The production of ethanol from renewable resources (corn, wheat, cellulosic biomass etc.) is currently a hot topic. Although there are clear economic benefits to developed countries, such as the USA and European nations, where the availability of feedstocks for fuel ethanol production could greatly reduce dependence on foreign oil, the cleaner burning and greenhouse gas neutrality of ethanol offer worldwide environmental benefits. If ethanol is to be an economically viable fuel, then its production process must be highly efficient. Today, it is significantly less expensive to produce gasoline than it is to produce an energetically equivalent amount of ethanol.
Much research effort is being focused on improving the economy of the current microbial-based ethanol production processes. Efficiency gains are largely dependent upon increasing the rate of ethanol production without sacrificing yield, while minimizing the cost of this improvement. Ultimately, improvements in the rate of ethanol production will be limited by the ethanol-producing microorganism being used. Currently, ethanol is produced by fermentation of glucose (typically derived from corn) by yeast, as a byproduct of the glycolytic pathway (Fig. 1) which yields energy for growth. Ethanol production by yeast eventually slows down and stops due to the toxic effect of ethanol. Also, a portion of the carbon source (glucose) is utilized by the yeast for biosynthesis rather than conversion to ethanol. These and other factors limit the efficiency of ethanol production by yeast. While strain improvements are possible, the constraints of maintaining a viable organism will eventually limit success in this area.
If it is true that practical limitations on strain improvement will be the major factor preventing further efficiency improvements in yeast-based ethanol production, then why not get rid of the microorganism altogether and use only the enzymes involved in the conversion of glucose to ethanol? This ‘cell-free ethanol’ production concept has a number of potential advantages over the conventional microorganism based process including greater process flexibility (i.e. ability to operate at higher temperature, or higher ethanol concentration), more freedom to manipulate enzymes (i.e. removing feedback inhibition mechanisms), and the ability to easily optimize the production process by altering enzyme levels. All of these could result in significant efficiency gains.
The concept of cell-free ethanol production is certainly not novel. In 1897 Eduard Buchner found that sugar could be converted to ethanol and carbon dioxide from cell-free extracts of yeast.1 He proposed that this conversion was carried out by protein which he termed ‘zymase’. This discovery won Buchner the Nobel Prize in 1907 and initiated the study of metabolism. However, Buchner's cell-free extracts fermented glucose to ethanol far less effectively than intact yeast.2
Little attention was paid to any possible industrial application of this idea until the work of Scopes3 who investigated cell-free ethanol production primarily, it seems, as an exercise to pave the way for production of more valuable products (i.e. pharmaceuticals). Scopes resolved the problems encountered by Buchner and others showing that sustained ethanol production was possible with a reconstituted cell-free system. One of the key issues was found to be the problem of ATP production by the cell free pathway. Inside a microorganism, ATP is hydrolyzed to provide energy for other cellular processes. With the cell-free system it is necessary to include a generic ATPase to carry out this task and allow for continued ethanol production. Scopes found that a sustained and high rate of ethanol production was only possible if the rate of ATP hydrolysis was closely matched to the rate at which glucose was being fed into the pathway. He also found that he could avoid this problem by including arsenate in the reaction medium. The enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH) can accept arsenate instead of phosphate as a substrate resulting in the synthesis of 1-arseno-3-phosphoglycerate which rapidly breaks down to arsenate and 3-phosphoglycerate (3PG). Thus the addition of arsenate allows an ATP production step (the phosphoglycerate kinase step) to be bypassed. This solves the ATP accumulation problem since in the presence of arsenate the pathway exhibits no net production of ATP.
So it seems that the question of whether or not ethanol can be made in a cell-free process has long been answered with a definitive ‘yes’. However, the question of whether or not this approach allows for more efficient (more economical) ethanol production has not yet been addressed. Since improvements in the ethanol process are largely dependent upon increasing the rate of ethanol production, this question has two parts; can ethanol be made faster using the cell-free process? and if so, can this rate improvement be accomplished in a cost effective manner?
A typical industrial ethanol fermentation exhibits a maximum rate of ethanol production of about 1.7 mmol L−1min−1 (E.J. Allain, unpublished data) whereas the Scopes cell-free system was found to be capable of producing ethanol at a much greater rate of approximately 9 mmol L−1min−1. However, since Scopes was attempting to mimic enzyme concentrations found inside the cell, the total amount of enzyme required to get this rate (∼20 g L−1) is likely not economically achievable. Still, the massive rate increase obtained with the cell-free system poses the question; What would the rate be if the total enzyme concentration is reduced to a more industrially realistic level?
The use of a mathematical model of the cell-free glycolytic system4 suggests that if the relative amounts of each individual enzyme remain constant, the rate of ethanol production is proportional to the total amount of enzyme added (i.e. the sum of the concentrations of each of the twelve enzymes involved in the pathway). This indicates that the cell-free system should be able to equal the current industrial rate of ethanol production with a total enzyme concentration of about 2.7 g L−1. This is approximately the same concentration of glycolytic enzymes one would observe if all of the yeast cells in a fuel ethanol fermentation were lysed and their contents released into the fermentation broth. This work suggests that in order to achieve an increase in rate over the microorganism based process, the total concentration of the enzymes in the cell-free process must be greater than 2.7 g L−1. While this is certainly more reasonable than 20 g L−1, it is still quite a lot of enzyme.
The above conclusions assume that the relative concentrations of each of the glycolytic enzymes in the cell-free process are the same as they would be inside a yeast cell. However, one of the potential advantages of the cell-free system is that the concentration of each individual enzyme can be readily changed. The glycolytic model was again used to examine the effect of changing the enzyme distribution. An optimization algorithm was developed that maintained total enzyme concentration constant but changed relative enzyme concentrations in an attempt to maximize ethanol production rate. The model predicted that using the same total concentration of enzyme (2.7 g L−1), a substantial increase in ethanol production rate could be achieved (E.J. Allain, in prep.). The optimized enzyme distribution rate of 3.7 mM min−1 is more than twice the ethanol production rate calculated using physiological enzyme levels. A total enzyme concentration of 1.3 g L−1 for the optimized enzyme distribution should match the current industrial microbial-based ethanol production rate. The model also predicts that increasing the total enzyme using this distribution proportionally increases the ethanol production rate, but the rate increase is much steeper than that observed with the physiological distribution (Fig. 2). A comparison of the relative amounts of individual enzymes between the physiological distribution and the optimized distribution is shown in Fig. 3.
Unfortunately, it is not yet possible to conduct an accurate economic analysis of cell-free ethanol production since it is not known how much the enzymes might cost if produced on an industrial scale. However, the modeling work suggests that ethanol can be made much faster and more efficiently using a cell-free process. Consider the 1.3 g L−1 of enzyme required to match the production rate observed in a conventional fuel ethanol fermentation. This same rate requires a dry weight yeast concentration of approximately 5 g L−1 (assuming a typical fuel ethanol fermentation has a yeast cell concentration of 4 × 1011 cells L−1 and the dry weight of a single yeast cell is 1.3 × 10−11 g). Thus the cell-free system is predicted to be roughly four times more efficient than the microbial-based process.
While these numbers are quite encouraging, the true potential of cell-free ethanol production is dependent upon being able to recycle the enzymes. One of the most straightforward ways to accomplish this would be to immobilize the enzymes on a solid support or gel such that they may be recovered after batch ethanol production or, better, utilized in a continuous process. Immobilization may also increase the stability of these enzymes in an industrial process.5 Another possibility takes advantage of the fact that with a cell-free system, the process conditions would no longer need to be constrained to keep the organism viable. Thus the reaction temperature could be raised from the conventional 32 °C up to a much higher temperature (80–95 °C). If one could find or engineer glycolytic enzymes that could handle these high temperatures (possibly from thermophilic organisms) then it may be possible to design a continuous process where ethanol is ‘distilled’ from the reaction as it is made. This would not only solve the enzyme recycling problem but the elevated temperatures could also significantly increase the ethanol production rate. The higher temperatures might also make it easier to hydrolyze recalcitrant feedstocks such as lignocellulose.
The use of a cell-free system for ethanol production offers further opportunities for efficiency increases that would be difficult or impossible to achieve using microorganism based production. Enzymes could be altered without worrying about consequences affecting the viability of the organism. For example, one of the key regulatory enzymes in glycolysis is phosphofructokinase. Since one of the substrates for this enzyme is ATP, high ATP levels should increase the rate of this reaction. However, phosphofructokinase also has an allosteric site that binds ATP. Binding of ATP to this allosteric site inhibits the enzyme. Removing the allosteric site could boost flux through the pathway resulting in an increased rate of ethanol production.
The successful implementation of cell-free ethanol production could open the door for similar processes to produce other valuable products such as chemical intermediates or pharmaceuticals from renewable resources. Perhaps research could find a way to take advantage of the ATP that is produced as a byproduct, utilizing this molecule to drive energy requiring biosynthetic reactions catalyzed by other cell-free pathways. Perhaps in the future we will see industrial scale cell-free biorefineries that produce a multitude of products from simple renewable feedstocks or waste materials. In any event, the potential for cell-free ethanol production combined with the high interest in renewable fuels suggests that this idea deserves serious consideration for further research.