Utilization of plant derived lactic acid bacteria for efficient bioconversion of brewers' spent grain into acetoin

Brewers ’ spent grain (BSG) is a major side-stream from the beer industry, with an annual estimated production of 39 million tons worldwide. Due to its high nutritional value, high abundance and low price, it has been proposed as an ingredient in human food. Here we investigated the ability of different lactic acid bacteria to produce the flavor molecule acetoin in liquid BSG extract, in order to broaden the possibilities of utilization of BSG in human food. All the investigated lactic acid bacteria (LAB) covering the Leuconostoc , Lactobacillus and Lactoccocus species were able to convert the fermentable sugars in liquid BSG into acetoin. Production levels varied significantly between the different LAB species, with Leuconostoc pseudomesenteroides species reaching the highest titers of acetoin with only acetate as the main byproduct, while also being the fastest consumer of the fermentable sugars present in liquid BSG. Surprisingly, the currently best investigated LAB for acetoin production, L. lactis , was unable to consume the maltose fraction of liquid BSG and was therefore deemed unfit for full conversion of the sugars in BSG into acetoin. The production of acetoin in Leu . pseudomesenteroides was pH dependent as previously observed in other LAB, and the conversion of BSG into acetoin was scalable from shake flasks to 1 L bioreactors. While all investigated LAB species produced acetoin under aerobic conditions, Leu . pseudomesenteroides was found to be an efficient and scalable organism for bioconversion of liquid BSG into a safe acetoin rich food additive.


Introduction
Brewers' spent grain (BSG) is a major side stream from the beer brewing industry, where it accounts for around 85 % of the total byproducts generated (Mussatto et al., 2006).BSG is the solid fraction, left after filtration of the beer wort.For every liter of beer produced, it is estimated that around 200 g of BSG is generated, yielding an annual estimated production of 39 million metric tons (Djukić-Vuković et al., 2016;Lynch et al., 2016).The primary use of BSG is currently as animal feed; however, as it is a renewable, low cost, and highly abundant sidestream, its utilization has been investigated for other purposes (Bolwig et al., 2019 andBuffington, 2014;).A promising approach is to utilize BSG in human feed and thereby to increase its value, as BSG is a high fiber (cellulose and hemicellulose) (30-50 %) and high protein (20-30 %) product, that contains several, potentially beneficial, compounds for human health such as phenolic compounds with antioxidant activities (Lynch et al., 2016).It has been proposed to include BSG in dairy products such as cheeses (Abd El-Moneim et al., 2018) and yogurts (Abd El-Moneim et al., 2015), in dairy alternatives (Madsen et al., 2021) and baked goods such as bread (Sahin et al., 2021) and cookies (Ajanaku et al., 2011).
Acetoin (3-hydroxybutan-2-one) is an important flavor compound in dairy products such as yogurt, cheese and butter.Its odor has been described as similar to yogurt and its taste as creamy and butter-like (Burdock, 2009).It is used as a food additive in many different food industries, including baked goods and dairy imitating products, amongst others (Xiao and Lu, 2014).Acetoin has been proposed as a promising renewable platform chemical (Zhang et al., 2016), and could be a way to increase the value of BSG by converting its available sugars into acetoin.Being able to convert the sugars in BSG into acetoin could be useful in biorefinery approaches to generate sustainable, pure, food-grade flavor ingredients, but could also be a way to enhance the flavor and thereby applicability of BSG for direct use in human food, especially given that many of the currently proposed BSG products involve baked and dairy products.A sustainable method to enrich BSG with acetoin without any additives can be through microbial fermentation.
Lactic acid bacteria (LAB) are known to produce acetoin and are responsible for the acetoin production in different dairy products such as yogurt and cheese (Xiao and Lu, 2014).Many species of LAB are labelled as Generally Recognized as Safe (GRAS) due to their wide historical use in various food matrices and are present in both the European Food Safety Authority's (EFSA) "qualified presumption of safety" (QPS) list and the Food and Drug Administration's (FDA) GRAS list, making them suitable for food fermentations.Furthermore, LAB can ferment sugars efficiently to fermentative products with high yields and can naturally utilize an extensive range of pentose and hexose sugars (Sauer et al., 2017), which makes them ideal for bioconversion of lignocellulosic derived materials such as BSG.The acetoin production of different LAB such as Leuconostoc (Levata-Jovanovic and Sandine, 1996), Lactoccocus lactis (Liu et al., 2020), and Companilactobacillus and Limosilactobacillus species (Comasio et al., 2019) have been investigated.Due to the relevance of these bacteria in the dairy industry, many of the characterizations are based on conversion of citrate or lactose into acetoin, as citrate is considered the substrate for acetoin production in milk, and many are based on LAB isolated from dairy sources.In this study, we seek to investigate acetoin production on a liquid fraction of BSG (LBSG) by utilizing LAB strains isolated from plant sources rather than dairy.We hypothesize that plant-isolated LAB can perform better in plant-based substrates, such as LBSG, due to a more diverse selection of genes for utilization of proteins, nutrients and carbohydrates present in plants.This hypothesis is backed by Gumustop and Ortakci (2023) who analyzed the genetic diversity of Leuconostoc pseudomesenteroides, and found a larger diversity in i.e. carbohydrate utilization genes in plantassociated strains.

Initial strain selection
The initial strain selection was based on the previous work done by Madsen et al. (2021).In this previous work, 171 LAB strains were screened on BSG agar plates and classified according to their ability to acidify the plates, monitored by color change after the addition of bromocresol purple.The idea behind the screening was to ensure that the selected strains were able to grow in BSG.Initially, 13 strains that were able to acidify LBSG agar were picked for further analysis of acetoin production in LBSG in the following study.This included six Leuconostoc mesenteroides, one Leuconostoc pseudomesenteroides, one Leuconostoc citreum, one Leuconostoc lactis, two Lactiplantibacillus plantarum, one Latilactobacillus curvatus and one Lactococcus lactis species.Later, it was decided to further investigate the acetoin production levels of Leuconostoc pseudomesenteroides species, and six more strains were included in the analysis.The well-studied laboratory strain Lactococcus lactis subsp.cremoris MG1363 was also included in the initial study as a reference, as it is a known acetoin producer under specific conditions (Nordkvist et al., 2003).

Liquid BSG preparation
The LBSG fraction was extracted from BSG from the Carlsberg Group's beer production facility in Fredericia, Denmark.The extraction was performed as described by M. F. Bjerregaard et al. (2019) by utilizing a custom-built rotary drum press, and the liquid fraction was extracted through a 300 μm filter (Comasio et al., 2019).The liquid fraction was frozen at − 40 • C immediately following the filtration, and was thawed and autoclaved (121 • C, 20 min) prior to use.The pH was adjusted with 10 M sterile NaOH, to a value of 6.0 before fermentation and the pH was measured with a WINLAB Excellent Line pH-Meter (Windaus-Labortechnik GmbH & Co., Germany).

Isolation, identification, and preparation of bacteria
The LAB were isolated from various sites in Denmark on either M17 or MRS medium, and incubated for 48 h at 30 • C. Following isolation, the strains were clean-streaked 3 times, and identified by MALDI-TOF mass spectrometry using the Microflex LT/SH smart benchtop system (Bruker, US).All identified strains reached a quality score > 2, sufficient for identification on species level, as recommended by the manufacturer.Cultures were stored in glycerol stocks (20 %) at − 80 • C and were plated on De Man, Rogosa and Sharpe (MRS, company, country) agar plates before use.Precultures were prepared by inoculating a single colony in MRS medium and incubating overnight at 30 • C.

Carbohydrate utilization
Strains were grown overnight in MRS medium at 30 • C.They were then washed twice in 0.9 % NaCl, after which they were assayed for carbohydrate utilization using the API® 50 CHL assay kit (bioMérieux), following the manufactures instructions.Strains were incubated in the assay strips for 48 h at 30 • C, after which they were evaluated to have no (− ), some (+) or strong (++) acidification capabilities in each of the available carbohydrates, based on the color change of the media.

Fermentation on BSG
Precultures grown overnight in MRS medium at 30 • C were centrifuged (2400 × g, 5 min, 5 • C) and washed in 0.9 % NaCl.A volume of preculture was added to LBSG to yield a theoretical bacterial OD600 value of 2.0, which was added to the fermentation medium in a 1/20 ratio to yield an initial OD600 value of 0.1.Fermentation in LBSG was performed in 100 mL shake flasks with 10 mL medium for aerobic fermentations or 50 mL sealed falcon tubes with 50 mL medium for microaerobic fermentations.The fermentations were performed for 24 or 48 h in shaking incubators set at 30 • C and 200 RPM in biological duplicates.

Bioreactor cultivations
Bioreactor cultivations were performed in 1 L Biostat A bioreactors (Sartorius, France) with 500 mL of LBSG under aseptic conditions.The pH and dissolved oxygen were measured using calibrated pH and dissolved oxygen (DO) sensors (Mettler Toledo, Swtizerland), respectively.Preculture was added as previously described to an initial OD600 value of 0.1.The cultivations were performed under aerobic conditions by continuous addition of 0.5 vvm atmospheric air, and the DO levels were controlled by adjusting stirring speed to ensure that oxygen was not limited.The pH was controlled by automatic addition of 5 M NaOH.

High-performance liquid chromatography
High-Performance Liquid Chromatography (HPLC) was used to quantify sugar, organic acid, ethanol and acetoin levels of cultures.The UltiMate 3000 HPLC (ThermoFischer Scientific, US) system was used, with an Aminex HPX-87H (Bio-Rad, US) organic acids column.Sugars and organic acids were detected and quantified using the Shodex RI-101 (Showa Denko K.K., Japan) refractive index (RI) detector, while acetoin levels were detected either using the same RI detector or UV absorbance at 290 nm with the UltiMate 3000 Diode Array Detector (ThermoFischer Scientific, US) in cases where RI was not sufficient for separation.The mobile phase used was 5 mM sulfuric acid (degassed) with a flow rate of 0.5 mL/min, and the chromatograms were detected for 30 min per sample.The column temperature was kept constant at 60 • C. Samples were diluted either 1:5 or 1:2 with 5 mM sulfuric acid, mixed by inversion and vortexing, followed by centrifugation for 5 min at 2400 ×g.The supernatant was filtered using a 0.22 μm syringe filter (Th.Geyer GmbH & Co. KG, Germany) and were then analyzed.Quantification and identification of metabolites were performed by measurement of analytically pure standards in 5 mM H 2 SO 4 .

Initial strain selection
All strains selected for screening in this paperincluding the initial 13 strains (see Section 2.1) -was additionally tested for their capabilities in carbohydrate utilization, in order to assess their ability to acidify the sugars naturally found in the BSG investigated in this study.The complete list of strains including identifications, source of isolation and classification of metabolism is presented in Table 1.

Characterization of acetoin production of LAB on BSG
The 13 initially chosen strains were cultivated in LBSG under microaerobic conditions in tubes and under aerobic conditions in shake flasks.Their organic acid, sugar, and acetoin production levels were quantified using HPLC.The presence of either meso-2,3 butanediol or optically active 2,3 butanediol was also detected but was not quantified due to the absence of optically pure standards.Under microaerobic conditions, none of the strains produced quantifiable levels of acetoin (data not shown), and microaerobic cultivations were omitted in the subsequent studies.The dependence of aeration for acetoin production in LAB has previously been demonstrated in L. lactis MG1363, where it is

Table 1
Strain ID, species identification and origin of isolation for LAB investigated in this study.Plant* indicates that origin is a plant, but exact isolation source was not noted.All strains were isolated from sites in Denmark.Acidification on BSG plate assays were performed by Madsen et al. (2021)  b Results from previous study (Madsen et al., 2021).
A.P. Wätjen et al. believed to be essential for NAD+ regeneration by NADH oxidase (Neves et al., 2002), as the acetoin pathway competes directly with the lactate pathway that is responsible for NAD+ regeneration.
The results from the aerobic shake flask cultivations after 24 h of cultivation can be seen in Fig. 1 and the full result set for 24 and 48 h are in Supplementary S1.All of the tested strains were able to ferment the glucose present in BSG extract within 24 h, however only Leu. pseudomesenteroides NFICC96 appeared to be able to consume all of the maltose after 24 h.Notably, L. lactis MG1363, L. lactis NFICC215, and Leu.lactis NFICC83 were unable to consume the maltose in LBSG at all, even after 48 h of cultivation.All the tested strains were able to produce acetoin on LBSG after 24 h of cultivation, although at different levels.As a result of different sugar consumption levels, acetoin titers should not be correlated to yield; however, comparing the titers after 24 h allows direct comparison of the performance on LBSG and was therefore used as the performance criteria.The highest producer Leu.pseudomesenteroides NFICC96 produced nearly 3 g/L of acetoin within 24 h of cultivation, while both L. lactis MG1363 and NFICC215 produced 1.6 g/L of acetoin even though they did not consume any of the maltose.The remaining Leuconostoc strains produced between 0.1 and 0.6 g/L of acetoin while Leu.lactis NFICC83 producing 0.6 g/L, Leu.citreum NFICC95 0.5 g/L, and Leu.mesenteroides strains varying between 0.1 and 0.4 g/L.Lactobacillus strains exhibited the lowest acetoin titers ranging from 0.1 to 0.2 g/L.Generally, strains that were low acetoin producers showed a low final pH value of around 4.1-4.2,while higher acetoin producers appeared to be above 4.4.
By-product formation in the form of acetate and lactate was also measured, and interestingly the LBSG already contained relatively high amounts of lactate and small amounts of acetate.This was also observed by Madsen et al. (2021) and was hypothesized to be due to contamination pre-processing of the BSG.Facultative heterofermentative LAB such as Lp.plantarum and L. lactis primarily produced lactate, with only small amounts of acetate, while obligate heterofermentative LAB such as strains belonging to Leuconostoc and Ll.curvatus generally produced less lactate and more acetate as expected.However, Leu. pseudomesenteroides NFICC96 did not appear to produce lactateon the contrary, the lactate concentration was lower than that of the unfermented LBSG, indicating that this strain likely consumes lactate.As Leu. pseudomesenteroides NFICC96 appeared to be the fastest consumer of LBSG, while producing the highest amounts of acetoin, it was decided to characterize the Leu.pseudomesenteroides species further.

Characterization of Leuconostoc pseudomesenteroides on liquid BSG extract
As only one Leu.pseudomesenteroides strain had been included initially, it was decided to investigate whether the production was strain-specific or if Leu. pseudomesenteroides generally were able to produce high levels of acetoin on LBSG.Four other Leu.pseudomesenteroides strains from our LAB strain collection were cultivated in LBSG as previously described for 24 h.The metabolite profile following the cultivation can be seen in Fig. 2. All of the tested Leu.pseudomesenteroides strains were able to produce higher levels of acetoin than we  A.P. Wätjen et al. previously observed with other Leuconostoc species, with NFICC99 exhibiting the highest titers while also being the only net lactate consumer, which was similar to the profile of NFICC96.However, as NFICC99 also has the potential to utilize arabinosecontrary to NFICC96 -(Table 1), it was chosen to investigate this strain further.
To characterize the acetoin production mechanism of NFICC99, we monitored its metabolite production and pH development during 24 h of cultivation (Fig. 3).It was seen that glucose was consumed and the maltose concentration increased slightly initially, likely due to the hydrolysis of maltodextrins present in LBSG.Lactate concentrations gradually increased at the early stage of fermentation, but decreased quickly following an increase in acetoin production rate, indicating a shift of the pyruvate flux towards the acetoin pathway rather than lactate.This is followed by a decrease in the lactate concentration and a further increase in acetate and acetoin levels.Acetoin formation and lactate consumption appeared to be correlated with the external pH value, and it was therefore decided to investigate the pH dependence of acetoin production, which can be seen on Fig. 4. In this case, the 48-h time point was chosen to balance differences in metabolite concentrations that would appear due to variations in growth rate at different pH levels.
A clear pH dependence of acetoin production was observed, where at high pH values, less acetoin and more lactate were produced, and vice versa at lower pH levels.However, at pH 5 and 4.6, less acetoin was produced than at pH 5.4, but less acetate was also made, indicating that a lower amount of sugar has been consumed, as 1 mol of acetate (+ethanol, which was not detected) will be generated from each mol of monomeric sugar consumed by obligate heterofermentative LAB, likely explaining the lower acetoin titer.As maltose and glucose concentrations were depleted entirely at all pH values except pH 4.6, the difference in the sugar consumption must come from the degradation of larger maltodextrins present in LBSG.Surprisingly, when pH was uncontrolled, the highest titers of acetoin were reached, concomitantly with the lowest concentration of total by-products.
To characterize the ability of Leu.pseudomesenteroides NFICC99 to produce acetoin on different carbon sources, we tried to cultivate it in MRS medium with different carbon sources (Results are available in S2 supplementary).However, none of the different carbon sources yielded acetoin titers near those observed in LBSG, even though higher total sugar concentrations were used.Instead, a high ethanol and lactate production was observed on glucose and maltose, indicating that the cells needed to get rid of excess NADH generated.Furthermore, higher acetoin titers were achieved in carbon sources such as arabinose and xylose that are usually considered less efficient carbon sources, where we also observed that growth was worse indicated by a lower OD600 value, which could indicate that the fast growth achieved in a heavily supplemented media such as MRS did not allow fast enough regeneration of NAD+ through the NADH oxidase.As acetoin production was previously shown to be pH dependent, and MRS is assumed to have a larger buffer capacity than BSG, we also tried different initial pH values down to pH 4.5, but this did not affect the acetoin titers by these strains significantly (Data not shown).

Discussion
In this study, we investigated the ability of different LAB to utilize the sugars in LBSG as a substrate towards acetoin production.Plant-isolated LAB were picked for investigation due to their higher potential for utilization of the plant-derived carbon and protein sources present in BSG.All of the tested LAB were able to produce acetoin and partly utilize the fermentable sugar substrates in BSG under aerobic conditions, there were significant differences in their abilities to do so.
L. lactis is currently one of the best-investigated LAB for acetoin production, and both engineered and natural high acetoin producing L. lactis strains have been reported in the literature (Liu et al., 2020).While both our plant-isolated L. lactis NFICC215 and the L. lactis laboratory strain MG1363 were able to achieve relatively high acetoin levels, they could not utilize the maltose fraction of BSG, rendering them less suitable for conversion in LBSG.These results were consistent with what was observed by Madsen et al. (2021) where it was seen that plant derived L. lactis strains were unable to degrade the maltose in BSG.Their inability to consume the maltose fraction is surprising, as MG1363 is known to be able to consume maltose (Solem et al., 2007).
For aerobic fermentation of BSG into acetoin, it was seen that Leu.pseudomesenteroides was the most suited organism.With Leu.pseudomesenteroides it was not only able to achieve the highest acetoin titers, but also the fastest consumption of the fermentable sugars in the LBSG.It was further demonstrated that the production was scalable to a 1 L bioreactor, presenting a quick and efficient method to enrich BSG with around 3 g/L (3000 ppm) of acetoin.Given the average maximum use of acetoin in foods of 750 ppm, this allows an easy method to enrich the nutritional value of certain foods with BSG without the need for adding acetoin.Bioreactor cultivations without pH control yielded the highest acetoin titers, concomitantly with the lowest amount of byproduct generation and should therefore be considered for utilization of BSG into acetoin.
Interestingly Leu.pseudomesenteroides strains have previously been shown to be unfavorable for acetoin and diacetyl production during cheese ripening, where it was shown to produce less acetoin and diacetyl than dairy Leu.mesenteroides strains (Pedersen et al., 2016).In a later pan-genomic Leuconostoc study this was hypothesized to be due to the absence of 2,3 butanediol dehydrogenase in some Leu.pseudomesenteroides strains, which is responsible for the conversion of diacetyl to acetoin and acetoin to butanediol (Frantzen et al., 2017).While it was not explained why they thought so, it could be because 2,3 butanediol formation requires a reducing cofactor (NADH/NADPH), which would be necessary for NAD+ regeneration in an environment where air transfer is limited.However, as the substrates and environment are assumed to be much different in BSG than in cheese, this is not surprising.
The importance of the composition of the medium was also demonstrated in our study, where Leu.pseudomesenteroides produced much less acetoin on MRS medium, even though some of the carbon sources were the same as in the LBSG.Using a genome-scale metabolic model of Leu.mesenteroides, Özcan et al. (2019) showed that acetoin and diacetyl formation was only favorable on glucose when a high glucose uptake was combined with a low growth rate, which might explain why no acetoin production was observed in rich MRS medium, as MRS is heavily supplemented with amino acids, vitamins, lipids, and glucose, allowing good growth.This hypothesis could be tried by monitoring the cell forming units (CFU) count of fermentations in different media.This also emphasizes the importance of screening for acetoin formation directly in the desired medium, rather than standardized media such as MRS.
The ability of Leu.pseudomesenteroides to efficiently consume the fermentable sugars could also be related to its ability to produce acetoin.Production of acetoin is involved in pH homeostasis in L. lactis, and the inability to utilize this pathway has been associated with the loss of ability to increase the internal pH of the cells under acidic stress, thereby reducing the viability of the cells (Zuljan et al., 2014).As LBSG has a low buffering capacity (pH 4.5 was reached after consumption of 4 g/L glucose), efficient pH homeostasis is expected to be essential for viability in LBSG.This was also observed when the metabolites of Leu.pseudomesenteroides were observed over time, that significant acetoin production started when pH was low and a slight increase in pH was observed following acetoin production (and lactate consumption).Cavin et al. hypothesized that degenerated ilv and leu operons observed in Leu.mesenteroides can be negative for growth (Cavin et al., 1999) and this was later proposed as an explanation to why Leu.pseudomesenteroides that does not have these genes at all exhibited faster growth (Frantzen et al., 2017), indicating that absence of genes can also be important for more efficient phenotypes.
An unexpected result was the ability of Leu.pseudomesenteroides to utilize lactate, which has previously not been described in Leuconostoc species.Lactate utilization has been observed in L. lactis under respiratory conditions (Zhao et al., 2013) in Levilactobacillus brevis under the presence of O 2 (Guo et al., 2017) but also under anaerobic conditions by Lentilactobacillus buchneri (Johanningsmeier et al., 2012) and is therefore not a unique ability of Leu.pseudomesenteroides.Lactate utilization could potentially influence acetoin production, as it previously has been shown that lactate is converted to pyruvate, which is also the precursor for acetoin biosynthesis.However, upregulation of pyruvate dehydrogenase and pyruvate oxidase were observed during aerobic lactate utilization in Lb. brevis, which switches the flux of pyruvate to acetate.This mechanism allows the generation of extra ATP and increases the pH of the medium in Lb. brevis (Guo et al., 2017), which might also be the case in Leu.pseudomesenteroides.We observed that strains that showed lactate concentrations below the lactate levels of LBSG also appeared to produce more acetoin and acetate, however as lactate and acetoin are made from the same precursor, this could also be a result of the fact that less lactate is produced when acetoin is made.
The ability of Leu.pseudomesenteroides to reduce lactate levels can potentially be utilized to alter the organoleptic properties of the LBSG for some applications.Roth et al. (2016) proposed that the high lactic acid content of BSG might negatively affect the flavor perception in bread, and Hansen and Schieberle (2005) showed that sourdough fermented with heterofermentative Fructilactobacillus sanfranciscensis had a pleasant odor, while those fermented with homofermentative Lp. plantarum had an unpleasant, acidic odor.This was later hypothesized to be due to formation of more lactic acid by homofermentative LAB, as lactic acid has a lower pKa value than acetic acid (Pico et al., 2015).On the other hand, it might be unsuited for some dairy products such as yogurts where lactic acid is considered an important flavor, and acetate might not be desired.

Conclusion
In this study, we investigated the ability of different plant-isolated LAB to produce acetoin in the liquid fractions of brewers' spent grain.Strains of Leuconostoc pseudomesenteroides appeared to be the most suited LAB for higher titer acetoin production, reaching around 3 g/L after 24 h of fermentation with acetate as the only major byproduct.The production was scalable to 1 L aerobically operated bioreactors, and thereby represents a sustainable and safe method to convert liquid BSG into a flavorful food additive for use in human food purposes.Furthermore, we quantified the production of acetoin and byproducts for L. lactis, Leu. mesenteroides, Leu. citreum, Leu. lactis, Lp. plantarum and Ll. curvatus, and hereby demonstrated different fermentation strategies to produce acetoin, depending on the final concentrations of acetoin and byproducts desired.

Fig. 1 .
Fig. 1.Characterization of acidification and metabolites after 24 h fermentation of LAB on liquid BSG.Bars indicate metabolite concentrations, and error bars are based on the standard deviation of biological duplicates.pH indicates the final pH value obtained after 24 h of fermentation.

Fig. 2 .
Fig. 2. Characterization of Leu.pseudomesenteroides metabolites after fermentation on BSG.Bars indicate metabolite concentrations, and error bars are based on the standard deviation of biological duplicates.BSG represents the organic concentrations in unfermented liquid BSG.Results are indicated for both after 24 h, and 48 h of fermentation.

Fig. 3 .
Fig. 3. Metabolite production of Leu.pseudomesenteroides NFICC99 over 24 h BSG fermentation.Left axis indicates metabolite concentrations, while right axis shows pH only.Error bars are based on the standard deviation of biological duplicates.

Fig. 4 .
Fig. 4. The pH dependence of NFICC99 for metabolite production on LBSG after 48 h.Bars indicate metabolite concentrations obtained when pH was controlled at different values.Error bars are based on the standard deviation of technical duplicates, and not biological duplicates.
and was established relatively by visual comparison as either high or low.N.d.,Not determined.Carbohydrate acidification was performed via API 50 CHL assay.Strains were evaluated to have (− ) no, (+) some or (++) strong acidification capabilities when fed the corresponding carbohydrate.
a Strain within the initial 13 screened.