In the literature, glycerol has been extensively employed as a carbon source in propionic
acid fermentations by P. acidipropionici. Dishisha et al. (2013]), employed glycerol and potato juice in a fermentative process using high-cell-density
sequential batches with cell recycle. In their study, final propionic acid concentrations
of 43.8 and 50.8 g L?1 were reached using glycerol and potato juice, respectively. In the literature, another
study (Liu et al.2011]) reports batch processes also using glycerol as a carbon source. In this study, the
final concentrations of propionic, acetic, and succinic acid reached 18.1?±?0.6, 0.54?±?0.09,
and 1.10?±?0.05 g L?1, respectively. In this same study, Liu et al. also report results using glucose as
a carbon source; the final concentrations of propionic, acetic, and succinic acid
reached 11.5?±?0.45, 2.57?±?0.12, and 0.55?±?0.03 g L?1, respectively. According to the results in the literature, the final concentration
of propionic acid was higher when using glycerol and potato juice as a carbon sources
in a high-cell-density process with sequential batches and cell recycle than that
obtained in our studies using sorbitol. However, when using glycerol in batch processes,
the final concentrations of propionic and succinic acid were lower than those obtained
using sorbitol as a carbon source, as in the present work. In addition, fermentations
employing sorbitol as a carbon source produced the most interesting results when compared
to those obtained with glucose.
The percentages of recovered carbon were 89.2?±?5.3, 72.3?±?3.9, and 74.5?±?4.2% in
the first, second, and third sequential batches, respectively. These results can be
explained by the following hypothesis: the cells only used approximately 10% of the
carbon source for maintenance, and the cell growth in the first batch and in the second
and third sequential batches used approximately 30% due to the increase in cell concentration.
In sorbitol fermentations, a cellular maintenance coefficient of 0.039 g g?1 h?1 was obtained in the first batch, which increased from 0.044 g g?1 h?1 to 0.051 g g?1 h?1 in the second and third sequential batches. (Table 3 – The meaning of abbreviations in Table 3 can be found in Table 1). These results are similar to those obtained by Goswami and Srivastava (2000]) using lactose (initial concentration of 47.7 g L?1) as the carbon source in a fed-batch experiment, where the cellular maintenance coefficient
was 0.038 g g?1 h?1. Table 3, shows acetic acid yields, YAA/S, of 0.029 g g?1 for the second and third batches. These results are similar to those obtained by
Zhang and Yang (2009]), using an adapted culture of P. acidipropionici in FBB fermentation (0.027?±?0.003 g g?1). However, Blanc and Goma (1987]), obtained a 0.140 g g?1 yield for acetic acid. Accordingly, when compared with household refuse enzymatic
hydrolysate as a carbon source, sorbitol fermentation showed lower values of acetic
acid yield. The most interesting result obtained for succinic acid yield, YSA/S in Table 3, is for the third batch (0.212 g g?1). Zhang and Yang (2009]) obtained lower results for YSA/S (0.073?±?0.002 g g?1). In the present study, YX/S was 0.366 g g?1 in the first batch, similar to the 0.362 g g?1 value obtained by Goswami and Srivastava (2000]) in their study with lactose as a carbon source. The definitions of the abbreviations
used in Table 3 can be found in Table 1.
Table 3. Yield coefficients and cellular maintenance (m) for the three sequential sorbitol
fermentation batches usingP. acidipropionicioperating in four independent vessels
In our study, the productivity of propionic acid was approximately 0.5 g L?1 h?1 (Table 4). (Blanc and Goma 1987]) obtained similar propionic acid productivity (0.4 g L?1 h?1) with sugar mixtures from hydrolysis of household refuse in batch experiments. Another
work (Liu et al., 2012]) reports lower propionic acid productivities using xylose (0.23 g L?1 h?1) and corncob molasses (0.28 g L?1 h?1) in fed-batch experiments. Dishisha et al. (2012]) studied propionic acid production from glycerol using immobilized cells on polyethylenimine-treated
Poraver (PEI-Poraver) and Luffa (PEI-Luffa). In their study, productivities of propionic
acid were 0.86 and 0.29 g L?1 h?1 using PEI-Poraver and PEI-Luffa, respectively. Blanc and Goma (1987]) obtained a productivity of 0.10 g L?1 h?1 for acetic acid while in the present work, the productivity of acetic acid decreased
from 0.05 g L?1 h?1 in the first batch to 0.03 g L?1 h?1 in the third sequential batch (Table 4). Other results are compared in Table 5. The definitions of the abbreviations used in Table 5 can be found in Table 1.
Table 4. Organic acids and cell growth productivity for the three sequential sorbitol fermentation
batches usingP. acidipropionicioperating in four independent vessels
In the present study, the average propionic acid/acetic acid (P/A) molar ratio was
increased from 9.7 in the first batch to 13.8 in the third sequential batch when using
sorbitol as a carbon source. Liu et al. (2011]) obtained a higher P/A molar ratio of 27.1 using glycerol and a lower P/A molar ratio
of 3.63 using glucose as a carbon source. Another study (Zhu et al.2010]) using glycerol as a carbon source reached a P/A mass ratio of 13.10 (P/A molar ratio
of 10.6) which is similar to the P/A molar ratio of the first batch (9.7) using sorbitol
as a carbon source. As expected, the propionic acid/succinic acid (P/S) molar ratio
decreased from 10.3 in the first batch to 3.7 in the third sequential batch, once
the concentration of succinic acid increased from 6.1?±?2.1 to 14.8?±?0.9 g L?1 and that of propionic acid decreased from 39.5?±?5.2 to 34.4?±?1.9 g L?1 (Table 6).
Table 6. Propionic acid/acetic acid (P/A) and propionic acid/succinic acid (P/S) productivity
ratios of three sequential sorbitol fermentation batches usingP. acidipropionicioperating in four independent vessels
The best propionic acid yield obtained, YPA/S, was 0.613 g g?1 for the first batch. This result was higher than the values found in the literature;
for example, Blanc and Goma (1987]), reached a lower propionic acid yield of 0.552 g g?1 using products of hydrolysis of household refuse; Zhang and Yang (2009]), using an adapted culture of P. acidipropionici in FBB fermentation, obtained a YPA/S value of 0.59?±?0.02; and Liu et al. (2011]), obtained a YPA/S of 0.475?±?0.017 and 0.303?±?0.012 from glycerol and glucose, respectively. Dishisha
et al. (2012]), obtained the best result for propionic acid yield (0.74 mol mol?1 or 0.595 g g?1) from glycerol. When using sorbitol, as we reported herein, it is possible to obtain
a superior yield of 1.51 mol mol?1 (0.613 g g?1) (see Table 5).
Substrate consumption rates and biomass rates were similar in each batch. The succinic
acid production rate increased from 0.022 g h?1 to 0.053 g h?1 while that of acetic acid (from 0.012 g h?1 to 0.007 g h?1) and propionic acid (from 0.137 g h?1 to 0.109 g h?1) decreased over the batches. The specific cell growth rate remained at a constant
value of 0.014 h?1 over the batches (Table 7; the definitions of the abbreviations used in Table 7 can be found in Table 1). Zhang and Yang (2009]) obtained a specific cell growth rate of 0.050?±?0.002 h?1 when working with the original culture and 0.16?±?0.02 h?1 with the adapted culture in free-cell fermentation.
Table 7. Instantaneous and specific rates of the three sequential sorbitol fermentation batches
usingP. acidipropionicioperating in four independent vessels
In conclusion, the final concentration of propionic acid obtained in this study using
sorbitol as a carbon source was higher than that obtained in other studies using glucose,
household refuse, and glycerol (in some specific operation modes) presented in the
literature. Employing sorbitol, an unexplored carbon source, in fermentation reactions,
allowed reducing the acetic acid yield when compared to glucose and household refuse
enzymatic hydrolysate as carbon sources. Furthermore, these results, all obtained
in quadruplicate, are important for the development of a continuous fermentation process
in the future.