It was thought that glucose and hydrolysed sago starch (HSS) could promote the growth of the cells better than raw sago starch (RSS) and gelatinized sago starch (GSS). However, the trends obtained in all cases were similar, although it was observed that in the log phase the uptake of the GSS and RSS forms was slightly faster than that of the glucose and HSS. Similar observation was reported by Shibata et al. ([2007]). They found that the performance of E. faecium using RSS was better than with corn, potato and wheat starches. It could be desirable to have RSS as the most appropriate form of starch to be used as direct substrate for the fermentation process; however, it is problematic to manage the RSS slurry at high concentration. Although Shibata et al. ([2007]) did not reported how they prepared the raw sago starch, and if they sterilized the sago slurry, then; the form of starch used perhaps was the gelatinised sago starch. The gelatinised sago starch can be used as substrate but a very low concentration because when the concentration of starch is increased the viscosity of the media is very high. The viscosity of the slurry increases proportionally with the concentration of the starch. In general, when the starch concentration increased, sedimentation occurred due to saturation or insolubility. This situation was visualized by[Zhang and Cheryan (1994]); therefore, these researchers used liquefied starch to avoid these problems and they improved the process by using the amylolytic Lactobacillus amylovorus strain. Shibata et al. used E. faecium No. 78 cultivated in RSS at a concentration no higher than 20 g/l, which is very low for an industrial application. Therefore, it could be reasoned that using batch and its extension RBF might be advantageous to produce higher LA concentration at a similar production rate.
The RBF mode was applied with the aim to reuse the cells in order to improve the overall productivity of the fermentation. One of the key points in fermentation technology is to maintain the cells in a stable and perpetual state of productivity and to avoid damage to the cells, the aim being to increase their reusability for longer periods. This pattern is well observed in ethanol fermentation where the yeast can be reused 400–600 times, but only if the concentration of ethanol is 7-10%. As reported by Amorim et al. ([2011]) higher ethanol concentration affects the viability of the cells and their reusability falls notably. Similarly, longer exposition of E faecium to a high concentration of LA resulted in lowered productivity. This fact was observed as the behaviour of the specific growth rate, which was affected by the LA. In Figure4 it was observed that the maximum specific growth (μmax) was found at a low concentration of starch and higher availability of a nitrogen source; for instance, in the first cycle to synthesize the necessary proteins for the metabolism of E. faecium; especially, the synthesis of amylases to produce glucose for growth was observed as biomass production. In addition, due to the low concentration of glucose the effect of osmotic stress was avoided, which could help in the productivity of the cells. In subsequent cycles, the growth rate decreased, and contrary to the first cycle, less availability of nitrogen source for a higher population of microorganisms decelerated the growth. Although the nitrogen source could affect the growing of the cells the general trend was a slow growing rate and maintained LA productivity.
On the other hand, as other studies have found, there is an optimum concentration of starch around 20–60 g/l, where the fermentation can go faster but is not limited by this concentration (Narita et al.[2004]; Naveena et al.[2005]; Okano et al.[2009];[Petrova and Petrov 2012]; Yun et al.[2004]). It is true that a higher concentration of starch to produce higher LA titter is possible, but the time process will also increase proportionately. To produce the highest concentration of 42.5 g/l of LA using E faecium and LSS it took 30 h, approximately (Figure5). Using raw starch, ([Petrova and Petrov 2012]) reported the highest amount of LA (37.3 g/l) from 40 g/l of raw starch during 48 h of fermentation using Lactobacillus paracasei B41. In our study it was observed that this concentration of LA has a strong effect on the growth of E. faecium, because after this concentration the production of LA was very limited. After the said concentration, the productivity decreased and the fermentations stopped with an asymptotic trend. Although using LSS form had some advantages because the effect of osmotic stress that could impair the cells as a result of glucose being used directly was avoided, the productivity production of LA was only around 36.3 g/l (Table1). The fact that E. faecium has a faster glucose uptake compared with other microorganisms, especially those having amylolytic abilities such L amylovorus, L. manihotivorans (Shibata et al.[2007]), could be valid only at a low starch concentration (less than 20 g/l).
On the other hand, if it is possible to increase the biomass by increasing the nitrogen source to improve the productivity; then it is possible to reduce the fermentation time and increase the LSS concentration. Here, the key point is to decrease the exposure time of the cells to LA. This same situation applies also for ethanol fermentation (Amorim et al.[2011]); when ethanol concentration increases, the yeast starts to lose activity due to a high percentage of the active biomass dying. Then, the numbers of cycles to reuse the cells becomes reduced and the overall productivity of the system fails. Shibata et al. ([2007]). only used a concentration of no more than 20 g/l and the dilution rate was low, because increasing the dilution rate is probably a way to wash out the amylolytic enzymes produced by E. faecium, although there are reports that amylase enzymes could bind to the cell membrane ([Walker 1965]; Anderson et al.[1989]). Then, the system becomes inefficient at a high dilution rate and a high concentration of sago starch and, even worse, because of the clogging of the membrane system at a high concentration of suspended solids ([Zhang and Cheryan 1994]).
Moreover, it was reported that sago starch was a difficult substrate for the amylolytic enzymes, due to the lack of suitable surface ready for the attack of the enzymes, and that it did not have natural pores or deep channels to facilitate the action of the enzymes (Uthumporn et al.[2010]; Nor Nadiha et al.[2010]). It was observed through SEM that sago starch has a very smooth surface and granule size of 6–50 μm (Figure7), having a distribution of 16% of particles with a size of 42 μm, which agrees fairly well with the finding of Nor Nadiha et al. ([2010]) and Nitta et al. ([2008]).
The characteristics of sago starch can be overcome with the use of Termamyl® SC to liquefy the sago starch with the aim of facilitating the subsequent attack of the amylase produced by E. faecium. The enzyme termamyl can only produce oligosaccharides and not glucose. In doing so and taking into account that the liquefaction process takes less than 2 h, and that the actual cost of Termamyl® SC is reasonably cheap, it can be used for this purpose.
For instance, the cost of Termamyl is around US$ 6715.60/ton and the dose of enzymes per ton of starch is around 0.625 kg, which means $4.20 US/ton of starch (Alagaratnum 2011, personal communication). Through a simple calculation and taking into account the price forecasted by Energetics[Incorporated (2003]) of about 550.0-1100.0 $US/ton, and an efficiency of 90% of conversion from starch to LA, as the percentage of the enzyme cost in the total cost of LA production could be considered as low (0.60%). Yet, this percentage in the total cost could be even lower if we consider the higher price for the LA food grade produced by Purac.
On the other hand it was observed that for this specific strain of E. faecium and from these results obtained, it was assumed that the LA concentration which induced the growth inhibition of E. faecium was in the range of 45 g/l (500 mM). Assuming that this is the growth inhibition concentration, with the LA pKa of 3.86, and using the following equations:
(1)
where [L-] is the dissociated and the non-dissociated LA, the fermentation was controlled at pH 6.5, and the concentration of non-dissociated LA that produced the growth inhibition of E. faecium was calculated at approximately 1.15 mM. This critical concentration of LA which affects the growth of E. faecium was lower compared to those reported for Lactobacillus rhamnosus (4.7 mM) and Lactobacillus helveticus (5 mM) (Gonçalves et al.[1997];[Timmer and Kromkamp 1994]). Therefore, it is suggested that E. faecium lowered the rate of LA production when it starts to reach a concentration close to 45 g/l and then modified its metabolism to synthesize exo-polysaccharide. E. faecium was able to grow at a rate as high as 0.4 h-1 in the first fermentation cycle. In the second cycle μ decreased to 0.27, then to 0.11, and from the fourth to eleventh cycle μ was almost constant at level of 0.028-0-0.02 h-1 (Figure4). Even at low growth rate the strain maintained its LA production rate at a level of 1 g/lh.