Challenges in enzymatic hydrolysis and fermentation of pretreated Arundo donax revealed by a comparison between SHF and SSF

Abstract

The perennial herbaceous crop Arundo donax is a potential feedstock for second-generation bioethanol production. In the present work, two different process options were investigated for the conversion of two differently steam-pretreated batches of A. donax. The pretreated raw material was converted to ethanol with a xylose-consuming Saccharomyces cerevisiae strain, VTT C-10880, by applying either separate hydrolysis and fermentation (SHF) or simultaneous saccharification and fermentation (SSF). The highest overall ethanol yield and final ethanol concentration were achieved using SHF (0.27 g g⁻¹ and 20.6 g L⁻¹ compared to 0.24 g g⁻¹ and 19.0 g L⁻¹ when SSF was used). The performance of both SHF and SSF was improved by complementing the cellulolytic enzymes with hemicellulases. The higher amount of acetic acid in one of the batches was shown to strongly affect xylose consumption in the fermentation. Only half of the xylose was consumed when batch 1 (high acetic acid) was fermented, compared to that 94% of the xylose was consumed in fermentation of batch 2 (lower acetic acid). Furthermore, the high amount of xylooligomers present in the pretreated materials considerably inhibited the enzymatic hydrolysis. Both the formation of xylooligomers and acetic acid thus need to be considered in the pretreatment process in order to achieve efficient conversion of A. donax to ethanol.

Introduction

Arundo donax, also known as giant reed, is a perennial herbaceous crop native to East Asia that has shown to thrive in Southern Europe. In a recent study the growth of A. donax was followed in a 12 years long-term field experiment, which showed that the high biomass yield and productivity makes A. donax a promising energy crop for bioethanol production [1], [2], [3]. Lewandowski et al. [4] have reviewed the literature of rhizomatous grasses as energy crops and stated that yields of 3–37 tons DM ha⁻¹ a⁻¹ have been reported for A. donax compared to 5–23 tons DM ha⁻¹ a⁻¹ for switchgrass. Perennial herbaceous energy crops like A. donax have the advantage of not requiring annual reseeding, need less energy input than annual crop land and reduce soil erosion [5]. A number of different lignocellulosic materials have been tested for their potential as feedstock for bioethanol production [6]. Lignocellulosic substrates are commonly classified as hardwood, softwood, agricultural residues, energy crops, weedy materials and municipal solid waste, and the choice of raw material is highly dependent on geographic location [5], [7]. Although A. donax has been discussed as energy crop for some time, there are few studies reporting assessment of the material for lignocellulosic ethanol production.

The high amount of xylan in herbaceous biomass such as A. donax is an issue. If the xylan is to be utilized for ethanol production, wild-type Saccharomyces cerevisiae cannot be used since it lacks a metabolic pathway for xylose utilization. Considerable efforts have been devoted to solve this issue by different metabolic engineering approaches (reviewed in e.g. [8]). The two major xylose utilizing pathways that have been incorporated into S. cerevisiae are (1) two fungal enzymes, xylose reductase (XR) and xylitol dehydrogenase (XDH), and (2) one enzyme from Piromyces spp., xylose isomerase (XI) [9], [10].

Separate hydrolysis and fermentation (SHF) and simultaneous saccharification and fermentation (SSF) are two principal process configurations for the production of bioethanol from lignocellulosic biomass. The choice of configuration will be determined by a balance of advantages and drawbacks associated with the two concepts for the feedstock in question which motivates why both process concepts were evaluated in the present study [7]. In SHF, enzymatic hydrolysis and fermentation are carried out in separate steps. This makes it possible to run each process under its optimum conditions, although end-product inhibition of the cellulolytic enzymes is a limiting factor [11]. Moreover, SHF offers the possibility of cell recycling [12], whereas in SSF it is not possible to separate cells and solid raw material particles.

In SSF, enzymatic hydrolysis of the pretreated raw material and fermentation are run in the same reaction vessel, which allow the released sugars from the hydrolysis to be rapidly consumed by the microorganism, thereby minimizing end-product inhibition of the cellulolytic enzymes. The lower capital costs together with lower process time for SSF also makes this configuration more attractive [13]. Furthermore, by using the complete slurry without separation of the fiber and liquid part one will minimize the loss (or dilution) of sugars associated with this separation, as pointed out in the original SSF patent from 1976 [14]. One disadvantage with this method is that the conditions for the enzymatic hydrolysis and the fermentation have to be the same, thus typically suboptimal for both of them. However, SSF has proven to be a more efficient strategy than SHF for several raw materials, e.g. spruce, wheat straw and corn stover [11], [15], [16], [17]. Importantly, SSF has proven to be advantageous when it comes to co-fermentation of glucose and xylose by recombinant S. cerevisiae, due to beneficial ratios between the two sugar concentrations enhancing the xylose conversion [18].

During the pretreatment, degradation products from the lignin and the monosaccharides are formed, which are inhibitory for both microbial metabolism and the enzymatic hydrolysis [19], [20]. The furans 2-furaldehyde (furfural) and 5-hydroxymethyl-2-furaldehyde (HMF) have received a lot of attention during the last years since they have been found to decrease the fermentation rate and extend the lag-phase of S. cerevisiae [21]. In the hydrolysis of hemicellulose to soluble oligo- and monosaccharides, which takes place during the pretreatment, acetyl groups on the hemicellulose are cleaved off, releasing acetic acid into the pretreatment liquid. Acetic acid is another compound with inhibitory properties shown to impair xylose utilization in a recombinant S. cerevisiae strain [22], [23]. Cellulolytic enzymes are inhibited by their end-products, glucose and cellobiose, which can be problematic for enzymatic hydrolysis of lignocellulosic materials. Recently it was shown that oligomers formed from degradation of xylan can inhibit the enzymatic hydrolysis [24]. This can potentially be a problem when using A. donax as feedstock since it contains a relatively high amount of xylan. Phenolic compounds originating from lignin have also shown to be inhibitory for microbial metabolism and cellulolytic enzymes [22], [25].

In the present study, pretreated A. donax was investigated for its suitability as feedstock for bioethanol production. Both SHF and SSF process options were evaluated, and since A. donax has high xylan content, a xylose-fermenting strain of S. cerevisiae expressing the genes for XR and XDH was used. As the conditions during the pretreatment can have a significant effect on the content of inhibitors in the final material, the effects of soluble compounds in the raw material on enzymatic hydrolysis and fermentation were studied in more detail. Here we present to the best of our knowledge, the first thorough investigation of A. donax as feedstock for bioethanol production in a comparative bioreactor based study.