Sequentially Pitching Lactic Acid Bacteria and Active Dry Yeasts for Sour Beer Production

Abstract

Lactic acid bacteria are key constituents in the souring process of traditional acidic beers, resulting from spontaneous or mixed fermentations. The development of such beer styles provides a balanced and complex-flavored beverage but is time-consuming and not easily repeatable. New time-efficient alternatives include successively pitching a souring bacteria and yeast, such that both have optimal fermentation conditions. In this study, each microorganism is assessed to comprehend its individual impact on the fermentation process (acidification and fermentation rates, attenuation, ethanol production) and its influence on physicochemical and organoleptic parameters. Lactic acid bacteria are responsible for the wort’s acidification, reaching a pH of 3.5 within 48 h, by converting up to 5 g/L of simple sugars into organic acids and other by-products. The two-step acidification and fermentation process alters the color and bitterness, the latter decreasing from an initial iso-α-acid concentration of 29 mg/L to an average of 13 mg/L. The sequential method shows that the decarboxylation and subsequent reduction of hydroxycinnamic acid precursors into their vinyl and ethyl derivatives is strain dependent. All tested lactic acid bacteria appear to possess the decarboxylation enzymes. However, only Lactiplantibacillus plantarum can reduce vinylphenols into ethylphenols.

Keywords:

• Hydroxycinnamic acid
• lactic acid bacteria
• Lactiplantibacillus plantarum
• Levilactobacillus brevis
• sour beer
• sequential fermentation
• Saccharomyces cerevisiae
• volatile phenols

Introduction

The growing popularity of sour beers among consumers and breweries is reflected by increased and diversified offers on the market.^[1,2] Different techniques exist to produce such beer types, including spontaneous souring, mixed fermentation, and inoculation-based methods.

Traditional sour beers are fermented by a multitude of microorganisms present in the ambient air, mainly wild yeasts (e.g., Saccharomyces, Brettanomyces, and Kloeckera spp.), acetic acid bacteria, and lactic acid bacteria (LAB).^[3–6] The synergies between these microorganisms yield a unique and complex flavor profile to sour beers.^[5,7] The resulting acidity originates from the bacteria’s carbon metabolism producing predominantly lactic acid and acetic acid. The most notable examples of classic sour beers are the Belgian lambics, the Flanders red ales, and the German Berliner weisse. However, the intricate microbial dynamic requires a prolonged fermentation time of up to several years, and repeatability is hard to achieve.^[4,7]

Currently, brewers are investigating time-efficient and consistent alternatives to produce sour beers in a controlled manner. Modern approaches, such as sour mashing or kettle souring, temporally separate the pitching sequence of bacteria and ­yeasts.^[1,2,8,9] This avoids microbial competition and provides successive ideal fermentation conditions, whereby the absence of hop and alcohol is favorable for bacterial growth and the remaining available nutrients for yeast development.

For the kettle souring method, wort with no hop is briefly boiled, cooled to a temperature between 30 and 50 °C, and inoculated with a souring microbe in the brewing kettle.^[9,10] The wort is fermented until the desired level of acidity, which typically requires 24 to 48 h. Subsequently, the acidified wort is boiled to stop the souring process and hopped for bitterness purposes. Once cooled to 20 °C, the wort is transferred to a fermenter and inoculated with brewing yeast to initiate primary fermentation.^[8–13]

Lactobacilli, which are part of the LAB group, such as Lactobacillus delbrueckii, Lactobacillus amylovorus, and Lactiplantibacillus plantarum (formerly Lactobacillus plantarum), are commonly used as commercial starter cultures for acidification purposes in mixed fermentation.^[14,15] Certain lactobacilli strains are obligate (e.g., L. delbrueckii) or facultative (e.g., L. plantarum) homofermentative, converting hexoses almost exclusively into lactic acid by the Embden-Meyerhof-Parnas pathway. Others are obligate ­heterofermentative (e.g., L. brevis) forming lactic acid, ethanol, carbon dioxide, and acetic acid by the pentose phosphate pathway.^[16] Moreover, LAB contribute to the organoleptic and nutritional improvement of fermented products.^[8,9,14]

The generated acidity impacts both wort and beer’s physicochemical and sensory attributes. Sour beers have a brighter color compared to beers made with non-acidified wort. Indeed, Maillard browning and caramelization are temperature-dependent, nonenzymatic reactions that occur faster at higher pH values.^[12,17] Peyer et al. reveal that acidic conditions cause a delay in yeast growth and performance without influencing the complete attenuation of sour beers.^[12] The acidity leads to a weak hop isomerization and a lower splitting rate of dimethyl sulfide precursors.^[18–20] Boiling of an acidified wort drives off volatile aromatics produced during the souring process, yielding beers that are described and criticized for lacking depth, complexity, and character.^[1,13] Nevertheless, the practice of souring wort before boiling emerges as the best compromise to obtain a sour beer with high acidity and minimal organoleptic failures while avoiding the risk of unintentional contamination of subsequent batches of non-sour beer.

Kettle souring, using LAB as pre-fermentation starters, is largely used as a sour beer production method.^[2,8,9] Little is documented about the sensory contribution from the lactobacilli in the final beer obtained by a sequential two-step fermentation. Therefore, this study aims to compare the acidification power of four lactobacilli, selected for their high hop sensitivity, and to study the ability of an array of Saccharomyces strains (lager, ale, neutral, fruity, phenolic) to ferment an acidified wort. Also, it seeks to comprehend the consequences of a kettle souring process on the sour beer’s aromatic profile (bitterness, phenols, acidity). To evaluate the sour beer’s technological and final quality attributes, each fermentation step is assessed regarding carbon degradation, acid production, and other by-product concentrations.

Experimental

Wort and beer production

For the wort preparation, malted pilsner (Weyermann®, Bamberg, Germany) was milled and mixed with water at a rate of 1:4 (w/v). At the mash-in, 0.02% (v/v) of lactic acid (Sigma®, Darmstadt, Germany) was added to reach a pH of 5.5. The mashing profile was as follows: 60 min at 63 °C, 30 min at 72 °C at the end of which an iodine mash conversion test was performed, and finally, 5 min at 78 °C. The wort was separated from the spent milled grains by means of a lauter tun. The spent grains were sparged with water until the wort reached a specific gravity of 1.054. The wort was boiled for 5 min and then cooled, either at 40 °C or 30 °C, ready for bacterial acidification.

Once the wort was cooled at the appropriate temperature, four dried bacteria selected from Institut Meurice’s bacteria collection (IM-b), namely one Levilactobacillus brevis (IM-b 372, referenced as Lb hereafter) and three Lactiplantibacillus plantarum (IM-b 393, IM-b 037, and IM-b 164 cited as Lp1, Lp2, and Lp3, respectively), were pitched at 5 g/hL into three vessels of 1.8 L each. The pH was regularly monitored, and the acidification process, which should reach a pH of 3.5, was stopped after 48 h. For each temperature condition, a non-acidified control sample was preserved.

The samples that were acidified at 30 °C were then boiled for 60 min, hopped with 1.0 g/L of Perle pellets (Castle Malting®, Beloeil, Belgium) containing 9.4% α-acids, and were subsequently cooled to 20 °C. The specific gravity was corrected to 1.054 by the addition of sterile distilled water.

The hopped non-acidified (control) and hopped acidified samples were inoculated separately with a total of 11 brewer’s yeast strains (Fermentis®, Marquette-lez-Lille, France). Saccharomyces pastorianus was pitched at 100 g/hL, and fermentations were carried out at 14 °C, whereas Saccharomyces cerevisiae was pitched at 50 g/hL, and fermentations were carried out at 20 °C. All trials were performed in triplicates of 100 mL. The ethanol produced was monitored, and the primary fermentation was halted when the increase in ethanol concentration was less than 5% (v/v) for two consecutive days.

Wort and beer analyses

After the lactic fermentation, after the second boil, and after the primary fermentation, several analyses were performed on each experimental condition according to the standard methods of Analytica EBC (European Brewing Convention).

Wort viscosities were obtained with a microviscometer Lovis 2000 M (Anton Paar®, Graz, Austria) of the falling ball type (EBC method 8.4).

Wort and beer specific gravities and extracts were measured with a DMA 4500 digital density meter (Anton Paar®, Graz, Austria) of the oscillating type (EBC methods 8.2.2, 8.3, 9.43.2 & 9.4).

Fermentable carbohydrates were measured using an Alliance HPLC (Waters^TM, Milford MA, U.S.A.) equipped with an Asahipak NH2P-50 4E (4.6 × 250 mm) column, maintained at 30 °C and a Waters^TM 2414 refractive index detector. The mobile phase was a 75/25 acetonitrile/water solution with an isocratic flow of 1 mL/min. Samples were diluted when necessary and filtered on a 0.45 µm membrane, and 10 µL aliquots were injected (EBC methods 8.7 & 9.27).

Organic acid concentrations were measured using an Alliance HPLC (Waters^TM, Milford MA, U.S.A.) equipped with a Bio-Rad HPX-87H (7.8 × 300 mm) column, maintained at 30 °C and a Waters^TM 2489 UV/Vis detector measuring at 210 nm. The mobile phase was a 0.003 M H₂SO₄ solution with an isocratic flow rate of 0.4 mL/min. Samples were filtered on a 0.45 µm membrane, and 20 µL aliquots were injected.

A Genesys 150 UV-visible spectrophotometer (Thermo Fisher Scientific®, Waltham MA, U.S.A.) was used to determine the sample’s color at 430 nm (EBC methods 8.5 & 9.6), bitterness at 275 nm after an isooctane extraction of the bitter substances, mainly iso-α-acids (EBC methods 8.8 & 9.8), free amino nitrogen (FAN) at 570 nm after a colorimetric reaction with ninhydrin (EBC methods 8.10.1 & 9.10.1), and total polyphenols at 600 nm after a reaction with ferric ions of an alkaline solution (EBC methods 8.12 & 9.11).

The pH was measured by a SevenCompact pH meter (Mettler Toledo®, Columbus OH, U.S.A.) (EBC ­methods 8.17 & 9.35).

The volatile phenol analyses were outsourced to an accredited enology laboratory (Laboratoires Dubernet, Montredon-des-Corbières, France), in which an approved SPME-GC/MS method was applied.

Beer alcohol contents were obtained with an Alcolyzer ME (Anton Paar®, Graz, Austria), a near-infrared spectroscopy meter (EBC method 9.2.6).

Statistical analyses

Statgraphics Centurion Version 18.1.12 was used for statistical calculations. One-way ANOVA was used to compare mean values between the samples. When F values were found to be significant, Fisher’s least significant difference (LSD) procedure was used to further determine any significant differences between the trials. The level of significance was determined at P < 0.05. Results are expressed as mean ± standard deviation.

Results and discussion

Wort acidification

The pH evolution during the lactic fermentation carried out at 30 °C (Figure 1) displays that all the selected LAB reach, from an initial wort pH of 5.5, the target pH of approximately 3.5 within 48 h. The acidification process was associated with the production of organic acids (Figure 2B), mainly lactic and acetic acids, at different concentrations depending on the LAB's carbon metabolism. Indeed, two of the homofermentative Lactiplantibacillus plantarum (Lp1 and Lp2) produced predominantly lactic acid with respectively 4.03 (±0.09) and 3.55 (±0.01) g/L after 48 h. The heterofermentative Levilactobacillus brevis (Lb) yielded 2.99 (±0.33) g/L of lactic acid and produced the most acetic acid, as well as 0.10% (v/v) of ethanol. Finally, Lactiplantibacillus plantarum (Lp3) was the least efficient as only 2.64 (±0.40) g/L of lactic acid was produced.

Figure 1. Acidification kinetics. pH levels in wort throughout lactic fermentation performed by different LAB strains (13.5°P wort, 5 g/hL pitching rate, 30 °C fermentation temperature, 48h fermentation time). Each value is expressed as a mean ± standard deviation of triplicates from three independent brews. Nomenclature: Lb: Levilactobacillus brevis; Lp1, Lp2, Lp3: Lactiplantibacillus plantarum. At the end of the acidification, a different annotation denotes a significant difference at P < 0.05 (Fisher’s LSD procedure).

Figure 2. Sugar (A) and organic acid (B) concentrations. Concentrations in the non-acidified (control) and acidified worts, lactic ­fermentation performed by different LAB strains (13.5°P wort, 5 g/hL pitching rate, 30 °C fermentation temperature, 48h fermentation time). Each value is expressed as a mean ± standard deviation of triplicates from three independent brews. Nomenclature: CT: non-acidified (control) wort; Lb: Levilactobacillus brevis; Lp1, Lp2, Lp3: Lactiplantibacillus plantarum. For each variable, a different annotation denotes a significant difference at P < 0.05 (Fisher’s LSD procedure).

During lactic fermentation, the LAB consumed a total of 3 to 5 g/L of glucose and fructose (Figure 2A), concurring with the theoretical carbon intake of each strain. Therefore, the remaining fermentable disaccharides and trisaccharides were available for the subsequent alcoholic fermentation.

All tested LAB had similar acidification behaviors at a fermentation temperature of 40 °C, except for Lp3, whose acidification power was less efficient (data not shown). Therefore, no further investigation was carried out regarding the latter.

Acidified wort characterization

Various parameters and their respective values for each retained condition are provided in Table 1. It reveals that no significant FAN quantity was consumed during the bacterial acidification, that total polyphenol (PP) concentrations were constant, and that viscosities, relative to the sample’s acidity, were negligibly modified.

Table 1. Non-acidified (control) and acidified wort parameters. Each parameter is expressed as a mean ± standard deviation of triplicates of three independent brews.

		OG	F. sugars	FAN	PP	Viscosity
Sample	pH	[°P]	[g/L]	[mg/L]	[mg/L]	[mPa.s]
CT	5.60 ± 0.09^a	13.58 ± 0.16^a	91.60 ± 9.28^a	234 ± 21^a	214 ± 20^a	1.8660 ± 0.0113^a
Lb	3.71 ± 0.10^b	13.54 ± 0.05^a	85.80 ± 4.84^ab	219 ± 19^a	192 ± 26^a	1.9040 ± 0.0021^b
Lp1	3.43 ± 0.06^c	13.48 ± 0.18^a	76.58 ± 4.43^b	213 ± 27^a	190 ± 07^a	1.9640 ± 0.0028^c
Lp2	3.62 ± 0.07^b	13.54 ± 0.19^a	86.78 ± 9.08^ab	223 ± 56^a	231 ± 81^a	1.9360 ± 0.0027^d

Nomenclature: OG: original gravity; F. sugars: fermentable sugars; FAN: free amino nitrogen; PP: total polyphenols; CT: non-acidified (control) wort; Lb: Levilactobacillus brevis; Lp1, Lp2: Lactiplantibacillus plantarum. For each variable, a different superscript in each column denotes a significant difference at P < 0.05 (Fisher’s LSD procedure).

Acidified and hopped wort characterization

Following the lactic fermentation, the acidified worts were boiled and hopped with a target of 30 International Bitter Unit (IBU). Consequently, all wort components were concentrated, iso-α-acids were solubilized, procuring bitterness, and thermal reactions such as Maillard browning occurred.

For an identical hop addition, the iso-α-acid concentration (Figure 3A) was reduced by 45–55% in the acidified worts in comparison to the non-acidified control wort. This observation concurred with the limited solubility of α-acids and their incomplete isomerization in acidic conditions.^[18,19]

Figure 3. Post-acidified and hopped wort bitterness (A) and color (B). Concentrations in the non-acidified (control) and acidified worts, lactic fermentation performed by different LAB strains (13.5°P wort, 5 g/hL pitching rate, 30 °C fermentation temperature, 48h fermentation time, 1h boiling time, 1 g/L added hop). Each value is expressed as a mean ± standard deviation of triplicates from three independent brews. Nomenclature: CT: non-acidified (control) wort; Lb: Levilactobacillus brevis; Lp1, Lp2: Lactiplantibacillus plantarum. For each variable, a different annotation denotes a significant difference at P < 0.05 (Fisher’s LSD procedure).

Color-wise, the control condition was slightly darker than the acidified worts (Figure 3B). The nonenzymatic Maillard browning was less efficient at a more acidic pH.^[17] Moreover, the availability of the reaction’s substrates was diminished in acidified worts. The reducing sugars were partially consumed by the LAB, and the amino groups were protonated at low pH, thus less nucleophilic and not accessible to react and form the Schiff base intermediate.^[17,21,22]

The volatile phenol profiles (Figure 4) illustrate that both LAB and thermal reactions were involved in the decarboxylation of hydroxycinnamic acid precursors into their vinyl derivatives and subsequently, in the reduction of vinyl to ethyl derivatives.^[23] During the boiling process of the control condition, the decarboxylation occured by thermal reactions as no LAB were present and provided 23 µg/L of 4-vinylphenol (4VP) and 37 µg/L of 4-vinylguaiacol (4VG). In the presence of LAB, 4VP and 4VG were produced by the decarboxylation of p-coumaric acid and ferulic acid, respectively. However, the relative activity of the phenolic acid decarboxylase (PDC) on the two precursors is strain dependent.^[24] When acidified with L. brevis, 4VP was predominantly present before boiling, at a concentration of 196 µg/L, whereas when acidified with either L. plantarum, 4-ethylphenol (4EP) was the major component. All tested LAB appeared to have a decarboxylase enzyme that initiated the hydroxycinnamic acids metabolism, but only L. plantarum had the enzymatic activity to reduce 4VP into 4EP. These observations concur with previous studies that demonstrate that the PDC activity is more widely distributed among LAB than the ability to reduce vinylphenols into ethylphenols.^[25] Additionally, an enzyme responsible for vinylphenol reduction, the VprA protein, has been identified and biochemically characterized in L. plantarum but remains unknown.^[26] The expression of the vprA gene is induced by both its substrate (4VP) and by p-coumaric acid. This could explain the lack of 4-ethylguaiacol (4EG) formation in favor of 4EP production.

Figure 4. Volatile phenol concentrations. Concentrations are expressed as a mean ± standard deviation of three independent ­measurements. Nomenclature: PreB: pre-boiling; PostB: post-boiling; 4VP: 4-vinylphenol; 4VG: 4-vinylguaiacol; 4EP: 4-ethylphenol; 4EG: 4-ethylguaiacol; CT: non-acidified (control) wort, Lb: Levilactobacillus brevis; Lp1, Lp2: Lactiplantibacillus plantarum. For each variable, except 4EG, a different annotation denotes a significant difference at P < 0.05 (Fisher’s LSD procedure). Statistical analyses are not provided for 4EG as its concentrations are below the quantification limit.

Since both L. plantarum had similar flavor-active volatile phenol profiles but with different intensities, only Lp1 was kept for further investigation.

Alcoholic fermentation kinetics

After the acidification process, all the remaining conditions were inoculated with 11 separate brewer’s yeasts, either top-fermented, of which some were diastaticus, or bottom-fermented strains. The alcoholic fermentations were performed once in triplicate and are illustrated in Figure 5. For a similar wort, the maltotriose negative S-33 had a considerably lower attenuation (cf. Table 2), the diastaticus species did not reach a fermentation limit in 12 days, and the neutral US-05 had an initial lag phase. The yeast’s activity is very slightly impacted by the low wort pH of 3.5 but was, however, not influenced by the pitching method: direct pitch or rehydrated (data not shown).

Figure 5. Alcoholic fermentation kinetics. Ethanol content throughout the alcoholic fermentation performed by different brewer’s yeast (13.5°P wort, 50 g/hL and 100 g/hL pitching rate for respectively ale and lager strains, 20 °C and 14 °C fermentation temperature for respectively ale and lager strains, 8–12 days fermentation time). Each value is expressed as a mean of triplicates from one experimental trial. Nomenclature: CT: non-acidified control; Lb: Levilactobacillus brevis; Lp1: Lactiplantibacillus plantarum.

Table 2. Physicochemical parameters of the fermented sour beer. Alcoholic fermentation performed by different yeast strains (13.5°P wort, 50 g/hL and 100 g/hL pitching rate for respectively ale and lager strains, 20 °C and 14 °C fermentation temperature for respectively ale and lager strains, 8–12 days fermentation time).

Sample	pH*	Alcohol*	AE*	ADF*	Color*	Bitterness*	FAN**	PP**
		[%v/v]	[°P]	[%]	[EBC]	[IBU]	[mg/L]	[mg/L]
CT_K-97	4.14 ± 0.11	5.33 ± 0.01	3.87 ± 0.01	71.86 ± 0.09	9.5 ± 0.1	24.4 ± 1.9	43 ± 3	200.1 ± 30.2
CT_S-33	4.42 ± 0.11	4.42 ± 0.01	5.44 ± 0.01	60.00 ± 0.07	9.2 ± 0.1	24.6 ± 1.9	83 ± 3	195.2 ± 29.5
CT_T-58	4.19 ± 0.11	5.39 ± 0.01	3.76 ± 0.01	72.66 ± 0.09	8.8 ± 0.1	24.6 ± 1.9	52 ± 7	196.0 ± 29.6
CT_US-05	4.11 ± 0.11	5.38 ± 0.01	3.65 ± 0.01	73.23 ± 0.09	9.1 ± 0.1	24.7 ± 2.0	49 ± 2	190.2 ± 28.7
CT_BE-256	4.11 ± 0.11	5.42 ± 0.01	3.63 ± 0.01	73.49 ± 0.09	9.3 ± 0.1	23.1 ± 1.8	42 ± 2	195.2 ± 29.5
CT_S-04	4.16 ± 0.11	5.41 ± 0.01	3.65 ± 0.01	73.33 ± 0.09	8.9 ± 0.1	22.9 ± 1.8	39 ± 3	193.1 ± 29.2
CT_BE-134	4.03 ± 0.10	6.60 ± 0.01	1.32 ± 0.00	90.34 ± 0.11	8.2 ± 0.0	22.5 ± 1.8	37 ± 1	197.6 ± 29.8
CT_WB-06	3.92 ± 0.10	5.96 ± 0.01	2.46 ± 0.01	81.85 ± 0.10	9.3 ± 0.1	25.3 ± 2.0	43 ± 3	201.7 ± 30.5
CT_S-23	4.15 ± 0.11	5.32 ± 0.01	3.89 ± 0.01	71.69 ± 0.09	9.8 ± 0.1	24.2 ± 1.9	49 ± 2	202.1 ± 30.5
CT_S-189	4.17 ± 0.11	5.29 ± 0.01	3.93 ± 0.01	71.39 ± 0.09	9.8 ± 0.1	23.7 ± 1.9	50 ± 1	200.1 ± 30.2
CT_W-34/70	4.17 ± 0.11	5.32 ± 0.01	3.90 ± 0.01	71.62 ± 0.09	9.8 ± 0.1	24.5 ± 1.9	55 ± 1	196.8 ± 29.7
Lb_K-97	3.40 ± 0.09	5.08 ± 0.01	3.92 ± 0.01	70.64 ± 0.09	8.5 ± 0.0	14.2 ± 1.1	68 ± 6	193.5 ± 29.2
Lb_S-33	3.51 ± 0.09	4.19 ± 0.01	5.52 ± 0.01	58.37 ± 0.07	8.6 ± 0.1	13.1 ± 1.0	55 ± 2	184.9 ± 27.9
Lb_T-58	3.43 ± 0.09	5.10 ± 0.01	3.82 ± 0.01	71.25 ± 0.09	8.1 ± 0.0	13.1 ± 1.0	61 ± 4	190.2 ± 28.7
Lb_US-05	3.41 ± 0.09	5.17 ± 0.01	3.74 ± 0.01	72.00 ± 0.09	8.0 ± 0.0	15.3 ± 1.2	74 ± 5	205.4 ± 31.0
Lb_BE-256	3.49 ± 0.09	5.15 ± 0.01	3.66 ± 0.01	72.34 ± 0.09	8.2 ± 0.0	12.3 ± 1.0	52 ± 2	184.9 ± 27.9
Lb_S-04	3.46 ± 0.09	5.24 ± 0.01	3.72 ± 0.01	72.33 ± 0.09	8.0 ± 0.0	13.7 ± 1.1	55 ± 2	187.4 ± 28.3
Lb_BE-134	3.39 ± 0.09	6.32 ± 0.01	1.48 ± 0.00	88.89 ± 0.11	7.8 ± 0.0	14.8 ± 1.2	48 ± 2	173.4 ± 26.2
Lb_WB-06	3.35 ± 0.09	5.89 ± 0.01	2.31 ± 0.01	82.66 ± 0.10	9.1 ± 0.1	16.3 ± 1.3	43 ± 4	205.0 ± 31.0
Lb_S-23	3.40 ± 0.09	5.07 ± 0.01	3.89 ± 0.01	70.77 ± 0.09	8.0 ± 0.0	12.4 ± 1.0	49 ± 2	200.9 ± 30.3
Lb_S-189	3.40 ± 0.09	5.04 ± 0.01	3.90 ± 0.01	70.57 ± 0.09	8.5 ± 0.1	13.0 ± 1.0	50 ± 1	193.5 ± 29.2
Lb_W-34/70	3.44 ± 0.09	5.09 ± 0.01	3.87 ± 0.01	70.97 ± 0.09	7.9 ± 0.0	13.2 ± 1.0	48 ± 3	184.1 ± 27.8
Lp1_K-97	3.26 ± 0.08	5.15 ± 0.01	4.02 ± 0.01	70.35 ± 0.09	7.2 ± 0.0	12.0 ± 0.9	42 ± 6	201.7 ± 30.5
Lp1_S-33	3.36 ± 0.09	4.27 ± 0.01	5.59 ± 0.01	58.51 ± 0.07	6.6 ± 0.0	10.9 ± 0.9	29 ± 4	198.4 ± 30.0
Lp1_T-58	3.37 ± 0.09	5.17 ± 0.01	4.03 ± 0.01	70.39 ± 0.09	7.1 ± 0.0	12.7 ± 1.0	33 ± 6	error ± error
Lp1_US-05	3.26 ± 0.08	5.16 ± 0.01	3.81 ± 0.01	71.56 ± 0.09	7.2 ± 0.0	13.1 ± 1.0	62 ± 4	189.4 ± 28.6
Lp1_BE-256	3.41 ± 0.09	5.26 ± 0.01	3.75 ± 0.01	72.25 ± 0.09	8.4 ± 0.0	11.9 ± 0.9	46 ± 3	183.7 ± 27.7
Lp1_S-04	3.31 ± 0.08	5.31 ± 0.01	3.80 ± 0.01	72.15 ± 0.09	8.2 ± 0.0	12.5 ± 1.0	42 ± 4	188.6 ± 28.5
Lp1_BE-134	3.29 ± 0.08	6.30 ± 0.01	1.91 ± 0.00	86.01 ± 0.11	6.8 ± 0.0	12.9 ± 1.0	44 ± 6	202.5 ± 30.6
Lp1_WB-06	3.25 ± 0.08	5.97 ± 0.01	2.38 ± 0.01	82.36 ± 0.10	8.9 ± 0.1	13.6 ± 1.1	44 ± 4	186.1 ± 28.1
Lp1_S-23	3.28 ± 0.08	5.28 ± 0.01	4.04 ± 0.01	70.76 ± 0.09	7.5 ± 0.0	11.9 ± 0.9	62 ± 9	187.4 ± 28.3
Lp1_S-189	3.29 ± 0.08	5.27 ± 0.01	4.06 ± 0.01	70.60 ± 0.09	7.8 ± 0.0	11.9 ± 0.9	57 ± 5	194.8 ± 29.4
Lp1_W-34/70	3.30 ± 0.08	5.30 ± 0.01	4.05 ± 0.01	70.75 ± 0.09	7.6 ± 0.0	11.6 ± 0.9	70 ± 7	194.3 ± 29.3

*Parameters are expressed as one value from a unique experimental trial ± instrumental uncertainty. **Each value is expressed as a mean ± standard deviation of three independent measurements from one experimental trial. Nomenclature: AE: apparent extract; ADF: apparent degree of fermentation; FAN: free amino nitrogen; PP: total polyphenols; CT: non-acidified (control) beer; Lb: Levilactobacillus brevis; Lp1: Lactiplantibacillus plantarum.

Green beer characterization

Table 2 provides fundamental physicochemical parameters that characterize the final fermented products.

The apparent degree of fermentation (ADF) reflects the yeast’s carbon metabolism and is strain dependent. It is observed that independently of the wort’s initial pH, all species reached their unique theoretical attenuation range. The alcohol produced differed as LAB partially converted simple sugars during lactic fermentation. The yeast’s sensory impact on the final sour beer lies in the ethanol content and the residual extract, dependent on whether the strain is diastaticus or not.

The FAN concentrations decreased significantly during the alcoholic fermentation going from approximately 210 mg/L to 50 mg/L for all experimental conditions, whereas the polyphenol concentrations remained constant around 200 mg/L.

As expected, the solubilized iso-α-acid concentrations (Figure 6) diminished during the yeast’s fermentation by an average of 15.5% for the control conditions and 9.5% for the acidified beers, as they were partially adsorbed by the yeast.^[27,28]

Figure 6. Non-acidified and sour beer bitterness. Iso-α-acid concentrations in the fermented products, alcoholic fermentation performed by different yeast strains (13.5°P wort, 50 g/hL and 100 g/hL pitching rate for respectively ale and lager strains, 20 °C and 14 °C fermentation temperature for respectively ale and lager strains, 8–12 days fermentation time). Each value is expressed as one value from a unique experimental trial ± instrumental uncertainty. Nomenclature: CT: non-acidified (control) beer; Lb: Levilactobacillus brevis; Lp1: Lactiplantibacillus plantarum.

During alcoholic fermentation, the color intensity of the control conditions reduced by 7.5–22.6%, dependent on the yeast strain employed. Indeed, yeasts can remove hue through the adsorption of colored polyphenols and melanoidins on their cell walls.^[29] The pigment adsorption mechanism is variable among S. cerevisiae strains, and the trait displays a polygenic inheritance.^[30] Conversely, the color of the acidified beers remained relatively constant, in comparison to the acidified wort. Acidic conditions modify the pigmented molecule’s polarity, altering the interactions with the yeast.^[31] These conditions could also change the yeast surface structure, thus further hindering the pigment adsorption process.

The 4VP concentrations were measured at 23 µg/L, 142 µg/L, and 15 µg/L after lactic fermentation, for respectively the control condition, the Lb acidified wort, and the Lp1 acidified wort. As illustrated in Figure 7A, these concentrations were enhanced when worts were fermented by phenolic off-flavor (POF) positive yeast strains (BE-134, T-58, and WB-06) capable of decarboxylating p-coumaric acid into its vinyl derivative. In comparison to the control beer, the proportion of 4VP produced by the yeast’s activity was diminished in acidic beer. As Lb does not have the enzymatic tools to reduce 4VP into 4EP, it must convert the p-coumaric acid precursor by means of an alternative metabolic pathway. As previously observed, Lp1 possesses the reductase to produce 4EP. This is supported by the fact that 4EP is present in beers soured with Lp1, independently of the yeast used for fermentation, at a constant concentration of approximately 175 µg/L (data not shown), identical to the acidified wort’s concentrations (Figure 4).

Figure 7. Volatile 4-vinylphenol (A) and 4-vinylguaiacol (B) concentrations. Concentrations in the fermented products, alcoholic fermentation performed by different yeast strains (13.5°P wort, 50 g/hL and 100 g/hL pitching rate for respectively ale and lager strains, 20 °C and 14 °C fermentation temperature for respectively ale and lager strains, 8–12 days fermentation time). Concentrations are expressed as one value from a unique experimental trial ± confidence interval of 95%. Nomenclature: CT: non-acidified (control) beer; Lb: Levilactobacillus brevis; Lp1: Lactiplantibacillus plantarum.

Regarding the 4VG concentrations (Figure 7B), the POF-positive strains were distinguishable as they had extreme concentrations. In comparison to the control beer, the proportion of 4VG produced by the yeast’s activity was constant when acidified with Lb and less when acidified with Lp1. This implies that Lb converts the ferulic acid precursor exclusively in its vinyl derivative, whereas Lp1 may use it to produce vanillin^[32] and other by-products.^[23] Nevertheless, when fermented with yeasts, other than POF-positive strains, 4VP and 4VG concentrations remained under their perception threshold of 200 µg/L and 300 µug/L, respectively.^[33]

Conclusions

The increasing interest in sour beer has incited research to develop an alternative process to reduce production time. One possible practice involves the successive pitching of a souring bacteria and yeast, in which each microorganism grows individually under optimal conditions. The advantage of a two-step fermentation is that the process is reliable and reproducible. Nevertheless, the major drawback is the lack of flavor complexity.

Although modern sour beer production has largely been studied; little is documented about the relative impact of each microorganism on sensory perception and physicochemical parameters. This study shows that the tested LAB can acidify a classic unhopped 13.5°P wort, from a pH of 5.5 to a pH of 3.5 in 48 h at 30 °C. The Saccharomyces behavior is not altered by these acidic conditions, and the alcoholic fermentations reach the expected alcohol yield. Additionally, bacterial acidification reduces the wort’s color and bitterness, which corroborates results from previous publications.^[12,13]

The volatile phenols are, possibly, the most rewarding results as each microorganism affects the phenol profile of the final sour beer. The Lb bacteria are able to decarboxylate the p-coumaric and ferulic acid precursors into vinyl derivatives, whereas both Lp1 and Lp2 strains can further reduce vinyl to ethyl derivatives. The POF-positive yeast strains accentuate the decarboxylation phenomena through the regulation of the pad1 gene,^[34] which provides clove-like, spicy, and smoky flavors to the beer. Additional research is required to comprehend the bacterial vprA gene’s regulation^[26] and its corresponding enzymes responsible for the reduction of vinyl to ethyl derivatives, as this is not yet known or characterized.

Previous studies have elaborated basic sensory profiles for sour beers^[8,9] excluding volatile compounds such as phenols. Further investigations could elucidate alternative metabolic pathways used by Lb bacteria to transform vinylphenols and identify their end products. Their detection could further contribute to the continuous descriptions of sour beers and complete the organoleptic profiles of such complex beverages.

Disclosure statement

No potential conflict of interest was reported by the authors.