Co-substrate model development and validation on pure sugars and corncob hemicellulosic hydrolysate for xylitol production

Effects of substrate concentration on biomass and xylitol production

Ten batch bioconversions using C. tropicalis cultivated in co-substrate of xylose and glucose with different concentrations of xylose (10 – 100 g/L) under a ratio of 10:1 were assessed. The experimental data of fermentative parameters are tabulated in Table 1. Both the mitigation of Yx/s and Qx as well as the increment of Yp/s1 and Qp incorporated with increased Yp/x could signify that the desired metabolic pathway of converting xylose to xylitol instead of biomass was achieved under the condition of relatively high substrate concentrations. As a result of it, although the increment of the initial substrate concentrations caused a longer cultivation time, the lowest biomass yield of 0.09 g/g incorporated with the highest xylitol concentration, yield, and productivity of 64.5 g/L, 0.78 g/g, and 0.90 g/L/h, respectively, were achieved by using xylose concentration of 90 g/L with additional 9.0 g/L glucose (Table 1). It can be concluded that the higher xylose concentration benefits the decrease of the fraction of xylose metabolized to generate biomass and increase the conversion efficiencies of xylose to xylitol. This phenomenon can be explained by a trade-off between xylitol production and biomass formation. As shown in Fig. 1, xylitol is directly converted from xylose, whereas the biomass formation from xylose needs through XDH followed by xylulokinase (XK) enzymatic reactions. XDH converts xylitol back to xylulose and XK phosphorylates xylulose to xylulose-5-phosphate (Xylulose-5-P), a pivotal step in xylose metabolism. Xylulose-5-P is an intermediate that can feed directly into the PPP by ribose-5-phosphate isomerase (RPI). The higher xylitol productivity was the result of a higher XR:XDH ratio28.

Table 1 Maximum xylitol concentration and its corresponding time and dried biomass concentration being produced as well as associated experimental kinetic parameters after cultivation with different xylose + glucose initial concentration levels maintained at 10:1 ratio. S1 is the xylose concentration (g/L); S2 is the glucose concentration (g/L); Pmax is the maximum xylitol concentration (g/L); x is the dried biomass concentration at maximum xylitol concentration (g/L).
Fig. 1
figure 1

Co-substrate xylitol production at a low level of glucose.

Additionally, the low level of glucose supports the conversion of xylose into xylitol. Xylose is directly converted into xylitol by XR in the presence of coenzyme, namely, nicotinamide adenine dinucleotide phosphate (NADPH) or nicotinamide adenine dinucleotide (NADH). NADPH coupling oxidation assay for XR revealed that wild type C. tropicalis exhibited higher specific XR activity when cofactored with NADPH rather than NADH by 7.8 times33. Continuous NADPH regeneration is a necessity in an efficient xylitol production system through the pentose phosphate pathway (PPP) with enzymatic coupling reduction processes of glucose-6-phosphate dehydrogenase (G6PDH) and 6-phosphogluconate dehydrogenase (6PGDH)20. The PPP is primarily an anabolic pathway that utilizes the C6 sugars of glucose to generate C5 sugars and reducing equivalents. The preference for using glucose as a carbon and energy source arises from its ease of metabolism and efficient energy extraction properties. The presence of a lower level of glucose could enhance NADPH supplement and further benefits xylitol production3. On the other hand, a low level of glucose can also enhance the secretion of extracellular xylitol as less of the xylitol formed is oxidized into xylulose by xylitol dehydrogenase (XDH) with NAD+ as an electron acceptor. The presence of glucose can hinder further oxidation by slowing down the glucose-6-phosphate (Glucose-6-P) production from xylulose via xylulose-5-phosphate (Xylulose-5-P)8. Consequently, the higher xylitol concentration could be achieved by the high fraction of xylose metabolized to xylitol and additional low level of glucose via enhancing NADPH supplement and decreasing xylitol oxidation. Since the highest xylitol yield and productivity were achieved by using xylose + glucose concentrations of 90 + 9.0 g/L, the subsequent hydrolysate experiment was performed with these initial values.

Our results are comparable to Saha et al.29, where they have subjected the Barnettozyma populi NRRL Y-12728 for xylitol production using xylose concentrations from 50 to 75 and 100 g/L. The yeast strain produced 32.2, 49.9, and 58.1 g/L xylitol with a yield of 0.65, 0.70, and 0.71 g/g and productivities of 0.45, 0.52, and 0.43 g/L/h at 50, 75, and 100 g/L xylose, respectively. Eryasar and Karasu-Yalcin.30 investigated strains of C. tropicalis, C. famata, C. guilliermondii, and D. hansenii for xylitol production, the highest xylitol concentrations of 83.28 and 54.07 g/L from 100 g/L xylose were achieved by using C. tropicalis M2 and C. tropicalis M43, respectively. Prakash et al.31 reported that maximum xylitol production of 68.6 g/L from 100 g/L of xylose with a yield of 0.76 g/g and volumetric productivity of 0.44 g/L/h were achieved by using D. hansenii. Xu et al.32 reported that the accumulation of xylitol reached the maximum at an initial xylose concentration of 100 g/L (65.8 g/L) and then decreased when the initial xylose concentration exceeded 100 g/L using C. tropicalis 31,949. C. tropicalis is a promising strain for xylitol production, the higher yield of 0.78 g/g and productivity of 0.90 g/L/h in this study may be attributed to the additional glucose.

Model development and searched parameters assessment

Five batch cultivations with initial xylose (S1) and glucose (S2) concentrations of 18 + 1.8, 36 + 3.6, 54 + 5.4, 72 + 7.2, and 90 + 9.0 g/L have been conducted and used in the estimation of parameters in a complete model through the parameter searching program. The optimal parameter constants are presented in Table 2, while the parameter values of maximum specific utilization rate of xylose (qsmax,1) and glucose (qsmax,2) as well as maximum specific production rate (qpmax) were functionally associated with their corresponding initial sugars concentration (Fig. 2). As can be seen from Fig. 2(A), qsmax,1/qsmax,2 was decreased with S2 up to 5 g/L after which a plateau was reached at ~ 0.4 with R2 of 0.996. The relationship between qpmax and S1 as shown in Fig. 2(B) with R2 of 0.998.

Table 2 Optimal kinetic parameters for all data sets.
Fig. 2
figure 2

The relationships between variables of the maximum specific xylose utilization rate (qsmax,1), maximum specific glucose utilization rate (qsmax,2), and maximum specific xylitol production rate (qpmax) with their corresponding sugars concentration. (A) relationship between qsmax,1/qsmax,2 with glucose concentration, (B) relationship between qpmax with xylose concentration.

The qsmax,1 was restricted by critical xylose concentration (S1,crit) as presented in Eq. (10). The relationships between qsmax,1/qsmax,2 with S2 and qpmax with S1 were fitted to the polynomial equations as shown in Eqs. (11) and (12), respectively.

$$q_{smax ,1} = left{ {begin{array}{*{20}l} {7.397 times 10^{ – 3} S_{1} + 0.4551;S_{1} < S_{1,crit} } hfill {q_{smax 1,crit} ;S_{1} ge S_{1,crit} } hfill end{array} } right.$$

(10)

$$frac{{q_{smax ,1} }}{{q_{smax ,2} }} = – 4.702 times 10^{ – 2} S_{2}^{3} + 8.889 times 10^{ – 1} S_{2}^{2} – 5.47S_{2} + 1.171$$

(11)

$$q_{pmax } – 1.972 times 10^{ – 6} S_{1}^{3} – 3.425 times 10^{ – 4} S_{1}^{2} + 2.112 times 10^{ – 2} S_{1} – 8.559 times 10^{ – 2}$$

(12)

The simulation curves from the model are depicted in Fig. 3 (B, D, F, H, J) alongside experimental data points for comparison. As evident from Fig. 3 (B, D, F, H, J) and Table 3 (Parameter search) with average RSStotal, R2, and MStotal values of 192, 0.982, and 12.8, respectively, the model demonstrated an acceptable fitting of the data within the xylose concentration range of 20 – 100 g/L.

Fig. 3
figure 3

Simulation (line) and experimental (point) data for batch xylitol production by C. tropicalis using xylose and glucose as substrate (g/L): (A) 10 xylose + 1 glucose, (B) 20 xylose + 2 glucose, (C) 30 xylose + 3 glucose, (D) 40 xylose + 4 glucose, (E) 50 xylose + 5 glucose, (F) 60 xylose + 6 glucose, (G) 70 xylose + 7 glucose, (H) 80 xylose + 8 glucose, (I) 90 xylose + 9 glucose, and (J) 100 xylose + 10 glucose. Representations: blue (xylose); yellow (glucose); pink (xylitol); green (dried biomass). The graph represents the mean value ± standard error (SE) values.

Table 3 Summary of related statistical parameters used in fitting quality assessment between the predicted profiles and pure sugars experimental data. S1 is the xylose concentration (g/L); S2 is the glucose concentration (g/L); RSStotal is the total residual sum of squares, R2 is the correlation coefficient, MStotal is the total mean square.

Model validation using optimized parameters

Model validation was carried out by employing the developed model and optimized kinetic parameters to simulate profiles of sugars consumption, biomass, and xylitol production. These simulated profiles were subsequently compared to actual experimental data or expected trends to evaluate the accuracy and effectiveness of the model. Interpolation and extrapolation validations were performed by applying the model to different experimental sets with initial xylose concentrations within and outside of the range of 20 – 100 g/L. Interpolation test was preformed using sets of xylose + glucose concentrations of 27 + 2.7, 45 + 4.5, 63 + 6.3, and 81 + 8.1 g/L. The extrapolation test was performed using 9 + 0.9 g/L. As evident from the simulation profiles shown in Fig. 3 (A, C, E, G, I) and Table 3 (Validation) with average RSStotal, R2, and MStotal values of 132, 0.975, and 8.81, respectively, it provided credible evidence that the model could accurately predict experimental results within the xylose concentration range of 10 – 100 g/L under a xylose and glucose ratio of 10:1.

Optimal kinetic parameters

Optimal kinetic parameters for substrate uptake, biomass production, and xylitol production are shown in Table 2. The proportioning factors represent the relative preference for sugars utilization and biomass production from either xylose or glucose. The factors for xylose (α = 0.59) and glucose (1 – α = 0.41) indicated that the specific xylose uptake was 59% of its maximum value, while that of glucose was 41% of its maximum when the two sugars were presented initially in a 10:1 ratio. The maximum specific growth rate (μmax) represents metabolism or substrate utilization, which is attributed to transport of substrate across the cells membrane, the faster flux is therefore required to achieve high μmax15. In the present model, μmax on xylose (μmax,1) was approximately 5 times that of glucose (μmax,2). It indicated that the yeast grew faster and was more efficient at utilizing xylose compared to glucose. This may be related to the two sugars transporter proteins – CtStp 1 and CtStp2 in C. tropicalis. Growth curve and sugars consumption profiles revealed uptake of both glucose and xylose simultaneously. However, the CtStp1 showed relatively less effect of glucose repression in mixed sugars and was a better transporter of xylose than CtStp234. The substrate affinity constant or Monod saturation constant (Ks) is a parameter to describe the affinity of microorganism for its substrate. It represents the substrate concentration at which half of the maximum specific growth rate was achieved. Ks associated with xylose utilization (Kss,1) was approximately 11 times that of glucose (Kss,2) and Ks associated with cells growth (Ksx) from xylose (Ksx,1) was approximately 7 times that of glucose (Ksx,2) indicating a high affinity for glucose than xylose35. As shown in Eqs. (2) – (5), the substrate limitation effect on xylose utilization (S1/(Kss,1 + S1) was higher than on glucose (S2/(Kss,2 + S2), whereas the effect of glucose limitation (S2/(Ksx,2 + S2) on growth was higher than that of xylose (S1/(Ksx,1 + S1). This could be a possible consequence of the preference not solely relied on facilitated diffusion but also on the cellular metabolic demand and the concentration gradient of the sugars in which xylose is the major substrate for xylitol production and glucose is served for cells growth. Additionally, Ks associated with xylitol production (Ksp) was 4.25 g/L. The xylose limitation effect on xylitol production (S1/Ksp + S1) indicated that a relatively higher level of xylose is required to drive xylose metabolism. This requirement can be explained from the conformation of structure of XR enzyme which contains a hydrophobic binding pocket whereas xylose exhibits significant hydrophilic characteristics6.

Generally, high xylitol production is achieved by using high substrate concentration. However, a decrease in xylitol yield was observed when the xylose concentration increased from 75 g/L to over 100 g/L, even though C. tropicalis is considered an osmotolerant yeast36. The substrate inhibitory constant (Ki) is a parameter used to describe the inhibitory effect of substrate on cells growth (Kix) and xylitol production (Kip). The constants of xylose (Kix,1) and glucose (Kix,2) associated with growth were 95.0 and 151 g/L, respectively. The constant of xylose associated with xylitol production (Kip) was 254 g/L. As shown in Eqs. (2) and (6), the factors of Kix,1/(Kix,1 + S1) and Kip/(Kip + S1) indicated that the growth was more sensitively affected by xylose concentration compared to xylitol production. Similar phenomenon was observed with xylose concentration over 200 g/L, the production of xylitol declined whereas the growth was inhibited37. This may be due to the yeast possessing metabolic pathways that can convert xylose into xylitol through enzymatic reactions which can be independent to active cells division or growth process.

During cultivation, the effects from xylitol concentration (P) occurred simultaneously. The rates of cells growth on xylose (rx,1), xylose utilization (ds1/dt), and xylitol production (dp/dt) were promoted when P was below the threshold levels of 14.0 g/L (Pix,1), 10.4 g/L (Pis,1), and 11.3 g/L (Pip), respectively. A linear depreciate occurred when P was between the threshold and inhibitory levels, until the inhibitory effect occurred at the P = 149 g/L (Pmx,1), 228 g/L (Pms,1), and 111 g/L (Pmp), respectively. This phenomenon can be explained by low levels of xylitol simulating XR28, whereas high levels of xylitol inhibit the XR reaction38. Theoretically, the factors 1 – (PPix,1)/(Pmx,1Pix,1), 1 – (PPis,1)/(Pms,1Pis,1), and 1 – (PPip)/(PmpPip) = 0 would occur when P = Pmx,1, Pms,1, and Pmp, respectively (Eqs. 2, 4, and 6), which indicated that the growth, consumption, and production were completely stopped by xylitol. However, this was not the case as xylose has been consumed to a certain level before P reaches inhibitory levels in a batch cultivation system. The slowdown and eventual cessation of growth as well as xylitol production, especially in the cases of pure sugars cultivation, were thus principally caused by xylose limitation. This caused the values of dx/dt or dp/dt gravitated toward zero as the variable of xylose concentration was directly proportional to the growth and xylitol rate equations (Eqs. 2 and 6). It was also possible for the values of dx/dt and dp/dt to descend and finally reaching zero due to accumulated xylitol concentration (P = Pmx or Pmp) in the fed-batch or continuous systems.

Model validation by extrapolation to xylitol production kinetics in hemicellulosic hydrolysate

The hydrolysate experiment was performed in batch culture system under a xylose and glucose ratio of 10:1. A lower xylitol concentration of 53.5 g/L with a yield of 0.64 g/g and productivity of 0.56 g/L/h was achieved using non-detoxified corncob hemicellulosic hydrolysate compared to pure sugars experiment. Specifically, the utilization of xylose and glucose, as well as the production of xylitol and dried biomass, were compared with pure sugars experiment as shown in Fig. 4. Higher dried biomass production was observed for the hydrolysate after 96 h. This phenomenon may be attributed to the utilization of incompletely depleted xylose and other sugars in hydrolysate as carbon source for yeast growth. Additionally, the lower xylose consumption rate (0.91 ± 0.05 g/L/h vs. 1.13 ± 0.06 g/L/h) incorporated with lower xylitol production rate (0.64 ± 0.03 g/L/h vs. 0.97 ± 0.02 g/L/h) between 12 – 48 h and incomplete xylose utilization after 72 h were observed from hydrolysate experiment. Differences in xylose utilization and xylitol production can be associated with the presence of inhibitors in hydrolysate. Similar effects of inhibitors on xylose consumption and conversion were observed when using C. tropicalis for xylitol production. The inhibitors could cause mitigation in biological activities and enzymatic functions. The activity of XR enzyme regulates the flux of xylose through the metabolic pathway. Higher XR activity typically leads to faster conversion of xylose to xylitol, promoting efficient xylose consumption by microbial cells39. Rafiqul et al.40 found that acetic acid, phenolics, furfural, and HMF in hemicellulosic hydrolysate significantly inhibited XR from C. tropicalis IFO 0618 with concentration-50 (IC50) values of 11, 6.4, 2.3, and 0.4 g/L, respectively. Kaur et al.41 reported that the addition of 2 g/L acetic acid decreased xylitol production rate by 4.18% compared to the control medium lacking inhibitor in cultivation by C. tropicalis OK165575. Therefore, a detoxification method is commonly employed to decrease the inhibitors concentrations and further improve the xylitol production. Kumar et al.42 found that 150 Da polymeric membrane was effective in removal of inhibitors from corn cob acid hydrolysate with simultaneous concentration of xylose. The removal of inhibitors like acetic acid (82.4%) and salts of acid (57.8%), respectively, and a xylitol yield of 62% were successfully achieved using C. tropicalis MTCC 6192. Ahuja et al.43 reported that ~ 93.23% furfurals and 94.62% phenolics in hydrolysate were removed by using activated carbon. A xylitol yield of 0.78 g/g was achieved from the detoxified hydrolysate by P. caribbica MTCC 5703. The removal of furfurals and phenolics effectively improves xylitol production. Moreover, during cultivation, although C. tropicalis can mitigate the inhibitory effects of furfural and HMF by converting them into less toxic compounds, this process requires NAD+, which is also an important cofactor used in xylose metabolism41,44.

Fig. 4
figure 4

Comparison of xylitol production using pure sugars and hydrolysate.

The predictability of the developed model was assessed in the hydrolysate experiment by comparing the experimental data with kinetics profiles. Figure 5 depicts the simulation curves of experimental data by the developed model for prediction of sugars utilization, biomass, and xylitol production. A relatively good fit was achieved with RSStotal, R2, and MStotal values of 356, 0.988, and 23.8, respectively. However, these statistical parameters were lower than those for the pure sugars fitting with RSStotal, R2, and MStotal values of 245, 0.993, and 16.3, respectively (Table 3), due to the underprediction of dried biomass after 96 h. The good fitting of sugars utilization and xylitol production indicated that the current model also demonstrates good predictability for hydrolysate medium, whereas the underprediction of biomass production after reaching the highest xylitol concentration suggested that further modifications to the growth rate equations were necessary.

Fig. 5
figure 5

Simulation (line) and experimental (point) data for batch xylitol production by C. tropicalis using corncob hemicellulosic hydrolysate. Representations: blue (xylose); yellow (glucose); pink (xylitol); green (dried biomass). The graph represents the mean value ± standard error (SE) values.