Analysis of proline in sour beer collected during barrel aging using liquid chromatography-mass spectrometry. Ashleigh Merrill, Jenna Watson, Kian Kruk, Emily Santa Ana, Teresa Longin, Chemistry Department, University of Redlands, Redlands, CA, 92373 Bryan Doty Sour Cellars, 9495 9th Street Unit B, Rancho Cucamonga, CA 91730
Confirming the Matrix Effect and Proline Concentrations
Peak Area for Proline
4.0E+07
3.0E+07 2.0E+07
Calibration curve for standards prepared in sour beer matrix
1.0E+07
y = 2.33E+05x - 2.56E+05 R² = 9.99E-01
0.0E+00 0
Time (min)
% Water
% ACN
0
99.0
1.00
9.99
89.3
10.7
12
10.00
90.0
14
99.00
1.00
20
99.00
1.00
Flow rate: 0.1 mL/min, column temp: 30°C, 1 L injection volume
Concentration as a Percentage by Mass %
Acetic Acid
0.0475%
Lactic Acid
0.579%
Ethanol
5.00%
We used a sour beer matrix to simulate the composition of sour beer samples. This matrix consisted of ethanol, lactic acid, and acetic acid, diluted in LCMS grade water. These percentages were determined by Alexis Cooper in a previous study at the University of Redlands.[10] Matrix matched standard solutions of dl-proline (Sigma-Aldrich, 99%), l-leucine (Sigma-Aldrich, >98%), and l-tyrosine (Sigma-Adlrich 99%) were prepared in the ranges of 2-140 ppm.
Quantifying Amino Acids [M+H]+ m/z 116-117
Proline
[M+Na]+ m/z 138-139
m/z of interest for proline
To compensate for background and overlapping peaks, the area in the chromatogram due to each amino acid was determined using the extracted ion chromatogram tool in the Agilent MassHunter software. Mass spectra of each amino acid standard were used to determine the principal masses in the chromatogram peak for a given amino acid. The MassHunter software then used those masses to extract the ion counts in the chromatogram due to those masses, providing an area in the chromatogram specific to a given amino acid. The following mass-to-charge ratios (m/z)were used for each amino acid: proline m/z 116-117 & 138-139; leucine m/z 132-33 and 154-155; tyrosine m/z 182-183 and 204-205.
40
50
120
Proline Concentration (g/mL)
Proline Concentration (g/mL)
120 100 80
60 40
20 0
0
50
100
Batch 113 Batch 130
150
200 250 300 Days from Brew
Batch 119 Batch 132
350
400
Batch 125 Batch 123
450
100
80 60 40 20 0
500
0
Batch 129
200 180 160 140 120 100 80 60 40 20 0 50
100
150
Batch 113
Batch 119
Batch 129
Batch 130
200 250 300 Days From Brew
350
Batch 125 Batch 132
400
450
500
Batch 123
Batch 129
50
100
Batch Batch 128 130
150
200 250 300 Days from brew
Batch 125
350
Batch 123
400
450
Batch 129
100 80 60 40 20 0 0
50
Batch 113 Batch 130
100 150 200 250 300 350 400 450 500 Days from Brew Batch 119 Batch 132
Proline
Batch 125 Batch 123
Batch 129
200 180 160 140 120 100 80 60 40 20 0 0
50 100 150 200 250 300 350 400 450 500 Days From Brew
Batch 113 Batch 129
Batch 119 Batch 130
Batch 125 Batch 132
Batch 123
30 25 20 15 10 5 0 0
50 100 150 200 250 300 350 400 450 500 Days from Brew
Batch 113 Batch 129
Batch 119 Batch 130
Leucine
Batch 123 Batch 132
Batch 125
Tyrosine
Batch 119
500 proline Batch 113
200 180 160 140 120 100 80 60 40 20 0 0
120
In addition to studying proline and leucine, we also used matrix-matched standards to quantify tyrosine concentrations in our samples. Plots of concentration versus days from brew for all seven batches for each amino acid are shown above. Batches 113 and 123 show very similar trends for all three amino acids. For both batches, proline levels increase slightly and then level off at relatively high concentrations after 150 days, while the concentrations of leucine and tyrosine are relatively moderate and remain steady over time. For batches 119 and 132, proline, leucine, and tyrosine concentrations tend to increase and then level off after about 150 days. However, batch 119 has relatively high concentrations of all three amino acids, while batch 132 has relatively low levels. For batch 125, the concentrations of all three amino acids increase slightly and then level off after about 100 days. However, the concentrations of proline remain relatively low with the concentrations of leucine and tyrosine remain high. Like batch 125, the levels of all three amino acids in batch 130 tend to increase over time and then level off after about 100 days. However, proline and leucine levels remain relatively low while the concentration of tyrosine is relatively high.
Proline results from Watson’s study[8] Proline results from this study Plots of the concentrations of proline vs. days from brew for the seven different batches obtained in this study using matrix-matched standards and obtained by Watson [8] using standards with water as the solvent are shown above. Both studies show that the proline concentrations tend to increase over time and then level off after about 150 days from brew for most batches. However, the proline concentrations obtained in this study are much higher than those obtained by Watson.[8] The overall concentration range for proline during the aging process in Watsons study was approximately 4.00 to 25.0 ug/mL, while the concentration range for our study was 11.9 to 109 ug/mL. This shows that the concentrations determined using matrix-matched standards give a much broader and larger range for proline concentrations in sour beer. Furthermore, the final concentration values found in this study fall within the final concentration ranges of proline in regular beer .[3-6] This confirms that matrixmatched standards must be used to produce accurate concentrations of amino acids in sour beer.
0
Beer Matrix Component
30
Leucine concentrations (g/mL)
Samples and standards were analyzed with an Agilent 6530 QToF LC/MS with a Zorbax C18 column. The analysis was done in positive ion mode with a 1 l injection volume.
20
Proline Concentration (g/mL)
Leucine concentrations (g/mL)
Detected Area for tryptophan peak
Sample Preparation and Experimental Conditions
LC gradient for amino acids
10
Picture Courtesy of Bryan Doty
To investigate the chemistry of sour beer during the aging process, Dr. Teresa Longin and Dr. David Soulsby of the University of Redlands partnered with Bryan Doty, head brewer at Sour Cellars. Doty collected samples every 2-4 weeks over the course of a year from seven different batches of sour beer. Emily Santa Ana studied the relationship between tryptophan and 1,2,3,4-tetrahydroharmane-3-carboxylic acid in the sour beer samples using liquid chromatography mass spectrometry (LC-MS).[7] Jenna Watson continued this research by quantifying levels of phenylalanine, leucine, and proline in the samples.[8] Watson’s results for proline were significantly lower than expected based on literature results for regular beer. We speculated that this could be due to interference from the organic acid matrix present in the sour beer. Kian Kruk [9] explored a possible matrix effect using tryptophan and tyrosine by adding a known amount of amino acid to solutions with varying Tryptophan: no matrix effect Tyrosine: matrix effect % by volume of sour beer matrix. The matrix was an average of organic 60000000 45000000 acid and ethanol concentrations found in the sour beer samples in a previous 40000000 50000000 study.[10] As shown in the figure, Kruk found that tryptophan’s LC-MS 35000000 40000000 30000000 signal was not affected by the sour beer matrix (right), but the LC-MS signal 25000000 30000000 for tyrosine was significantly suppressed in a sour beer matrix (left). 20000000 20000000 15000000 Tryptophan’s retention time is significantly longer than that of the organic 10000000 10000000 5000000 acids, while tyrosine’s retention time is only about a minute longer than that 0 0 of the organic acids. This suggests that if an amino acid’s retention time is % Sour Beer Matrix closer to that of the organic acids, then that amino acid’s signal will be % Sour Beer Matrix suppressed. The retention time of proline is almost the same as that of the organic acids, so proline is likely to experience a significant matrix effect. The goal of this study is to confirm that matrix effect for proline and to use matrix-matched standards to more accurately quantify proline and other amino acids. Detected Area for tyrosine peak
y = 1.47E+06x + 5.50E+06 R² = 9.46E-01
5.0E+07
Sour Beer Studies at the University of Redlands: Previous Work
1 mL of each sample was filtered through a 0.45 mm nylon filter into a sample vial.
Calibration curve for standards prepared in water
6.0E+07
To confirm the matrix effect for proline, we created a “mock beer” sample consisting of 40 g/mL of proline in a sour beer matrix. We also created proline standards using sour beer matrix as the solvent and another set of standards using water as the solvent. We analyzed the mock beer sample using LC-MS along with both sets of standards. We used the calibration curves from the two sets of standards to convert the proline peak area of the mock beer to the concentration of proline to see which set of standards reproduced the known concentration of proline in the mock beer. The water-based calibration curve gave back a concentration of 2.39 ug/mL while the sour beer matrix calibration curve gave a much more accurate concentration of 39.8 ug/mL for the 40.0 ug/mL mock beer sample. This results demonstrates that there is a significant matrix effect for proline and confirms that we must use matrix-matched standards to determine accurate concentrations of amino acids in our sour beer samples. The standard curves themselves (on the left) show a significantly diminished peak area of proline (as measure in ion counts) in the standards made with the sour beer matrix, further demonstrating that the organic acids suppress the proline signal.
Tyrosine concentrations (g/mL)
7.0E+07
Leucine concentrations (g/mL)
In recent years, sour beer has taken over the alcohol industry with its complex flavor profile.[1] The brewing process for sour and regular (non-sour) beer starts with water, malted barley, wheat, hops, and a mixture of yeast. For sour beer, brewers add bacteria like Lactobacillus or Pediococcus which produce organic acids during fermentation. Brewers then age the sour beer in barrels for several year. Yeast and bacteria can also produce trace compounds such as amino acids that provide subtle flavors. During the fermentation process the concentrations of amino acids vary due to the consumption or production of amino acids by microbes in beer.[2] In both regular (non-sour) and sour beer, amino acids aid in the development of umami, smokey, flavors. Consequently, amino acids are essential to the final flavor profile of a brew. Amino acid concentrations in regular beer have been well studied using high performance liquid chromatography (HPLC). [3,4,5,6]. In these studies, proline concentrations ranged from 31.8-663 ug/mL, leucine concentrations ranged 2.60 – 84.8, and tyrosine ranged from 1.00 – 71.3, indicating that amino acid concentrations can vary widely among various types of beers. All these studies focused on non-sour beers and looked at finished products rather than exploring amino acid concentrations during the aging process. In contrast, amino acids concentrations in sour beer during the aging process or in the final product are not well studied. Given the complexity of the process for creating sour beer, brewers would benefit from knowing how amino acid concentrations can vary over time for various brewing, fermenting and aging conditions.
Amino Acids in Sour Beer vs Regular Beer Results & Discussion Proline Concentration (g/mL)
Introduction
50
100
Batch128 130 Batch
150
200 250 300 days from brew
Batch 125
Batch 123
350
400
Batch 119
450
500
Batch 113
Leucine results from Watson’s study[8] Leucine results from this study Since Watson studied leucine concentrations [8], we also quantified leucine concentrations used matrix-matched standards. As for proline, the concentration trends for our study match Watson’s study. However, the concentration ranges are much higher with matrix-matched standards, suggesting that leucine also experiences a matrix effect due to organic acids in sour beer.
Conclusions and Future Work This work confirmed that the presence of organic acids in the sour beer matrix suppress the LC-MS signal for proline, establishing that we must use matrix-matched standards to obtain accurate concentrations for most amino acids. Our results for proline showed similar trends in concentration vs. days from brew to the results obtained by Watson [8], but we found much higher concentrations of proline. To further compare our results to Watson’s results, we quantified leucine in the sour beer samples using matrix-matched standards, and our trends in leucine concentration versus days from brew were similar to those obtained by Watson [8] but, again, our concentrations were higher. Finally, we used matrix-matched standards to find concentrations of tyrosine in the sour beer samples. In general, concentrations of the three amino acids increased over time and then leveled off for most batches. However, for batch 129, the concentration of tyrosine abruptly fell to undetectable levels while the concentrations of proline and leucine remained constant. In addition, batch 130 showed relatively low levels of leucine and proline but relatively high levels of tyrosine compared to other batches. This confirms that the biochemical processes controlling concentrations of amino acids can vary depending on the brewing conditions and the amino acids. We plan to use matrix-matched standards to study concentrations of more amino acids in sour beer to see if other amino acids exhibit some of the same variable trends as tyrosine. We will also compare trends for amino acids to trends in concentrations of organic acids and ethanol found by Cooper.[10] These comparisons will help shed light on the biochemical processes involved in fermentation and aging. Brewing of sour beer is an intricate process affected by numerous variables like starting ingredients, extraction and mashing times, barrel wood, temperature, humidity, and wild fermentation. These unknown variables contribute to a wonderfully complex brew but can result in quality inconsistency and difficulties with reproducibility. Having a more complete understanding of how the trace compounds in sour beer fluctuate throughout aging will allow brewers to understand how the brew changes in response to its starting conditions and environment. This work can give chemists and brewers a better understanding of why certain sour beers has different flavor profiles which gives sour beer brewers more control over the flavor of their final product.
Batch 129 shows very odd behavior for tyrosine. The concentrations of all three amino acids are relatively moderate and remain steady over time until day 385. At this point, the concentration of tyrosine drops dramatically to an undetectable level while proline and leucine concentrations remain constant. This was not due to an instrumentation issue, as the batch was run twice in the LCMS to ensure the instrumentation was not at fault. This trend will need to be further investigated.
Several past studies have determined amino acid concentrations in regular beers such as pilsners, lambics, pale ale, and Eastern European wheat beers.[3This study Kabelova [3] Erbe and Schad and Ferreira and 6] The table on the right summarizes ranges of Bruckner [4] Bollig [5] Gudio [6] concentrations proline, leucine, and tyrosine in the Concentration 28.0 – 104 31.8 – 250 91.7 - 663 224 - 392 45.7 studies on regular beers and the ranges we found using range for matrix-matched standards. Our concentrations for proline proline and tyrosine fall into the general range for (g/mL) proline found in regular, finished beer. This agreement Concentration 62.0 - 172 2.60 - 29.6 19.9 - 84.8 7.06 - 74.8 12.3 - 18.1 gives us confidence that using matrix-matched range for standards provides accurate concentrations for amino leucine acid concentrations in our sour beer samples. It also (g/mL) 5.70 - 71.3 1.00 - 71.3 24.0 – 54.0 6.21 – 10.6 suggests that microbial activity involving these amino is Concentration 4.20 – 25.8 range for similar in sour and regular beer. Our concentrations of tyrosine leucine tend to be a bit higher than the concentrations (g/mL) found in regular beer, so it’s possible that microbial activity involving leucine is slightly different between sour and regular beer. We plan to investigate other amino acids to see how ranges in our sour beer samples compare to those for regular beer. All the variation between the amino acids shows that there are indeed different concentrations for different amino acids within one batch and different trends in amino acid concentrations across batches. Each batch of sour beer has its own ecosystem due to the various amounts of added microbes from brewers, or from the barrel the beer was brewed in. This phenomenon will need to be further studied to give brewers more control over their final product, ensuring correct flavor production within sour beer brewing. We plan to look more closely at the conditions Bryan Doty used for brewing various batches to see if there are correlations between brewing conditions such as the relative amounts of wheat and barley, the types of hops used, etc. and amino acid trends. We also have information about trends in organic acid and ethanol concentrations from previous work[10] and will look at those trends for correlations with amino acid concentrations.
Acknowledgements & References We wish to thank the Hedco Foundation for the financial gift that provided the HPLC-QToF-MS used in this study. 1. Hurt, J. Beer Drinkers Are Growing Sweet on Sour Beers. https://www.forbes.com/sites/jeanettehurt/2020/03/11/beer-drinkers-are-growing-sweet-on-sourbeers/?sh=5b23f2795176. 2. Dysvik A.; La Rosa S. L.; De Rouck G.; Rukke E.; Westereng B.; Wicklund T.; Ercolini D. Microbial Dynamics in Traditional and Modern Sour Beer Production. American Society for Microbiology 2020, 86, e00566-20 3. Kabelová, I.; Dvořáková, M.; Čížková, H.; Dostálek, P.; Melzoch, K. Determination of Free Amino Acids in Beers: A Comparison of Czech and Foreign Brands. Journal of Food Composition and Analysis, 2008, 21, 736-741. 4. Erbe, T.; Bruckner, H.; Chromatographic determination of amino acid enantiomers in beers and raw materials used for their manufacture, Journal of Chromatography A, 2000, 881(1-2), 81-91. 5. Schad, G.J.; Bollig, B. Fast and Simple Determination of Free Amino Acids in Beer. The Column, 2015, 10. 6. Ferreira, I., Guido, L.; Impact of Wort Amino Acids on Beer Flavour: a Review. Fermentation. 2018, 4-23. 7. Santa Ana, E. Identifying and Quantifying Trace Compounds in Sour Beers Using High Pressure Liquid Chromatography-Mass Spectrometry. B.S. Thesis, University of Redlands, Redlands, CA, 2021 8. Watson J. Using Liquid Chromatography Mass Spectrometry to Quantify Amino Acid Content in Sour Beer. B.S. Thesis, University of Redlands, Redlands, CA, 2022. 9. Kruk, K. Quantification of Amino Acids in Sour Beer Using NMR and LC-MS. B.S. Thesis, University of Redlands, Redlands CA, 2024. 10.Cooper A. CRAFT Beer: Using quantitative 1H NMR and Complete Reduction to Amplitude-Frequency Tables (CRAFT) analysis to determine analyte development in sour beers. B.S. Thesis, University of Redlands, Redlands, CA, 2020