All the chemicals were purchased from Sigma-Aldrich (Steinheim, Germany) unless otherwise stated.

The yeast strain P. fermentans WC 1507, belonging the collection of the Laboratory of Microbial Biotechnologies (Department of Life Sciences, University of Modena and Reggio Emilia) and already known for its xylose-to-xylitol conversion ability (Raimondi et al., 2023), was routinely cultured in YPD (20 g/L glucose, 10 g/L tryptone, 10 g/L yeast extract) broth at 30°C in aerobic conditions. The inoculum for free cells fermentation experiments and the biomass for immobilized-cells processes were prepared in MY medium (20 g/L xylose, 3 g/L yeast extract, 2 g/L (NH₄)₂SO₄, 3 g/L KH₂PO₄, 1 g/L K₂HPO₄, and 1 g/L MgSO₄ × 7H₂O), where the yeast was incubated aerobically at 30°C. Unless otherwise specified, xylitol production was conducted in a xylose-enriched MY medium containing 120 g/L of xylose. Any modifications required for specific experiments were explicitly mentioned.

Aspergillus oryzae FST 76–2, which is part of the collection of the Department of Agricultural Chemistry, Edaphology, and Microbiology at the University of Cordoba (Spain), was used to produce mycelium pellets for yeast immobilization. The strain was routinely cultured at 28°C on agar plates containing SM medium (17 g/L corn meal agar, 1 g/L yeast extract, 2 g/L glucose, 20 g/L agar). Mycelial pellets were generated using fungal pellet medium (FPM) consisting of 60 g/L glucose, 3 g/L yeast extract, 3 g/L NaNO₃, 1 g/L K₂HPO₄, 10.2 g/L MgSO₄ ∙ 7 H₂O, 0.5 g/L KCl, 0.02 g/L FeSO₄ ∙ 7 H₂O, pH = 5.5.

Bioreactor batch fermentations with suspended cells of P. fermentans WC 1507 were carried out to determine the effect of medium composition (i.e., xylose and ammonium concentration) and process parameters (i.e., the pH and oxygenation) on xylose-to-xylitol transformation. Fermentation runs were performed in 500 mL stirrer tank bioreactors (Mini Bio, Applikon Biotechnology, Delft, the Netherlands), filled with 350 mL of xylose enriched MY medium and inoculated (5%, v/v) with an overnight grown seed culture. Cultures were maintained at 32.5°C and aerated with 0.5 v/v/min filter-sterilized air. Foaming was prevented by the addition of 150 μL of a mixture (1:1, v/v) of Xiameter 1,520 (Dow Corning, Midland, MI, United States of America) and polypropylene-glycol.

In the MY medium, xylose was utilized at the concentration of 90, 120, and 150 g/L, with or without (NH₄)₂SO₄. To investigate the effect of the oxygenation on the xylose-to-xylitol conversion, two different conditions were compared: i) dissolved oxygen tension (DOT) higher than 20%, obtained by cascade-controlled stirring between 1,100 and 1700 rpm; ii) free DOT, obtained applying constant stirring at 1,100 rpm. To study the effect of pH on the bioconversion performance, four different pH conditions were tested: pH higher than 2.5, 3.5, 4.5, and 5.5, with pH decrease prevented by the automatic addition of 1 M NaOH. The stirring was constant at 1,100 rpm and the process was monitored for 144 h. All the processes were periodically sampled to determine growth, xylose consumption, and xylitol generation.

The biomass of P. fermentans WC 1507, cultured for 24 h in MY medium with 20 g/L xylose, was entrapped in alginate beads (AB) and in mycelial carrier (MC), and alginate-coated AB and MC (hereinafter referred to as cAB and cMC).

To produce the AB, the culture of P. fermentans WC 1507 was mixed (1:1 v/v) with 40 g/L sodium alginate. The suspension was pumped with a peristaltic pump and dropped through a syringe needle into a gently stirred 20 g/L CaCl₂ solution. The AB were kept in the calcium solution at room temperature for 30 min, then they were washed twice with water and stored in physiological solution (9 g/L NaCl) at 4°C. According to the initial concentration, alginate 2% (w/v) AB were obtained.

The support for immobilization in MC consisted of inactivated mycelial pellets of A. oryzae FST 76–2. To produce the pellets, the fungal spores from an 8-days grown SM plate were suspended in 1 mL of FPM, vortexed, sonicated in a water bath for 10 min, and inoculated at the concentration of 6 ×10³ spores/mL in a 1 L flasks containing 250 mL of FPM. After 3 days of aerobic incubation at 28°C, the mycelial granules or pellets were washed with water, and sterilized in autoclave for inactivation. The mycelial pellets, the cells of P. fermentans WC 1507 previously collected by centrifugation, and sterile water were mixed with the ratio of 1:1:5 (w/w/v) and vortexed. To force the yeast cells towards the core of the fungal pellets, the suspension was subjected to vacuum for 10 s, then the MC were washed with water and stored in physiological solution at 4°C (Lúquez-Caravaca et al., 2023).

To produce cAB and cMC, 30 g of AB or MC were submerged in 2 g/L sodium alginate for 5 min, drained in a sterile sieve, and rinsed with 200 mL of sterile water. Hydrogel film coating was obtained transferring the beads into a gently stirred 20 g/L CaCl₂ solution. After 30 min at room temperature, the cAB and cMC were washed with water and stored in physiological solution at 4°C. Photographs of the carrier beads were acquired and analyzed utilizing ImajeJ (https://imagej.net/), in order to measure their size and monitoring swelling during the bioconversion.

Xylitol production by immobilized cells of P. fermentans WC 1507 was evaluated in flasks experiments. 2.5 g of immobilized yeast cells were added to 100 mL flasks filled with 25 mL of MY medium containing 120 g/L xylose. The flasks were incubated at 200 rpm in an orbital shaker at 28°C for 116 h. Aliquots of medium were collected every 24 h to analyze xylose and to monitor pH and cells release.

To determine the reusability across consecutive production runs, the immobilized biomass was recovered and pooled at the end of the incubation, washed with sterile water, and added to fresh medium to begin anew. At the end of each run, some immobilized biomass was withdrawn to determine the number of entrapped cells (cells/g carrier), cells vitality (%), and beads swelling (mm). Five replicate flasks were set up for the first run and thus decreased to three replicates for the third one.

Bioreactor batch fermentations with immobilized cells of P. fermentans WC 1507 were carried out in a laboratory-scale airlift bioreactor. The apparatus was set up by modifying the Mini Bio apparatus, with the removal of the stirring shaft and the installation of a draft tube with a diameter of 30 mm. The process was performed in 350 mL of MY medium with 120 g/L of xylose and 35 g of AB loaded with approx. 2.0 ✕10⁸ yeast cells per g. Mixing was realized by providing the culture with 4 v/v/m of air. The pH was kept at values >3.5 by automatic titration with 1 m NaOH. Aliquots of medium were collected every 24 h to analyze xylose and to monitor cells releasing.

To quantify xylose and xylitol concentration, samples were clarified by centrifugation (9,500 xg for 5 min) and analyzed by HPLC (1,200 System, Agilent Technologies, Waldbronn, Germany) equipped with a refractive index detector. Isocratic elution was carried out at 60°C with 0.6 mL/min of 5 mM H₂SO₄ through an ion exclusion column (Aminex HPX-87 H, Bio-Rad, Hercules, CA, United States of America), according to Amaretti et al. (2018). The xylose to xylitol conversion yield (Y_P/S) was calculated on mass basis as the ratio between xylitol produced and xylose consumed (Amaretti et al., 2020). Volumetric productivity was calculated by dividing xylitol titer by time of fermentation (hours).

To quantify the free cells in the supernatant or those entrapped in the carriers, after dissolution or disruption, the yeast cells were counted in Thoma’s chamber and assayed for vitality. For alginate beads, 1 g of AB or cAB was placed in 9 mL of 10 g/L sodium-citrate solution and vortexed until complete dissolution. For disruption of mycelium carrier, 1 g of MC or cMC was broken with a tissue grinder in 9 mL of physiological solution or 10 g/L sodium-citrate solution, respectively. Vitality was evaluated by mixing (1:1, v/v) the cells suspension and a solution containing 0.1 g/L methylene blue and 20 g/L dihydrate sodium citrate. After 5 min incubation at room temperature, at least 200 cells were observed by microscope, the colorless and the blue ones being considered vital and non-vital, respectively (Kwolek-Mirek and Zadrag-Tecza, 2014).

All the reported values are means of at least three separate experiments. t-test and ANOVA followed by Tukey post hoc analysis were utilized for the comparison of means. Differences were considered statistically significant for p < 0.05.

When cultured batch-wise in a xylose-based medium within a stirred tank bioreactor, P. fermentans WC 1507 produced xylitol essentially during the growth phase, confirming a previous study assessing the ability of the strain in preliminary shake-flasks experiments (Raimondi et al., 2023) (Figure 1A,B). The application of cascade control of stirring, aimed at preventing the occurrence of oxygen transport limitations, resulted in an improved growth rate without affecting the final turbidity of the culture. On the other hand, improved oxygenation depressed both xylose consumption and xylitol production. Under limited oxygenation conditions, at the entrance into the stationary phase after 48 h, 80.5 g/L xylose were consumed and yielded 57.3 g/L xylitol, with a Y_P/S of 71%. When the DOT was kept >20%, significantly lower values (p < 0.05) were achieved, 45.9 g/L xylose being consumed to yield 14.3 g/L xylitol, with a Y_P/S of 31% (Figure 1C). This result confirmed previous evidence indicating microaerophilic conditions as the most appropriate for xylitol production by other yeast species. Yeasts can use xylose for cellular respiration in aerobic conditions. After entering the cell, a xylose reductase NAD(P)H-dependent reduces it to xylitol. While the excess of produced xylitol can be excreted out of the cells, the remain is oxidized to 5-xylulose through a xylitol dehydrogenase NADP⁺-dependent, then 5-xylulose can be metabolized through the pentose phosphate pathway and glycolysis (Moysés et al., 2016; Zha et al., 2021). When an excess of oxygen occurs, NADP⁺/NAPDH ratio increases and, therefore, intensify the xylitol-to-xylulose conversion, reducing the accumulation and excretion of xylitol.

Throughout the growth and production phase, the pH of the culture progressively decreased to 2.7. The effect of pH was thus investigated, by automatically titrating acidity and preventing the pH to drop below 3.4, 4.5, or 5.5. Compared with the free-pH culture, pH values >3.5 and 4.5 maximized xylose consumption, leading the substrate to get nearly exhausted during cultivation (Figure 2). Xylitol generation was the highest (p < 0.05) in cultures with the pH > 3.5, that reached 79.4 g/L xylitol, a value that was slightly but significantly higher than that obtained by cultures with the pH > 4.5 (73.4 g/L). The cultures with pH > 5.5 were the less efficient in producing xylitol, yielding the lowest titer (40.8 g/L). The cultures with uncontrolled pH presented the highest Y_P/S, followed in decreasing order by those with pH > 3.5, 4.5, and 5.5. These results suggested that acidic pH was necessary for xylitol production and that progressively higher values had the effect of decreasing the efficiency of the transformation. Yeasts have been observed to produce xylitol across a broad pH range, with high pH that negatively influences the activity of the xylose membrane transporter, while an acid environment can affect the cell maintenance requirement as well as the redox balance of the bioreduction (Converti and Dominiguez, 2001; Dasgupta et al., 2017). In the context of second-generation biorefinery development, xylitol production can be attained through the utilization of lignocellulosic hydrolysates as a feedstock (Vanmarcke et al., 2021). These substrates may be characterized by the presence of inhibitory molecules, such as acetic acid, and it is expected that the acidity of the environment could lead to increased acetate toxicity (Trček et al., 2015), potentially resulting in a shift in pH preference towards a less acidic range. To achieve an optimal xylitol titer, an overall balance among these factors is necessary and should be investigated, preferably also using real industrial feedstock.

In order to assess the medium composition, batch fermentations were carried out in the presence/absence of ammonium and yeast extract as nitrogen source and with different initial xylose concentrations (90, 120, and 150 g/L) (Figure 3). At all initial xylose concentrations, the highest amounts of both consumed xylose and generated xylitol were obtained with the organic nitrogen source (p < 0.05), that had a cumulative effect with the inorganic one. These results confirm that xylitol production by P. fermentans WC 1507 goes along with growth and that nitrogen restriction is detrimental for production. On the other hand, high xylose concentration is necessary for xylitol generation, that reached its maximum, in terms of substrate consumption, product generation, and Y_P/S when the culture was grown with 120 g/L. Growth under microaerophilic conditions may force the strain to reduce a great amount of xylose to xylitol, in order to oxidize part of the reduced cofactors generated with the catabolism. However, P. fermentans WC 1507 did not grow nor produce xylitol in anaerobiosis (data not shown), indicating that the strain relies on respiration to gain energy for its cellular processes.

Among the various conditions described in this study, the most successful process resulted in a xylitol titer of 79.4 g/L, with a Y_P/S of 66.3%, and a volumetric productivity of 1.3 g/L/h. Except for the processes exploiting C. tropicalis, which, to the best of our knowledge, have shown the best performance thus far (Oh and Kim, 1998; Kwon et al., 2006), the batch process described in this study achieved among the highest values of xylitol titer, Y_P/S, and volumetric productivity compared to other batch processes involving non-engineered yeast strains (Table 1). In particular, P. fermentans WC 1507 outperformed the wild-type P. fermentans used in a similar process by Prabhu et al. (2020) and exhibited performance similar to the selected mutant obtained through random chemical mutagenesis by the same authors. Therefore, the strain P. fermentans WC 1507 could represent a promising starting point for both strain and process improvement aiming at further enhancing the ability of this yeast.

Since xylitol production by P. fermentans WC 1507 presented limited oxygen demand, it was well suited to be carried out with immobilized cells, also offering the advantage of possible reuse of the biomass. Immobilized cells of the yeast, in the forms AB, MC, cAB, and cMC, were employed for three consecutive runs in flasks of MY medium containing 120 g/L xylose (Figure 4). The first run was started with 1.3–2.1 × 10⁷ entrapped cells per g, resulting in asuspension of 1.8–3.4 × 10⁷ immobilized yeasts per mL of culture. The swelling of the beads and the count of free and entrapped cells were evaluated at the beginning and at the end of each run (Figure 4A; Supplementary Figure S1).

At the end of each incubation run, the counts of cells entrapped within AB, MC, cAB, and cMC were significantly more numerous (p < 0.05) compared to the initial counts, determining a progressive increase of entrapped cells. For instance, the 3^rd run initiated with 5.4–11.7 × 10⁷ immobilized yeasts per mL of culture (Figure 4A). This effect was more evident for mycelium immobilized systems MC and cMC, where immobilized cells became significantly more numerous than in alginate beads, reaching up to 1.8 × 10⁸ cells/mL. At each run, the increase of entrapped cells was always accompanied by the release of yeasts in the medium. The cells directly liberated by the carriers or subsequently grown in the medium reached the concentration of 6.0–6.2 × 10⁸ cells/mL at the end of the third run. At this time point, the cells entrapped in AB, MC, cAB, and cMC represented 17.7%, 9.0%, 20.1%, and 22.5% of total cells. The alginate beads exhibited progressive swelling over the course of the three runs, starting with a mean diameter of 3.4 mm and reaching 5.1 mm. The coating of AB and MC exerted a slight but significant effect on preventing the release of yeasts only in the first two runs. This effect tended to attenuate, and no significant differences were observed at the end of the third run in both alginate and mycelium-based systems.

The kinetics of xylose consumption and xylitol generation by immobilized cells were linear (Supplementary Figure S2). In agreement with previous observations (Dominguez, 1998), with all the carriers and particularly with AB and cAB, the first run was the less efficient in terms of conversion rate, xylose consumption, and xylitol generation, while the second and the third ones proceeded at higher similar rates. At the end of the third run, the AB and cAB performed better (p < 0.05) than MC and cMC in terms of xylose consumption, xylitol generation and Y_P/S (Figure 4B). After 116 h of incubation, 102.5 g/L xylose was consumed by P. fermentans WC 1507 immobilized in AB and cAB, yielding 63.0 g/L xylitol with a Y_P/S of 61.5%.

While the Y_P/S values observed with both immobilized and free cells were similar, the immobilized cells exhibited lower final xylitol titer and volumetric productivity. Presumably, mass transfer limitations impaired the diffusion of the substrate in the beads. Literature reports cases where xylitol production was more efficient with the cells immobilized within calcium alginate compared with the free ones (Dominguez, 1998; Prakash et al., 2011; Soleimani and Tabil, 2013). An accurate comparison between free and immobilized cells and between our study and previous ones is challenging due to variations in experimental setups. Therefore, operational parameters such as the loading of cells, the size of beads, and the number of reutilization cycles should be specific targets for process improvement. The addition of an outer layer coating the carriers, specifically in cAB and cMC, neither improved nor reduced the bioconversion performances. The film likely did not introduce additional limitations to substrate mass transfer, opening up the perspective to exploit this strategy to enhance the tolerance of immobilized cells against inhibitory substances (Willaert, 2011; López-Menchero et al., 2021).

AB immobilization was selected for utilization in laboratory-scale bioreactor experiments. The stirred tank bioreactor resulted inadequate to preserve the integrity of the beads (data not shown), therefore the bioreactor was modified into an airlift configuration. The best process parameters (i.e., the pH > 3.5 and the absence of oxygen control) and medium composition (120 g/L xylose, 3 g/L yeast extract, and 2 g/L ammonium sulfate) previously identified in batch cultures were utilized. Three consecutive runs were carried out with the same AB beads. Between each run, the exhausted medium was replaced with new fresh medium. At the start of the first run, 3.4 × 10⁷ immobilized cells/mL were loaded. Throughout the experimental runs, a gradual rise in entrapped cells was observed, concomitant with an increase in the number of free cells (Figure 5A). This trend resulted in entrapped cells comprising 12%–17% of the total cells in the culture by the end of each run. These distribution among entrapped and free cells were similar to those observed in shake flasks. The alginate beads displayed a gradual increase in size during all three runs, exhibiting a progressive swelling phenomenon. The average AB diameter after these runs reached 5.2 mm, similar (p > 0.05) to the size achieved in the shake flasks experiment.

Likewise flasks experiments, the three consecutive runs yielded a progressive increase in xylitol production, that passed from a final titer of 38.7 g/L, generated in the first run with a Y_P/S of 37.5% and volumetric productivity of 0.33 g/L/h, to a titer of 63.0 g/L xylitol, generated in the third run with a Y_P/S of 61.5% and a volumetric productivity of 0.52 g/L/h (Figure 5B). Nonetheless, unlike in flasks experiments, the conversion in the laboratory scale airlift reactor did not proceed linearly, but presented a sigmoidal trend, with xylitol production progressively decelerating and approaching a plateau between 96 and 116 h of incubation, even though around 30–38 g/L of xylose were still available (Figure 5C). The bioreactor likely performed less effectively than flasks because of oxygenation, since the DOT reached values below 10% only at 32 h and recommenced to increase toward saturation at approx. 60 h. When a lower air flow was applied to achieve a DOT close to 0%, as required for xylitol production, it proved inadequate to maintain the immobilized biomass in suspension and circulating within the airlift system. Therefore, further experiments are necessary to find the optimal aeration condition and the airlift design, in order to guarantee an adequate culture mixing and, at the same time, the suboptimal oxygenation levels necessary for xylitol production.