Six 2-L sequential batch reactors (SBR) were simultaneously operated to evaluate the adaptation of an MMC to synthesise AA in co-production with LA under a feast and famine regime. The first condition (SBR1), assayed in triplicate, was fed with synthetic hemicellulose hydrolysate, where the composition in percentage (% g compound/g VS) was a residual stream of lignocellulosic process fractionation (Vallejos et al., 2017): 78% xylose, 9.0% acetic acid, 5.06% furfural, 4.7% arabinose, 2.3% glucose, 0.8% cellobiose, and 0.14% hydroxymethylfurfural. A second condition (SBR2), assayed in triplicate, was operated using acetic acid as the only carbon source. Both substrates were selected as models for comparing the use of non-fermented (SHH) versus acetate substrate as the carbon source to produce high added-value compounds (AA and LA) instead of PHB when the MMC are subjected to a feast and famine regimen. SBR1 and SBR2 were supplemented with the nutrients and minerals necessary for bacterial growth, including: 132 mg of peptone/L, 68 mg of beef extract/L, 112 mg of (NH₄)₂SO₄/L, 49 mg of KH₂PO₄, 66 MgSO₄/L, 170 mg of NH₄Cl/L, 92 mg of K₂HPO₄/L, 45 mg KH₂PO₄/L, 600 mg MgSO₄/L, 100 mg EDTA/L, 70 mg CaCl₂/L, and 2 mL of a trace element solution (Huang et al., 2017). 10 mg Thiourea/L was also added to prevent nitrification. The trace elements solution was composed of 1,000 mg FeSO₄·6H₂O/L, 150 mg H₃BO₃/L, 150 mg CoCl₂·6H₂O/L, 120 mg MnCl₂·4H₂O/L, 120 mg ZnSO₄·7H₂O/L, 60 mg Na₂MoO₄·2H₂O/L, 30 mg KI/L, and 30 mg CuSO₄·5H₂O/L.

SBR1 and SBR2 were inoculated with activated sludge from the municipal wastewater treatment plant in Temuco, Chile. Both reactors were operated under the same conditions: 4 L/min airflow, 30°C, 60 rpm, and a C/N/P ratio of 30/0.9/0.1 (in mmol/L). The strategy applied was the alternation of F/F periods through SBR operation to obtain MMC enriched with microorganisms capable of producing high added-value compounds. Two cycles per day were carried out in each SBR until a stable operation was reached. One cycle consisted of: a feed period (6 min) followed by an aerobic reaction (704 min) and 10 min of withdrawal (Huang et al., 2017; Wang et al., 2017). No settling phase was performed, and all excess biomass was withdrawn with the mixed liquor (Dionisi et al., 2005; Cabrera et al., 2019). Thus, the biomass retention time was equal to the hydraulic retention time. 0.25 L of substrate was fed into each reactor in each cycle using organic loading rate (OLR) of 60 Cmmol/L*day. At the end of each cycle, 0.25 L of the liquid was withdrawn by digital peristaltic pumps using a Compact DAQ system (cDAQ-9,178 chassis, National Instruments, Austin, TX, United States), with a routine programmed using the LabView software (National Instruments). SBR1 and SBR2 were operated for 100 consecutive cycles (50 days) since stable operation was reached from cycle 80 in a previous study (Pinto-Ibieta et al., 2020). To monitor the evolution of AA, LA, PHB, and biomass growth in both SBR, samples were taken at the end of a cycle once a week. Once a stable operation was determined, intermediate samples were taken during feeding cycles 80, 84, 88, and 90 from both reactors. The samples were taken in triplicate every hour to monitor the evolution of AA, AL, PHB, biomass, and carbon source uptake (pentoses and acetate) as a function of the time in the cycle. The mean and standard deviation of these cycles were calculated to corroborate the stability of the operation.

The MMC obtained from the operation of SBR1 and SBR2 were cultivated separately in batch reactors to evaluate their adaptability to the use of SHH and maximise their ability to produce LA, AA and/or PHB in a batch stage, regardless of the substrate used during the selection phase. SHH (described in point 2.1) was the only carbon source used in these batch assays. The maximum accumulation capacity of compounds was assayed without the addition of nitrogen, since it has been reported that restricting this nutrient in the culture could lead to the accumulation of storage compounds by the cell (Guerfali et al., 2018; Chong et al., 2019). Six assays, in triplicate, were carried out to evaluate the effect of increasing SHH doses (30, 75, and 120 Cmmol/L) on the production of the compounds of interest (Table 1). 80 mL of adapted MMC (with a concentration of 3 mg VS/L), obtained from each SBR reactor described in Section 2.1 were used as inoculum in Erlenmeyer flasks containing 120 mL of mineral medium without a nitrogen source (see Section 2.1), incubated for 24 h at 150 rpm and 30°C. ANOVA tests were performed to evaluate the significance of the differences observed for AA and LA accumulations at the different experimental conditions.

LA, AA, and PHB were quantified and identified according to Lappalainen and Dong (2019) and Serafim et al. (2004). A GC/MS Clarus 600 (Perkin Elmer) with an ELITE 1701 column 30 m × 0.25 mm × 0.25 um was used for identification, and a GC/FID Clarus 600 (Perkin Elmer) with CAPILAR NUKOLTM SUPELCO 30 m × 0.25 mm × 0.25 um for quantification. The calibration curve was obtained by injecting a series of standards at different concentrations of AA, LA, and PHB (Sigma Aldrich). The substrate consumption was measured by determining the concentration of reducing sugars and acetate in filtered samples (0.22 μm pore size PVDF membrane, Merck). To quantify the total reducing sugars, the dinitrosalicylic acid (DNS) reagent method was used (Zhang et al., 2012; Prasertsung et al., 2017). Acetate was determined by gas chromatography in a flame ionisation detector (Clarus 400, PerkinElmer), using a NukolTM capillary column (Sigma-Aldrich, Darmstadt, Germany). VS and soluble chemical oxygen demand (SCOD) were quantified using a standard technique (AWWA, 2005).

Figure 3A shows the concentrations of LA, AA, and PHB reached for the different SHH concentrations evaluated in the batch experiments conducted with the MMC previously selected in SBR1. For the three SHH concentrations evaluated, it was possible to accumulate the three added-value compounds identified in SBR1: LA, AA, and PHB. The highest concentration of SHH (120 Cmmol/L) produced the highest accumulation of each compound, obtaining yields of 23 ± 1, 16 ± 2, and 6 ± 2% g compound/g VS of LA, AA, and PHB, respectively (Figure 3A). When using 75 Cmmol/L, the LA, AA, and PHB accumulation concentrations were 20 ± 2, 14 ± 1, and 4 ± 1% (g compound/g VS), respectively; no significative difference from the latter was found in the values obtained at 120 Cmmol/L (Supplementary Table S1). The higher concentrations observed at 120 Cmmol/L could indicate that SHH was not inhibitory at this concentration (de Oliveira et al., 2019). The lowest accumulations were obtained at 30 Cmmol/L, with yields of only 15 ± 3, 9 ± 2, and 5 ± 2% (g compound/g VS) for LA, AA and PHB, respectively (Figure 3A), values which showed a significative reduction compared to those obtained at 120 Cmmol/L (Supplementary Table S1). This decrease could be caused by a limitation in the substrate, which could leading to a microbial growth limitation (Alburqueque et al., 2010; Pinto-Ibieta et al., 2021).

The use of MMC from SBR2 instead of SBR1 in the batch experiments resulted in general in a higher concentration of PHB and lower concentration of AA and LA compared to the results obtained using SBR1 (Figure 3). Specifically, by feeding with SHH at 120 Cmmol/L, it was possible to accumulate LA, AA, and PHB at concentrations of 18 ± 2, 7 ± 1, and 15 ± 2% (g compound/g VS), respectively (Figure 3B). On the other hand, at 30 Cmmol/L, the production of both LA and PHB was 5% (g compound/g VS), while no production of AA was observed. As in the reactors inoculated with the MMC obtained from SBR1, the statistical analysis (ANOVA) for the reactors inoculated with the MMC obtained from SBR2 only showed a significant difference when comparing 120 versus 30 Cmmol/L (Supplementary Table S2).

These results indicate that the simultaneous production of LA, AA, and PHB was possible using both adaptations of MMC under batch conditions (from SBR1 and SBR2), although the long-term adaptation of the MMC to the SHH in SBR1 favoured the accumulation of AA and LA. It may be noted that the origin and adaptation of the MMC could be a key point for the accumulation of high added-value compounds, as it is indicated by the statistical differences observed in the accumulation extents obtained from the MMC from SBR1 and SBR2 (Supplementary Table S3).

Figure 4 shows the evolution of pentoses, biomass, LA, AA, and PHB during the culture time for the assays at 120 Cmmol/L using MMC from SBR1 (Figure 4B) and from SBR2 (Figure 4B) as the inoculum. Both reactors resulted in similar production and composition of the accumulated compounds despite of the different MMC used (Figure 4). Pentoses and acetic acid were rapidly consumed while AA, LA, and PHB were produced at the same time. The maximum accumulation of AA and LA was reached at the time when the pentoses were exhausted (around 3 h). Likewise, the maximum PHB was reached once the acetic acid was exhausted in the reactors (Figure 4). Similar behaviours were observed for AA, LA, and PHB, with concentrations decreasing after maximum generation. These results could indicate AA and LA storage in some kind of carbon source similar to PHB. Therefore, MMC adapted from SBR2 were not able to synthesise LA and AA using acetate, but the same microorganisms fed with SHH were capable of synthesising LA and AA, confirming that the production of AA and LA depends on the use of SHH as carbon source. This was expected since the metabolic pathways described for both AA and LA production begin with hexoses and/or pentoses as substrate (Lee et al., 2019).

Although both MMC types were able to synthesise LA, AA, and PHB when cultivated in batch conditions, the culture yields were different. As expected, MMC adapted in SBR1 with SHH were able to produce higher yields of AA and LA than those from SBR2. On the other hand, the co-production of LA and PHB was favoured in the batch assays conducted with MMC selected in SBR2. Therefore, the kind of substrate used in the selection stage could stimulate the production of specific compounds in a subsequent batch stage. The possible benefits of this co-production should be evaluated considering the yields of each compound, their final properties, the difficulties, and costs of separation, etc. In addition, it is important to identify the microbial communities and their metabolic pathway involved in AA and LA synthesis.