The investigation presented herein aimed at constructing and characterizing a recombinant S. cerevisiae
strain with enhanced capacity to ferment mixtures of glucose and pentose sugars. The strain harbored and improved fungal pathway for arabinose and xylose utilization. In fact, in anaerobic conditions the strain fermented both L-arabinose and D-xylose. This was not the first time that a recombinant S. cerevisiae
harboring a fungal pentose utilization pathway was obtained [51
]. However, the present report shows an improvement of more than one order of magnitude in L-arabinose utilization when compared to previous similar attempts [51
]. Consistently, the aerobic growth rate on L-arabinose as sole carbon source also showed an estimated 10 fold increase [52
The currently best L-arabinose fermenting S. cerevisiae
strains have been obtained by expression of the bacterial L-arabinose pathway alone [53
] (Figure ), or in combination with a D-xylose isomerase pathway (Figure ) [32
]. For these constructs, co-fermentation of the two pentose sugars has only been achieved by extensive and carefully controlled evolutionary engineering protocols [32
]. In particular, it was demonstrated that evolutionary engineering of combined traits such as fermentation of two sugars easily drifts towards one of the traits, as soon as it becomes preferred over the other. In a mixture of D-xylose and L-arabinose, the selection process towards the utilization of D-xylose accelerated as soon as a preference for D-xylose arose, to the disadvantage of L-arabinose utilization. Thus, co-fermentation of L-arabinose and D-xylose was the result of a delicate equilibrium of different selection pressures, with a yet un-clarified molecular explanation.
In contrast, the present study showed that the presence in TMB3664 of the additional activities needed for L-arabinose fermentation did not affect D-xylose fermentation at all. At the rates observed in the current study, arabinose fermentation rather improved ethanolic fermentation of D-xylose, indeed representing a pure net advantage for the strain in sugars mixtures (Table ).
L-arabinose and D-xylose co-utilization by S. cerevisiae
has previously been demonstrated by combining a bacterial L-arabinose isomerase pathway and a fungal reduction/oxidation D-xylose pathway [30
] (Figure ). However, AR expressing strains convert L-arabinose to L-arabitol [30
], which represents a dead end in the metabolism of such strains and, in addition, can inhibit the bacterial enzymes for L-arabinose utilization [30
]. Thus, a solution to this problem was to exploit the natural efficiency of AR to convert L-arabinose to L-arabitol, and to further channel L-arabitol into central metabolism via the fungal L-arabinose utilizing pathway (Figure ). Therefore, strain TMB3664 was constructed by expressing an improved fungal pentose utilization pathway, combining a mutated AR engineered for enhanced NADH preference with strictly NAD+
/NADH dependent ALX, LAD and XDH. Indeed, more than 50% of the L-arabitol generated by AR in strain TMB3664 was further channeled into central metabolism in anaerobic mixed sugar fermentation. In addition, the yield of the second by-product xylitol was reduced to less than 10%.
The enzymes involved in the reduction-oxidation steps of fungal pentose sugar assimilation display different co-factor preferences [45
]. By-product accumulation [55
], poor aerobic growth [52
] and poor anaerobic ethanol production [51
] not only in recombinant S. cerevisiae
strains, but also in natural L-arabinose consuming yeast species such as Candida arabinofermentans
and P. guilliermondii
, have been ascribed to the difference in co-factor preference of the L-arabinose metabolizing enzymes [45
]. The activity of the heterologous enzymes expressed in S. cerevisiae
TMB3664 was 5 to 50 times higher than previously reported for similarly constructed recombinant S. cerevisiae
] and comparable to the enzyme activity levels found in natural L-arabinose fermenting yeast [45
]. The pathway reconstructed in TMB3664 combines a mutant AR, with increased relative coenzyme preference for NADH [37
], with an entirely NAD+
/NADH dependent downstream pathway [27
]. This partially relieved the strain from NAD+
] and allowed it to channel L-arabinose and D-xylose towards the formation of ethanol and even biomass, with reduced by-product formation.
L-arabinose metabolism of the natural pentose fermenting yeasts C. arabinofermentans
and P. guilliermondii
has been characterized in relation to sugar concentration and aeration [56
]. Although these yeast strains were not grown under strict anaerobic conditions, it may still be useful to compare the product formation pattern of TMB3664 with those of C. arabinofermentans
and P. guilliermondii
grown in different aeration conditions including severe oxygen limitation [56
]. C. arabinofermentans
consistently produced less L-arabitol than P. guilliermondii
, except under the most extreme oxygen limitation, when both yeast strains displayed an L-arabitol yield on consumed L-arabinose of approximately 0.6 g/g. Considering that, in contrast to P. guilliermondii, C. arabinofermentans
exhibits moderate but detectable NADH-dependent ALX activity, in addition to high NAD+
-dependent LAD activity [45
], it is tempting to rank the three yeast strains in terms of co-factor balance of their pentose utilization pathway and to relate this to the strains' L-arabitol and xylitol formation. Thus, TMB3664, harboring the mutated AR and the exclusively NADH-dependent ALX from A. monospora
, would be considered the most co-factor balanced of the three strains, and in fact it produces the lowest yield of L-arabitol and xylitol and has the highest ethanol production rate [45
]. This comparison delineates a trend and points out the importance of equilibrated co-factor utilization for minimal by-product formation in pentose fermenting recombinant S. cerevisiae
strain based on a reduction/oxidation pathway for pentose metabolism.
In conclusion, the present work describes an L-arabinose and D-xylose co-fermenting recombinant S. cerevisiae strain expressing a pentose utilization pathway of entirely fungal origin. Anaerobic product formation by such strains such strains is largely influenced by the co-factor preference of the enzymes constituting the pathway.