Several studies on ethanol production by wild-type strains of Z. mobilis on industrial starch-based raw materials have been reported. Bringer et al. [69] investigated an industrial-scale process and Poosaran et al. [70] evaluated a cassava-derived starch hydrolysate. In the latter case in a batch culture at controlled T = 30 °C and pH = 5.0, fermentation using Z. mobilis ZM4 gave an ethanol yield of 95% theoretical, a productivity of 6 gL-1 h-1 and a final ethanol concentration of 114 gL-1. Under the same conditions, a strain of Sac — charomyces uvarum gave an ethanol yield of 90% theoretical, a productivity of 4gL-1 h-1 and a final ethanol concentration of 106gL-1 for a cassava starch suspension (23% glucose equivalent). A comparative batch and continuous culture study with starch hydrolysate using yeast and Z. mobilis 29191 has also been reported by Beavan et al. [71].

Extensive studies with various strains of Z. mobilis have been reported by using sugar cane syrup and molasses [72-76] and for sugar beet mo­lasses [77,78] with evidence of yield reductions on sucrose based media due to production of the fructose polymer levan as by-product [3,6] and rate reductions due to high salt concentrations in the molasses. Improved produc­tivities were reported following membrane desalting of high salt-containing sugar cane molasses [72].

Most recently, Davis et al. [79] studied the fermentation of a hydrolyzed waste starch stream from flour wet milling using both Z. mobilis ZM4 and an industrial ethanol-producing strain of S. cerevisiae. With glucose concen­trations in the range 80-110gL-1, Z. mobilis ZM4 demonstrated superior fermentation characteristics. In a repeated batch process (five cycles), rapid concentration of the cells and increased productivities were achieved by cell settling between batches using the flocculent strain Z. mobilis ZM401 (ATCC 31822) as characterized by Skotnicki et al. [80]—see Fig. 6.

Similar flocculent mutants of wild-type Z. mobilis strains CP4 and ATCC 29191 have been isolated by Lawford et al. [16] and Fein et al. [81] using a spe­cially designed chemostat. These strains were deposited with the ATCC as strains 35 000 and 35 001, respectively. The use of such flocculent cultures was demonstrated to increase volumetric productivity by as much as ten-fold [82]

and may have considerable potential in future large-scale processes for more stable fermentations.

Recombinant strains of Z. mobilis developed for xylose utilization have been evaluated on various agricultural residues including oat hull hydrolysate produced by the Iogen process [40]. Oat hull hydrolysate contains glucose, xylose and arabinose in a mass ratio of 8 : 3 : 0.5. Synthetic hydrolysate (6% w/v glucose; 3% w/v xylose; 0.75% w/v acetic acid) at initial pH 5.75 was mixed with either 2 ml L-1 corn steep liquor (CSL) or 1.2 g L-1 di-ammonium phosphate as N source and used for evaluation of ethanol production. From the results it was concluded that the highest productivity was achieved with Z. mobilis ZM4 (pZB5). In this and other studies, CSL was also found to be an effective nutrient source to replace yeast extract in the fermentation media for Z. mobilis [83-85].

Further studies were reported by Mohagheghi et al. [41] with an integrant strain (designated Z. mobilis Fig. 8b) derived from ZM4 (pZB5) using over­limed corn stover hydrolysate. The hydrolysate contained 16 gL-1 glucose, 69gL-1 xylose and 11 gL-1 acetic acid at pH = 5.0. This medium was sup­plemented with 100 gL-1 glucose and diluted to various concentrations prior to fermentation. The authors found that up to 50 g L-1 ethanol was produced by the integrant strain with diluted 80% corn stover hydrolysate. Yields of 83-87% theoretical (based on sugars utilized) were reported.

One of the potential issues for large-scale Z. mobilis fermentations is whether or not contamination control is needed particularly in the presence of ethanol-tolerant strains of Lactobacilli. Such contamination constitutes a problem in many yeast-based processes and can reduce yields by an es­timated 2-5%. However, its impact is reduced as pH decreases to 3.0-3.5 towards the end of batch fermentation (in the absence of pH control). “Acid washing” of the residual yeast at this pH or lower is often used to minimize contamination in yeast subsequently used in a repeated batch process. Z. mo­bilis is more sensitive to low pH than S. cerevisiae and contamination was identified as a problem by Bringer et al. [69] in their study on an industrial — scale process for conversion of starch to ethanol using Z. mobilis although Lawford and Rousseau [63] demonstrated that lactic acid in such circum­stances is not likely to be inhibitory to Z. mobilis. Interestingly, although rarely observed in Z. mobilis fermentations due to the usual high metabolic flux rates in the ED pathway, conditions have been reported which can pro­mote lactic acid synthesis in Z. mobilis [37,85].

The issue of contamination control was addressed directly by Grote et al. [86] in which a continuous culture of Z. mobilis ZM4 was directly contaminated with a 10% (v/v) inoculum of Lactobacillus sp. isolated as an ethanol-tolerant contaminant from an industrial plant (Grain Processing Corporation, Muscatine, Iowa). It was found at D = 0.1 h-1 under condi­tions of glucose limitation, pH control at 5.0 and ethanol concentrations of 60-65 gL-1, that the addition of the contaminant caused only a temporary disturbance in the process. Steady state conditions with no evidence of sus­tained contamination were regained within five to six generations. These results suggest that contamination is not likely to be a significant problem once an active culture of Z. mobilis is established providing that the pH is maintained above 3.5-4.0. A similar conclusion was reached in a recent study [87] using an acid-tolerant strain of Z. mobilis under non-sterilized feed and operating conditions.

3