabolized into ETHP by Brettanomyces, or somehow chemically degraded into ETHP)?
6. Does storage temperature play a role in the degradation of THP off-flavors (cold vs room temp vs hot)?
7. What genetic phenotypes determine high THP production in some strains of Brettanomyces and not others?
8. What role does acetic acid bacteria play in THP production?
9. What are the unidentified transient forms of THP, and do they apply to beer or just wine?
10. Does the release of lysine from yeast autolysis when aging on trub increase THP potential (some say it does in wine and cider )?

* Due to the specialized GC/MS equipment needed for measuring forms of THP that most labs do not have, certain answers will be difficult to obtain. Some studies have used alkaline strips as a way to smell the aroma of THP, and while not as precise as the specialized GC/MS lab equipment, could still help further the understanding of THP. These strips were prepared in the following way: knowing that mousy off-flavour has a lingering sensory impact, the technique of alkaline paper strip assessment was adapted from Heresztyn (1986a). Paper strips (Whatman No. 1, 4–5 mm × 50 mm) were prepared by soaking in NaOH (0.1M) and drying overnight at room temperature. The alkaline paper strips were then briefly dipped into cell-free samples (centrifuged) and immediately assessed for the mouse-like odour by sniffing .

MTF Threads and Other Forum Posts

Below is a list of discussions on internet forum threads that may shed light on specific strains and individual experiences. Keep in mind that many of the opinions and experiences are anecdotal, although commonalities and shared experiences may prove to be useful and accurate.

• Joe Idnoni from House Cat Brewing and Brian Ogden from Attaboy Beer report that Caseinate de Potassium (a fining agent used in wine-making) was used to significantly reduce THP off-flavors in a kettle sour beer. Note that casein is a milk allergen, and might require declaration as a milk allergen (refer to your local government requirements) .
• Tariq Ahmed of Revel Cider discusses their protocol for reducing THP in spontaneously fermented ciders.
• MTF thread detailing experiences with certain strains and procedures (03/24/2017).
• Milk The Funk thread on THP flavor spikes and other sensory testing approaches.
• Milk The Funk thread on anecdotal experiences of THP forming in beer lines.
• General Milk The Funk Thread on March 10, 2015.
• General Milk The Funk Thread on Aug 25, 2015.
• General Milk The Funk Thread on 09/17/2015.
• Homebrewtalk thread started by "ne0t0ky0".
• Homebrewtalk thread started by "loctones", comments by Michael Tonsmeire.
• Babblebelt thread with comments by Chad Yakobson.
• Michael Tonsmeire observes that adding fresh Saccharomyces at bottling time seems to reduce THP production, versus force carbonating.
• Kyle Weniger from Joseph James Brewing Co noticed significantly improved THP reduction storing a split batch of beer at room temperature versus storing cold.
• Chad Yakobson on The Sour Hour, 3/11/2015 (around 50 minute mark) - indicates that he has been researching THP for 2 years, and is continuing this research. Little else is mentioned.