Hydrolysis and fermentation of sweetpotatoes for production of fermentable sugars and ethanol

Industrial Crops and Products 42 (2013) 527–537

Contents lists available at SciVerse ScienceDirect

Industrial Crops and Products journal homepage: www.elsevier.com/locate/indcrop

Hydrolysis and fermentation of sweetpotatoes for production of fermentable sugars and ethanol William H. Duvernay a , Mari S. Chinn a,∗ , G. Craig Yencho b a b

Department of Biological and Agricultural Engineering, North Carolina State University, Weaver Labs, Campus Box 7625, Raleigh, NC 27695-7625, USA Department of Horticultural Science, Room 214-A Kilgore Hall, Box 7609, 2721 Founders Drive, North Carolina State University, Raleigh, NC 27695-7609, USA

a r t i c l e

i n f o

Article history: Received 13 March 2012 Received in revised form 9 June 2012 Accepted 19 June 2012 Keywords: Industrial sweetpotatoes Enzymes Hydrolysis Fermentation Glucose Ethanol

a b s t r a c t Liquefaction, saccharification, and fermentation of FTA-94 industrial sweetpotatoes (ISPs) were examined using ␣-amylase and glucoamylase for the production of ethanol. Starch degradation and sugars produced over time were examined for (1) ␣-amylase (Liquozyme SC) at different loading rates (0.045, 0.45, and 4.5% KNU-S/g dry ISP) during liquefaction; and (2) three glucoamylases (Spirizyme Fuel, Spirizyme Plus Tech, and Spirizyme Ultra) at different loading rates (0.5, 1.0, and 5.0 AGU/g dry ISP) during saccharification. The majority of starch, 47.7 and 65.4% of dry matter, was converted during liquefaction of flour and fresh sweetpotato preparations, respectively, with the addition of 0.45 KNU-S/g dry ISP of Liquozyme SC after 2 h (66.4 and 80.1% initial starch in dry matter, respectively). The enzymes used during saccharification increased starch breakdown, but was more effective in conversion of short chain carbohydrates to fermentable sugars. The addition of 5.0 AGU/g of Spirizyme Ultra after 48 h produced 795.4 and 685.3 mg glucose/g starch with flour and fresh preparations, respectively. Yeast fermentation on hydrolyzed starch was examined over time with and without the addition of salt nutrients. Yeast converted all fermentable sugar (e.g. glucose, fructose, maltose) and produced 62.6 and 33.6 g ethanol/L of hydrolysate for flour (25% w/v, substrate loading) and fresh (12.5% w/v, substrate loading) ISP, respectively, after 48 h without salt addition. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Corn is the predominant feedstock used to produce ethanol in the United States. Production continues to increase and approximately 10.8 billion gallons were produced from corn in 2009 (EERE, 2010). To diversify and improve biofuel production while reducing U.S. dependence on corn as a sole fuel ethanol source, significant efforts to convert lignocellulosic biomass as a potential feedstock for the production of ethanol have been implemented. While research and development of lignocellulosic biomass conversion processes is essential to U.S. progress toward replacing fossil fuels with renewable energy, emphasis is also being placed on finding alternate and abundant starch-based biofuel sources that do not directly compete with food and feed production. Sweetpotatoes (Ipomoea batatas, Morning-glory family), are an important staple crop produced globally in terms of total biomass. During 2010, an estimated 106,569,572 MT of sweetpotatoes

∗ Corresponding author at: Department of Biological and Agricultural Engineering, North Carolina State University, 277 Weaver Labs, Raleigh, NC 27695-7625, USA. Tel.: +1 919 515 6744; fax: +1 919 515 6719. E-mail addresses: [email protected] (W.H. Duvernay), [email protected] (M.S. Chinn), craig [email protected] (G.C. Yencho). 0926-6690/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.indcrop.2012.06.028

valued at $5.07 billion international dollars were produced globally (FAOUN, 2012). China is the dominant producer of sweetpotatoes, cultivating roughly 80% of the world’s crop, while Uganda and Nigeria produce roughly 3% of the global crop and rank 2nd and 3rd respectively. In the United States sweetpotatoes are considered a specialty crop and the nation currently ranks 8th in terms of total global production and crop value (FAOUN, 2012). Several high starch “industrial” sweetpotatoes (ISP), have been developed by the sweetpotato breeding and genetics program at NC State. These breeding lines have been selected for their biomass potential in replicated experiment station and on-farm yield trials. University and USDA-ARS studies have projected that sweetpotatoes can produce ethanol yields of 4500–6500 L/ha (500–700 gallons/acre), compared to (2800–3800 L/ha (300–400 gallons/acre) for corn (Hall and Smittle, 1983; Mays et al., 1990; Ziska et al., 2009). Therefore sweetpotatoes may offer an alternative material that can be converted to sugars needed for ethanol production and/or sugar-based value added chemicals. Sweetpotato is a very drought resistant crop requiring low chemical and fertilizer inputs and they can be grown in marginal soils, providing additional use benefit (George et al., 2011). Currently, they are predominantly grown in the Southeastern region of the United States. In 2007, 100,600 acres of sweetpotatoes were planted in the United States, with over 82% of production located in North Carolina, Louisiana, Mississippi, and

528

W.H. Duvernay et al. / Industrial Crops and Products 42 (2013) 527–537

Alabama alone (NC SweetPotato Commission, 2008). In the U.S., industrial sweetpotatoes are not intended for use as a food crop (typically not sweet in taste) and have been bred for higher dry matter content to generate up to 50% higher yields in starch than the common orange-fleshed sweetpotato (Nichols, 2007). These traits of ISPs offer potentially greater fermentable sugar yields from a sweetpotato crop for industrial conversion processes and the opportunity to increase planted acreage (even on marginal lands) beyond what is in place for food. Several technological advances are needed to establish a viable ISP-based bioproducts production system in the U.S. (George et al., 2011). Breeding and cultural management strategies for ISP production, which are very different from that of tablestock sweetpotatoes, are currently being researched to address these key field components. Our group and the work presented here are focused on examining industrial starch processing elements important to the economic production and bioconversion of ISP’s into fuels and chemicals. Starch conversion practices have involved three primary approaches to generating fermentable sugars: acid hydrolysis, temperature and pressure extrusion, and enzymatic hydrolysis. Acid hydrolysis offers an effective means to convert starch; however recovery of the acid is not efficient and as a result must be disposed of as wastewater (Farone and Cuzens, 1996). Another problem associated with acid hydrolysis is the production of hydroxymethylfurfurals (HMF), a dehydration product of d-glucose, which is known to inhibit growth and alcohol fermentation by yeast (Kim and Hamdy, 1985). High pressure extrusion significantly stabilizes conversion against temperature-induced inactivation. However, in a study done on the activity of ␣-amylase from barley malt at different temperatures, increasing pressure hindered the catalytic activity of ␣-amylase and a deceleration of the conversion rate was detected at all investigated temperatures (Buckow et al., 2007). Enzymatic hydrolysis of starch is biologically based, and can offer a simpler, more environmentally friendly method to produce useful sugars and ethanol. The process for enzymatic hydrolysis of corn starch has been optimized and is widely used for bioethanol production in the US. However, the starch in sweetpotatoes is considered more complex than cereal starches (corn, rice), containing 19.1% amylose in total starch (Collado et al., 1999). Root and tuber starches also contain longer and more complex amylopectin chains, making it more challenging to hydrolyze into fermentable sugars (Hoover, 2001; Srichuwong et al., 2005). In addition, structure and composition differences related to fatty acid and protein content influence the extent of starch hydrolysis observed using enzymes for various starch sources (Zhang and Oates, 1999; Wu et al., 2006). Much of the previous research on sugar production from sweetpotatoes has been limited to enhancing the sweetness and taste as a food commodity (Vanden et al., 1986; Koehler and Kays, 1991; Ridley et al., 2005) and a few studies using acid-based methods (Azhar and Hamdy, 1981; Kim and Hamdy, 1985; Shin et al., 2004). Some studies have also looked at enzymatic conversion for viscosity reduction during liquefaction and simultaneous saccharification and fermentation (SSF) for ethanol production (Zhang et al., 2010; Srichuwong et al., 2012). Yet there is still a need to establish a more defined biologically based approach to sweetpotato starch conversion and evaluate the enzymes and processing conditions suitable for effective fermentable sugar production. This study investigated processing parameters of liquefaction, saccharification, and fermentation of the ISP clone FTA-94, a high starch sweetpotato line bred for the production of glucose feedstocks and ethanol. The effects of ␣-amylase and glucoamylase loading rates, glucoamylase source and incubation time on extent of starch hydrolysis and sugar production (maltotriose, maltose and glucose) from fresh and flour preparations of the FTA-94 ISPs were evaluated. The most

promising treatment combinations were implemented to determine the need for supplemental nutrients to support yeast fermentation of ISP hydrolysates and potential ethanol production. 2. Materials and methods 2.1. Enzymes Two classes of amylase enzymes were used to convert starch into its simple sugars, ␣-amylase and glucoamylase. The enzyme ␣-amylase acts on the inner portions of the two main polymers of starch, amylose (␣-1,4 bonds) and amylopectin (␣-1,6 bonds), to form glucose chains of varying length, while glucoamylase produces free glucose by cleaving both ␣-1,4 and ␣-1,6 glycosidic bonds on longer chains of amylose and amylopectin (van der Maarel et al., 2002). The ␣-amylase used for liquefaction of the starchcontaining mashes was Liquozyme SC (Novozymes North America, Inc., Franklinton, NC, stored at 4 ◦ C, density 1.25 g/mL). It has an optimal pH of 5.5 with a broad pH tolerance, an optimal temperature of 85 ◦ C, and an activity of 120 KNU-S/g enzyme. One KNU-S/g (kilo novo unit) of ␣-amylase is the amount of enzyme that breaks down 5.26 g of starch per hour. Three different glucoamylase enzymes were chosen for saccharification studies: Spirizyme Fuel (750 AGU/g protein), Spirizyme Plus Tech (400 AGU/g protein), and Spirizyme Ultra (900 AGU/g protein) (Novozymes North America, Inc., Franklinton, NC). All three enzymes have an optimal temperature of 65 ◦ C, approximate density of 1.15 g/mL, and an amyloglucosidase unit of activity (AGU), representing the amount of enzyme able to hydrolyze 1 ␮mol of maltose per minute at 37 ◦ C and a pH of 4.3. 2.2. Industrial sweetpotato preparation FTA-94 ISPs were grown on the Horticultural Crops Research Station in Clinton, NC (35.023◦ N; 78.278◦ W). They were planted on May 23, 2007, and harvested on October 11, 2007. Two sweetpotato preparations were examined, fresh and flour (dry). Sweetpotatoes that were used for fresh preparations were stored at 14.5 ◦ C until use. They were then washed, diced into chunks, and ground in a food processor to produce 3–5 mm pieces just prior to use in experiments. Sweetpotatoes were washed, cut into slices, and dried at 70 ◦ C for 60 h for flour preparations. The slices were then broken into smaller pieces and ground through a 3 mm screen using a Wiley mill. Wet basis moisture content was determined for diced and oven-dried roots using an oven drying method (105 ◦ C, 24 h). 2.3. Experimental design and statistical analysis For liquefaction, the effects of incubation time, substrate preparation (fresh and flour ISP), and ␣-amylase loading rate on the change in starch content and the change in reducing sugars were investigated over time. Two sets of controls were prepared (1) sweetpotato only, no enzyme addition and (2) sweetpotato, enzyme addition and no incubation (to determine initial sweetpotato condition). For saccharification, the effects of incubation time, substrate preparation (fresh and flour ISP), type of glucoamylase and enzyme loading rate on the change in starch content and sugar production (maltotriose, maltose, and glucose) were investigated over time. Controls that contained no added enzyme and were not incubated at hydrolysis temperatures were prepared to determine initial sweetpotato condition. Liquefaction only controls were also prepared to represent the initial condition for saccharification. All treatment combinations and controls were completed in triplicate. Main and interaction effects of experimental factors were evaluated using analysis of variance (ANOVA) of the General Linear

W.H. Duvernay et al. / Industrial Crops and Products 42 (2013) 527–537

Model (GLM) in SAS (version 9.3.1, SAS Inc., Cary, NC). Pairwise ttest comparisons were performed to determine if each treatment effect was significantly different. Assessments of statistical significance were made at ˛ = 0.05. 2.4. Liquefaction Flour liquefaction preparations contained 3 g of dried ISP and 12 mL of distilled deionized water (ddH2 O), resulting in a 25% (w/v) ratio in 50 mL Falcon tubes. Fresh liquefaction preparations contained 10.3 g of fresh ISP (3 g dry, initial moisture content 70.9% wet-basis) and 17 mL of ddH2 O in Falcon tubes (50 mL), resulting in a 12.5% (w/v) ratio. A 12.5% (w/v) ratio was used for fresh preparation to allow for full substrate coverage, sufficient supernatant collection after liquefaction and minimize viscosity issues. Liquozyme SC was added at levels of 0, 0.03, 0.3, and 3.0% volume of enzyme/g dry ISP (0, 0.045, 0.45 and 4.5 KNU-S/g dry ISP, respectively). Samples were incubated at 85 ◦ C for 3 h and sampled destructively at 0, 1, 2, and 3 h. At each sampling interval, treatments were centrifuged (4 ◦ C, 2731 × g, 15 min), the pH of each was recorded and 2 mL of supernatant was removed and stored at −80 ◦ C until reducing sugar analysis. The remaining liquid from the liquefied slurry was decanted and the residual solid was resuspended and washed twice as described in Bridgers et al. (2010) prior to immediate analysis for starch content as alcohol insoluble solids (AIS). 2.5. Saccharification Considering results from liquefaction experiments, Liquozyme SC was added at a level of 0.3% volume enzyme/g dry ISP to both flour (3 g dry; 25%, w/v) and fresh (9.6 g fresh–3 g dry; initial moisture content 68.8% wet basis; 12.5%, w/v) preparations in 50 mL Falcon tubes and incubated at 85 ◦ C for 2 h. The glucoamylase enzymes (Spirizyme Plus Tech, Spirizyme Fuel, and Spirizyme Ultra) were added to appropriate treatments at levels of 0.5, 1.0, and 5.0 AGU/g dry ISP, based on their specific amyloglucosidase activities. Sodium azide was then added at a loading rate of 0.2% (w/v) to prevent contamination. Treatment combinations were incubated at 65 ◦ C for 0, 24, 48, and 72 h. At 24 h intervals, treatments were destructively sampled by centrifugation (4 ◦ C, 2731 × g, 15 min), pH was recorded, and 2 mL of supernatant was removed and stored at −80 ◦ C prior to sugar analysis. The remaining liquid from the saccharified slurry was decanted, and the residual solid was washed twice as described by Bridgers et al. (2010) and saved for subsequent starch (AIS) quantification. 2.6. Starch quantification Quantification of initial starch and the remaining solid fraction after liquefaction and saccharification was determined using a modified Alcohol Insoluble Solids (AIS) method (Ridley et al., 2005; Duvernay, 2008). Protein and fibrous fractions of the sweetpotato dry matter were not measured so final results report the change in starch content as a fraction of total dry matter, assuming the enzymes are not degrading protein and fiber (difference between initial and final AIS values). 2.7. Fermentation Fermentation of the ISP sugar hydrolysates was completed to study the need for supplemental nutrient addition and measure sugar consumption/ethanol production over time. Using the most promising conditions found from analysis of both liquefaction and saccharification experiments, 0.3% volume enzyme/g dry ISP of Liquozyme SC was added to both flour (5 g dry ISP; 25%, w/v) and

529

fresh (14.91 g fresh–5 g dry ISP; initial moisture content 66.5% wet basis; 12.5%, w/v) preparations (exposed to UV light, 30 min) in sterile Falcon tubes (50 mL) and incubated for 2 h at 85 ◦ C, followed by saccharification with Spirizyme Ultra for 48 h (5.0 AGU/g dry ISP, 65 ◦ C, added aseptically). Sodium azide was not added to minimize the possibility of growth inhibition of yeast during fermentation. Saccharification tubes were centrifuged (4 ◦ C, 2731 × g, 15 min) and 9 mL of supernatant was aseptically removed and added to Borosilicate glass disposable culture tubes (18 mm × 150 mm, Fisherbrand, Fisher Scientific, Chicago, IL). Culture tubes with hydrolysates were autoclaved (121 ◦ C, 15 psig, 15 min). Depending on the treatment (salt, no salt), either 1 mL sterile ddH2 O or 1 mL of sterile salt solution was added to the tubes in triplicate. The salt solution contained (per L): 1.32 g NH4 Cl, 0.11 g MgSO4 , 0.069 g CaCl2 , and 8.5 g yeast extract. Tubes were then inoculated with dried yeast (Ethanol Red Yeast, Saccharomyces cerevisiae, Lesaffre) at a concentration of 0.1% (w/v). Controls without yeast were completed in duplicate for each ISP preparation. Culture tubes were capped and incubated at 35 ◦ C and sampled (0.5 mL aliquots) every 24 h for a period of 72 h. Samples were stored at −80 ◦ C until HPLC analysis.

2.8. Sugar analysis Reducing sugars for liquefaction samples were quantified using a modified colorimetric dinitrosalicylic acid assay method (Chinn, 2003). Dilutions were made for all samples, including controls, based on preliminary trials. One milliliter of 0.1 M sodium acetate buffer (pH 5.0), 3 mL of DNS reagent, and 0.5 mL of each sample were combined and boiled in a water bath (100 ◦ C) for 5 min. Glucose calibration standards (0–2 g/L) were also completed to estimate reducing sugars as glucose equivalents. Absorbances were read at 540 nm using a Pharma Spec UV-1700 UV–Visible Spectrophotometer (Shimadzu Corporation, Kyoto, Japan). Absorbance values for samples were converted into milligrams of reducing sugars per gram of dry ISP. Individual sugars (maltotriose, maltose, glucose, fructose) and ethanol concentrations for samples were quantified by high performance liquid chromatography as described by Bridgers et al. (2010).

3. Results 3.1. Liquefaction—starch quantification Alcohol insoluble solids (AIS) for both fresh and flour ISP preparations represent the decrease in starch percentage relative to the initial starch dry matter fraction. The main effects of incubation time (hour) and enzyme loading (enzyme) on AIS for flour preparations after liquefaction were statistically significant (Pvalue < 0.05), while the interaction did not show a significant effect. The change in starch content of flour preparation treatments during liquefaction increased with increased enzyme loading and ranged from 23.3% to 52.7% loss in dry matter as starch (Fig. 1a). The average initial starch content, however, was 66.4%, leaving at least 14% of starch not converted. Incubation for 2 h produced statistically higher changes in starch than 1 h, resulting in the most effective change among all enzyme loadings (42.64%, P < 0.05). Three hours of incubation time produced slightly higher changes in starch content, but the difference on average was not statistically greater than 2 h treatments. The main effect of enzyme loading suggested that on average across all incubation times, each increase in enzyme loading was significantly different than the other, with a loading of 3.0% volume enzyme/g dry ISP of Liquozyme SC producing the highest average starch change of 51.22% (P < 0.05).

530

W.H. Duvernay et al. / Industrial Crops and Products 42 (2013) 527–537

Fig. 1. Average change in starch (% of dry matter lost) over time during liquefaction of flour and fresh FTA-94 industrial sweetpotatoes using Liquozyme SC: (A) flour preparations and (B) fresh preparations. Error bars represent one standard deviation of the experimental replicates.

Similar to flour preparations, the main effects of incubation time (hour) and enzyme loading (enzyme) on AIS of fresh sweetpotato preparations after liquefaction were statistically significant (P < 0.05), while the interaction was not statistically significant. Fresh sweetpotato liquefaction treatments resulted in starch content changes that ranged from 48.9% to 68.3% loss in dry matter as starch (Fig. 1b). The initial starch percentage was considerably higher for fresh sweetpotato preparations than flour at an average of 80.1% starch in dry matter. However, similar to results observed for flour, the fresh treatments did not convert all starch to sugars, leaving at least 13% of starch unhydrolyzed. On average across all enzyme loading rates, 3 h of incubation statistically produced the highest changes in starch percentage (63.86%; P < 0.05). Considering the average of each enzyme loading across all incubation times, the 3.0% volume enzyme/g dry ISP Liquozyme SC loading resulted

in the greatest change in starch but was not statistically different than 0.3% volume enzyme/g dry ISP (64.4% dry matter loss as starch; P > 0.05). For starch conversion, fresh samples performed considerably better than flour. The interaction of hour, enzyme, and preparation of the FTA-94 ISP showed that the flour preparation treatment (3.0% Liquozyme SC, 2 h incubation) with the highest change in starch content (51.2%) was significantly less than all fresh treatments excluding those with no enzyme added. Fresh preparations at the same loading rate and incubation time converted 14.1% more of the starch dry matter than flour samples. Differences in starch degradation may have been related to the initial starch contents. The average initial starch content for both flour and fresh were 66.6 and 81.7%, respectively, making more starch available in all fresh samples.

W.H. Duvernay et al. / Industrial Crops and Products 42 (2013) 527–537

531

Fig. 2. Average change in reducing sugars over time during liquefaction of flour FTA-94 industrial sweetpotato preparations using different Liquozyme SC loadings. Error bars represent one standard deviation of the experimental replicates.

3.2. Liquefaction—reducing sugars Reducing sugars of various sizes are formed when starch is converted to fermentable sugars. Typically, higher reducing sugar levels indicate a higher proportion of smaller sugar molecules in the mixture, which are easier to ferment. Reducing sugars for both were measured as a change in milligrams of sugars per gram of dry ISP, representing the increase in sugars from time zero to each respective incubation time. The analysis of variance for reducing sugars generated from flour treatment combinations during liquefaction showed significant main effects of incubation time (hour) and enzyme loading (P < 0.05). The flour treatments produced changes in reducing sugars ranging from 18.5 to 160.8 mg/g dry ISP (Fig. 2). The main effect of incubation time averaged across all enzyme loadings did not show a statistically significant change in reducing sugars after 2 h of incubation. Incubation for 3 h, however, was significantly greater than hour 1 (P < 0.05). Observing the main effect of enzyme loading showed that the 0.3% volume enzyme/g dry ISP loading produced the most reducing sugars, with significantly higher values than both 0.0 and 0.03% (v/w) treatments (P < 0.05). Results for change in reducing sugars from fresh sweetpotato preparations during liquefaction ranged from 120 to 287.1 mg/g dry ISP, and were significant only for the main effect of enzyme loading (P < 0.05). Statistically, a loading of 0.3% volume enzyme/g dry ISP produced higher values (235 mg/g dry ISP) than other treatments and controls, except the 3.0% (v/w) loading. Reducing sugars in fresh preparation treatments were statistically higher for every enzyme loading and incubation time than flour preparations. This is possibly due to structural changes in the flour ISP starch as a result of drying, which would reduce binding sites available for enzyme activity (Bondaruk et al., 2007). 3.3. Saccharification—starch quantification Changes in starch content during saccharification studies were quantified using alcohol insoluble solids and represent the percent of starch dry matter that was hydrolyzed to soluble sugars. Both fresh and flour ISP saccharification treatments showed slight improvements in the amount of starch converted into sugars beyond conversion during liquefaction. The main effects of both

incubation time (hour) and glucoamylase source (enzyme) and the interaction of hour and enzyme were statistically significant (P < 0.05) for AIS in saccharification flour preparations. Change in starch as a percent of dry matter lost for flour treatments during saccharification ranged from 48.6 to 53.3% (Fig. 3a). The initial starch content measured for saccharification treatments were on average 64.3%, indicating that after liquefaction approximately 11% of insoluble starch remained to be converted. The change in starch dry matter was statistically significant for each enzyme and all times beyond time zero (P < 0.05). The highest change in starch content for Spirizyme Fuel occurred at 72 h (53.1%), while Spirizyme Plus Tech achieved its highest change in starch at 48 h (51.7%). Spirizyme Ultra produced statistically similar changes over time (∼52%). The main effects of incubation time (hour), glucoamylase source (enzyme), and enzyme load for saccharification of fresh sweetpotato preparations were statistically significant (P < 0.05) as well as the interaction of hour and enzyme (P < 0.05). Fresh sweetpotato treatments saccharified with glucoamylase produced a broader range of starch content changes from 53.3 to 61.8% than flour sweetpotato treatments (Fig. 3b). Within each glucoamylase source, incubation for 24 h converted up to 5.3% more starch beyond liquefaction. For fresh treatment preparations the average initial starch content was 72.3%. This amount of initial starch was considerably less than the starch content observed during previous liquefaction studies completed on fresh ISPs, and influenced the amount of starch that was converted as measured by AIS, where lower initial starch values resulted in smaller changes in starch content (% of dry matter). Spirizyme Fuel, which had the lowest initial starch content, produced statistically lower changes in starch content (53.8–56.5% loss of dry matter) than Spirizyme Plus Tech and Spirizyme Ultra at all incubation times (55.0–61.4% loss of dry matter; P < 0.05; Fig. 3b). Spirizyme Plus Tech produced 1.4–2.3% higher changes in % dry matter than Spirizyme Ultra for all incubation times except time zero (P < 0.05). 3.4. Saccharification—soluble sugars Soluble sugars (maltotriose, maltose, glucose) resulting from saccharification of ISP flour preparations over time are presented

532

W.H. Duvernay et al. / Industrial Crops and Products 42 (2013) 527–537

Fig. 3. Average change in starch (% of dry matter lost) for the interaction of time and glucoamylase enzyme source across loading during saccharification of flour and fresh FTA-94 industrial sweetpotatoes: (A) flour preparations and (B) fresh preparations. Measured initial starch content (ISC) of the ISP preparations are indicated by the dashed line. Errors bars represent the standard error associated with each mean value.

Fig. 4. Concentration of HPLC sugars during saccharification of flour FTA-94 industrial sweetpotatoes over time using three different glucoamylase enzyme sources and loadings: (A) Spirizyme Fuel, (B) Spirizyme Plus Tech and (C) Spirizyme Ultra. Error bars represent one standard deviation of the experimental replicates.

in Fig. 4. An ideal conversion process would successfully convert all starch to simple glucose units for subsequent fermentation processing. For the analysis of variance of glucose, maltose and maltotriose produced from flour during saccharification, all main and interaction effects of incubation time (hour), glucoamylase source (enzyme) and glucoamylase loading (load) were statistically significant for glucose, maltose and maltotriose (P < 0.05), except for the main effect of hour for maltose conversion (P = 0.6835). Glucose production during saccharification of flour ISP treatments ranged from 80.0 to 549.6 mg/g dry ISP (123.6–865.5 mg/g starch) with the addition of glucoamylase. Without the addition of glucoamylase (controls), levels of glucose remained below 50 mg/g dry ISP throughout the 72 h incubation (data not shown). Spirizyme Fuel at all loadings converted flour-based starch to glucose with concentrations ranging from 80 to 493 mg/g dry ISP (Fig. 4a). When only 0.5 AGU/g dry ISP was added, glucose levels increased from 80.0 to 209.6 mg/g dry ISP as time increased from 0 to 72 h; however 48 h was not statistically different than 72 h of incubation (P > 0.05). The 1.0 AGU/g dry ISP loading increased glucose levels to 316.2 mg/g dry ISP after 48 h, but fell to 269.9 mg/g at 72 h of incubation. The 5.0 AGU/g dry ISP loading level produced its highest levels of glucose at 48 h (514 mg/g dry ISP; 794.2 mg/g starch), which was also highest among the Spirizyme Fuel treatment combinations (P < 0.05). Spirizyme Plus Tech produced glucose concentrations from flour preparations that ranged from 237.2 to 549.6 mg/g dry ISP (Fig. 4b). With the 0.5 and 1.0 AGU/g dry ISP loading rates, there was no statistical increase in glucose as incubation time increased from time zero. The 5.0 AGU/g dry ISP loading rate after 72 h of incubation produced a statistically higher level of glucose than any

other treatment combination of time and loading for Spirizyme Plus Tech with 549.6 mg/g dry ISP (865.5 mg/g starch; P < 0.05). Glucose released after saccharification with Spirizyme Ultra ranged from 116.0 to 513.8 mg/g dry ISP (Fig. 4c). For each enzyme loading, glucose levels increased with time from 0 to 48 h followed by a decrease after 72 h incubation. Glucose levels at hours 24 and 48 for 0.5 and 5.0 AGU/g enzyme loadings were not statistically different. On average, the most glucose produced by Spirizyme Ultra was 513.8 mg/g dry ISP (795.4 mg/g starch) and resulted after 48 h of incubation at an enzyme loading of 5.0 AGU/g dry ISP. Over the course of 72 h of ISP flour saccharification, maltose production values ranged from 0 to 275.9 mg/g dry ISP. At a loading rate of 5.0 AGU/g dry ISP, maltose concentrations increased up to 63.2 mg/g dry ISP for all three glucoamylases. This 5.0 AGU/g dry ISP loading rate produced maltose values that were significantly smaller than any other enzyme loading for each glucoamylase (P < 0.05), suggesting that the glucoamylase activity played an important role in digesting maltose into glucose. Maltotriose concentrations were very small for all treatment combinations. Nearly half of the values observed were zero, with others only reaching a high concentration of 23.3 mg/g dry ISP at time zero and a Spirizyme Plus Tech loading of 0.5 AGU/g dry ISP. The highest values of maltotriose at each incubation time were produced with no addition of glucoamylase. Conversion of fresh ISP starch into sugars using glucoamylase during saccharification was successful for all three enzymes (Fig. 5). However, the sugars were not broken down as extensively as with flour. The analysis of variance tables for glucose, maltose and maltotriose for fresh sweetpotato preparations for the main and

W.H. Duvernay et al. / Industrial Crops and Products 42 (2013) 527–537

533

and 48. Interestingly, the highest value of glucose for each type of glucoamylase for all enzyme loadings and incubation times was recorded at time zero with a loading of 5.0 AGU/g dry ISP (Fig. 5). Maltose concentration values during saccharification of fresh ISPs ranged from 0 to 377.2 mg/g dry ISP (Fig. 5). As enzyme loading increased for all three enzymes within each time step, the levels of maltose decreased significantly by as much as 140 mg/g dry ISP (P < 0.05). However, over time maltose levels increased for each enzyme loading and source treatment combination. For Spirizyme Ultra maltose values increased significantly from 0 to 24 h (P < 0.05). Levels of maltotriose for fresh ISP treatments remained low throughout the entire 72 h saccharification, increasing from 0 to only 28.7 mg/g dry ISP (Fig. 5). When glucoamylase was not added, values for maltotriose ranged from 26 to 32 mg/g dry ISP for all incubation times. Maltotriose concentrations, although all were not significant, decreased with increases in enzyme loading. 3.5. Fermentation

Fig. 5. Concentration of HPLC sugars during saccharification of fresh FTA-94 industrial sweetpotatoes over time using three different glucoamylase enzyme sources and laodings: (A) Spirizyme Fuel, (B) Spirizyme Plus Tech and (C) Spirizyme Ultra. Error bars represent one standard deviation of the experimental replicates.

interaction effects of incubation time (hour), glucoamylase source (enzyme) and glucoamylase loading (load) were statistically significant for glucose and maltose response variables (P < 0.05). The main and interaction effects of hour, enzyme and load were all statistically significant for maltotriose (P < 0.05) with the exception of the full three way interaction (P > 0.05). Glucose concentrations ranged from 70.5 to 398.5 mg/g dry ISP (102.8–537.1 mg/g starch), with the exception of one treatment which reached an average value of 508.6 mg/g dry ISP (685.3 mg/g starch; 5.0 AGU/g dry ISP Spirizyme Ultra, 0 h). Without the addition of glucoamylase, glucose levels ranged from 0 to only 47.4 mg/g dry ISP for all incubation times, with most of the values remaining below 30 mg/g dry ISP (data not shown). For most treatment combinations, glucose levels significantly increased as enzyme loading increased. Glucose values for enzyme loadings of 0.5 and 1.0 AGU/g dry ISP with Spirizyme Fuel were statistically similar while an enzyme loading of 5.0 AGU/g dry ISP produced a significantly higher level of glucose at all time points (P < 0.05). Glucose values for Spirizyme Fuel loadings of 5.0 AGU/g dry ISP were 384.6 mg/g dry ISP at time zero and ultimately decreased to 302.4 mg/g dry ISP at 72 h. For Spirizyme Plus Tech at time 0, glucose levels significantly increased as enzyme loading increased from 0 to 5.0 AGU/g dry ISP (P < 0.05). At 24 and 48 h of incubation, 0.5 and 1.0 AGU/g dry ISP of Spirizyme Plus Tech were statistically similar to each other (308.5 and 303 mg/g dry ISP, respectively), however 5.0 AGU/g dry ISP produced significantly higher glucose values. Spirizyme Ultra performed similar to Spirizyme Plus Tech, with significantly higher glucose values as enzyme loading increased at hours 0 and 72, and statistically similar values for loadings of 0.5 and 1.0 AGU/g dry ISP at hours 24

Results for glucose, ethanol and fructose concentrations over time during fermentation of flour and fresh ISPs are shown in Fig. 6. Flour-based ISP sugars fermented with and without salt supplementation produced similar changes in sugars and ethanol and the presence of salts did not significantly improve the fermentation. Samples without salts reached an average of 62.6 g ethanol/L hydrolysate after 48 h, and increased by 4.6% after 72 h. Samples with salts produced the peak value of 67.8 g ethanol/L hydrolysate after 48 h. Glucose values for samples without and with salt supplementation were 116.7 and 117.5 g/L, respectively, at the beginning of the fermentation period. After 48 h, levels of glucose significantly decreased to 7.1 g/L for samples without salts, and completely disappeared for samples with salt. Fructose increased up to 17.8 g/L on average after 24 h for fermentations with and without salts and was completely converted by hour 72. Fermentation of fresh ISP sugars with and without salt supplementation produced similar sugar and ethanol trends relative to flour treatments, where the presence of salts did not significantly improve the fermentation yields (Fig. 6b). Cultures produced the highest level of ethanol by 48 h of fermentation, 34.9 and 33.6 g/L, for salts and no salts respectively. Initial glucose levels for fresh ISP sugars were 61.2 and 63.1 g/L, respectively, for samples with and without salts and by hour 48 no glucose was detected. Fructose was not initially observed in the fresh ISP fermentations. However at 24 h, samples without salts had produced 6.4 g/L of fructose, and an average of 5.3 g/L was created in samples with salts as a result of yeast metabolism (Boulton and Quain, 2001). Fructose was not detected beyond hour 48. Differences in starting material loading rates for ISPs will affect sugar yields and ultimately final product concentrations in the broth. Although the initial sugar concentrations for fresh and flour preparations were different (due to difference in initial liquefaction and saccharification substrate loadings, 25% (w/v) flour, 12.5% (w/v) fresh), the yeast performance in terms of ethanol yields was similar between the preparations. The presence of salts seemed to help increase ethanol yield, however the ethanol production rate, 1.4 g/L/h for fresh samples and 0.7 g/L/h for flour samples, and overall production of ethanol were not significantly different, suggesting that the addition of salts during fermentation of ISP hydrolysates is not necessary. 4. Discussion The ISP’s evaluated in this project have higher dry matter and starch contents than conventional sweetpotatoes and have been bred specifically for biofuel and industrial chemical production, not

534

W.H. Duvernay et al. / Industrial Crops and Products 42 (2013) 527–537

Fig. 6. Ethanol, glucose, and fructose concentrations measured over time during fermentation of flour and fresh FTA-94 industrial sweetpotatoes: (A) Flour, 0.1% yeast, no salt, (B) Flour, 0.1% yeast, salt extract, (C) Fresh, 0.1% yeast, no salt and (D) Fresh, 0.1% yeast, salt extract. Error bars represent one standard deviation of the experimental replicates.

for food consumption. However, before ISP’s can be considered as a viable new industrial crop for the production of fermentable sugars and ethanol in the U.S., the fundamental work needed to define a bioprocess tailored to their unique characteristics is required. Two key issues relevant to this goal are the enzymatic hydrolysis of ISP-based starch into sugars and the conversion of these sugars into useful products such as ethanol. Our research was designed to address these needs. Based on AIS values for liquefaction and saccharification, flour ISP preparations had lower initial starch content compared to fresh ISP preparations. Flour sweetpotato preparations were dried at 70 ◦ C for 60 h to prevent decomposition of the sweetpotato during storage. Drying the sweetpotato may have contributed to loss of starch content among all flour preparations. A study performed on the effect of drying conditions on the quality of dried white potato cubes showed that initial starch contents of potato dried at 70 ◦ C had an average of 67% starch content (% dry basis) compared to approximately 75% starch content found for raw, unblanched potato cubes (Bondaruk et al., 2007). These results are comparable to an average of approximately 66% initial starch content found for flour ISP preparations during liquefaction and saccharification in this study. Bondaruk et al. (2007) also found that blanching the potato cubes before drying caused a smaller reduction in starch content (73% compared to 75% for raw potato). This suggests that in the future drying preceded by a blanching step may help improve the initial starch content of flour ISP preparations. If flour

preparations are desired, more research must be performed to determine the most effective approach for preserving starch. A two-step hydrolysis (liquefaction and saccharification) and analysis of both starch and sugars is important to defining an effective ISP starch conversion process. During the liquefaction process, the incubation temperature (85 ◦ C) initiated gelatinization of the sweetpotato starch, where excess water in the aqueous slurry causes radical swelling and loss of crystalline order in the starch granule. This destabilized the amylopectin crystallites and allowed the added ␣-amylase greater accessibility to starch polymers (Donovan, 1979; Stevens and Elton, 1981; Evans and Haismann, 1982; Biliaderis, 1991). The Liquozyme SC was effective at converting starch to soluble dextrin sugars, resulting in a 79.4 and 85.3% change in starch after liquefaction. The change in starch content (% of dry matter lost) was not largely improved after saccharification with glucoamylase relative to liquefaction. The glucoamylase did, however, dramatically increase the amount of soluble sugars, specifically glucose, available for fermentation by at least 5.6, 3.9, and 6.1 times more glucose on average for Spirizyme Fuel, Spirizyme Plus Tech and Spirizyme Ultra, respectively. This suggests that the ␣-amylase used during liquefaction affects the quantity of all of the insoluble starch, while the glucoamylase used during saccharification allows for the generation of smaller, more easily fermentable sugars. Quantification of individual sugars in addition to change in starch content gave a more complete representation of the extent of hydrolysis.

W.H. Duvernay et al. / Industrial Crops and Products 42 (2013) 527–537

Liquozyme SC loadings of 0 to 3.0% (v/w) were tested to determine whether increasing enzyme activity would result in corresponding increases in reducing sugar levels. The 0.3% (v/w) loading level proved to exhibit the most effective use of available enzyme to convert starch into sugars. The addition of 3.0% (v/w) of ␣-amylase did not result in the expected increase in starch loss and reducing sugars. This may have been due to saturation of available binding sites on inner portions of amylose and amylopectin by native amylases (e.g. ␤-amylase) and the addition of “excess” exogenous ␣-amylase, resulting in unused enzyme remaining dormant in the solution (Hageniman et al., 1992). In addition, sweetpotato starch contains high amounts of amylopectin containing bonds which are not digestible by ␣-amylase alone. The excess enzyme added during liquefaction would not have been effective in improving hydrolysis of ISP starch fractions containing these bonds (Zhang and Oates, 1999). The effects of increased loadings of glucoamylase during saccharification did not show indications of reduced enzyme activity due to saturation, as increases in glucoamylase up to 5 AGU/g dry ISP resulted in increased production of glucose. Enzyme loadings of glucoamylase used for saccharification were based on preliminary studies in our lab (data not shown). Loadings up to 0.75 AGU/g dry ISP in previous experiments did not significantly increase reducing sugars, so the experiments presented were designed to test higher loadings of glucoamylase. Considering enzyme cost and activity and processing time for the saccharification treatments using flour preparations, 5.0 AGU/g dry ISP of Spirizyme Ultra incubated for 48 h was selected as the most promising, producing 513.8 mg/g dry sweetpotato of glucose and converting approximately 80% of the starch to single glucose units (94% dextrose equivalent based on reducing power of glucose from maltotriose, maltose, and glucose concentrations). Azhar and Hamdy (1981) performed a study on the acid hydrolysis of sweetpotatoes and found a dextrose equivalent value of 83%, though only 1% (w/v) sweetpotato was used and extremely high levels of hydroxymethylfurfurals (HMFs) were detected. In comparison to acid hydrolysis methods for starch conversion, higher amounts of sugars can be recovered from starch using environmentally friendly commercial enzymes without limitations of HMFs. A treatment combination of 5.0 AGU/g dry ISP of Spirizyme Ultra incubated for 48 h also increased the number of reducing sugars by converting nearly all of the complex oligosaccharides to glucose and increased the total amount of composite sugars from 1.2 to 3.0 mmol/g dry sweetpotato over the liquefaction control (no glucoamylase added). A study on the saccharification of sweetpotato mashes using native amylases (mainly ␤-amylases) for 32 h of incubation found that at 70 ◦ C, 1.99% fructose, 3.45% glucose, and 18.37% maltose (% dry matter) were converted leaving 35.7% AIS (McArdle and Bouwkamp, 1986). Comparing these values to the over 51% glucose (% dry matter) produced using commercial ␣-amylase and glucoamylase in our study shows that although endogenous enzymes can effectively produce sugars from sweetpotato starch using only heat treatments, significantly higher amounts of fermentable sugars (with a greater extent of hydrolysis) can be obtained with the help of adding additional enzymes. Despite the successful conversion of starch to sugars during liquefaction and saccharification, at least 10% of starch remained unhydrolyzed for fresh and flour ISP preparations, suggesting inhibition effects during the hydrolysis process. The inability of glucoamylase to help complete the starch digestion can be partially attributed to competitive product inhibition caused by the high concentration of sugars produced during hydrolysis (Weselake and Hill, 1983; Robyt, 1984; Houng et al., 1992; Hill et al., 1997). In a study investigating the semi-continuous hydrolysis of raw sweetpotato starch by a crude fungal source of glucoamylase, results showed similar inhibitions of starch

535

digestion to those found in this work with increasing glucose concentrations, where glucose was produced at a high rate during the first 12 h of incubation (up to 4%, w/v, glucose). However, the rate decreased rapidly during the next 12 h, indicating an inhibition effect of glucose on starch digestion (Noda et al., 1992). Results on hydrolysis of sweetpotatoes from this study showed that during saccharification, glucose levels reached 13.9 and 6.7% (w/v) of glucose for flour and fresh preparations, respectively. Glucose concentrations observed were high enough to induce product inhibitions on glucoamylase activity (Noda et al., 1992). Aside from product inhibition, incomplete conversion may have occurred during saccharification due to the protein content of ISPs and their potential binding to amylose and amylopectin fractions, resulting in a limitation in binding sites and residual starch (Wu et al., 2006). Results of fresh ISP preparations during saccharification showed that for all three glucoamylases, samples taken after enzyme addition and before saccharification incubation (time 0) produced the highest glucose values of all incubation times. These values also significantly increased as glucoamylase increased to 5.0 AGU/g dry ISP. The glucoamylases quickly acted on the starch and sugars to produce glucose with high activity despite sample processing at 4 ◦ C (∼1 h prior to storage at −80 ◦ C). Results for flour preparations also showed significant increases in glucose levels at time 0 with increasing glucoamylase loadings, though as incubation time increased to 48 h, glucose levels continued to increase. The higher rate of hydrolysis for fresh treatments compared to flour could have been a result of the lower loading rate (12.5%, w/v), which provided a greater volume of liquid, and the natural water content of the ISP. The increased water content, presence of bound and absorbed water in the pore spaces of the sweetpotato and easily accessible starch granules (compared to more complex granules of flour treatments caused by heating during preparation) may have allowed for a faster conversion of starch to sugars (Bondaruk et al., 2007). Higher amounts of maltose were also observed for fresh treatments over the entire course of saccharification when compared to flour treatments. In the presence of high levels of glucose and high dissolved solids, excess glucoamylase loading has been known to catalyze the reverse condensation reaction. This causes maltose, isomaltose, and other byproducts to form from glucose (Crabb and Shetty, 1999; Roy and Gupta, 2004). These byproducts are referred to as ‘reversion products’ and are known to decrease overall glucose yield. The increase in maltose and decrease in glucose for fresh treatments may be related to this phenomenon as the highest glucose levels from starch and complex carbohydrates were observed nearly instantaneously.

5. Conclusion It has been shown that the use of commercial enzymes during liquefaction and saccharification of FTA-94 ISPs has a substantial effect on the conversion of starch to fermentable sugars. Provided that sugar is available for conversion as primarily glucose, it appears that yeast performance on ISP hydrolysates is independent of the initial preparation (e.g. fresh, flour). For hydrolysis steps, enzyme cost and activity and energy requirements for preprocessing should be considered when making choices about conversion treatments of fresh and flour ISP preparations. Differences in physical properties (e.g. bulk density, hygroscopic interactions with water, viscosity) between the two preparations can greatly impact material handling and final process operations as well, making the choice to use fresh or flour materials dependent on application, equipment size, facility space and storage capabilities. The most promising enzyme loadings for sugar production were

536

W.H. Duvernay et al. / Industrial Crops and Products 42 (2013) 527–537

0.45 KNU-S/g dry ISP (0.3%, w/v) of Liquozyme SC for liquefaction and 5.0 AGU/g of Spirizyme Ultra for saccharification. The majority of the starch was converted during liquefaction of flour and fresh ISP preparations with the addition of 0.45 KNU-S/g dry ISP of Liquozyme SC after 2 h of incubation, leaving at most 18.7% starch dry matter to be converted. Additionally, saccharification was able to increase the breakdown of starch; however its main function was converting the soluble dextrins resulting from liquefaction to fermentable sugars. The addition of 5.0 AGU/g of Spirizyme Ultra after 48 h of incubation was able to produce 862.2 and 743.9 mg of simple sugars/g of starch with flour and fresh preparations, respectively. The sugar yields observed from ISPs after liquefaction and saccharification are similar to yield values for corn and other cereal starches reported in the literature (Karakatsanis and LiakopoulouKyriakides, 1998). For corn-derived starch, 94–99% conversion has been achieved within 48 h using enzymatic methods (DettoriCampus et al., 1992; Arasaratnam and Balasubramaniam, 1993; Karakatsanis and Liakopoulou-Kyriakides, 1998; Mojovic et al., 2006). Differences observed can easily be attributed to composition of the starting material, including the type of plant (grain vs. storage root), substrate loading rates and experimental conditions (Mojovic et al., 2006). Fermentation with Ethanol Red Yeast was able to generate 62.6 and 33.6 g ethanol/L of hydrolysate for flour (25%, w/v, substrate loading) and fresh (12.5%, w/v, substrate loading) ISP, respectively, after 48 h of fermentation without salts, proving that industrial sweetpotatoes can offer a viable alternative starch source for industrial application of ethanol production. Acknowledgements The authors would like to thank Dr. Mike Boyette for the use of post-harvest processing equipment during experimentation and Novozymes North America, Inc. (Franklinton, NC) for providing the “first generation” enzymes used in this research effort. Mr. Kenneth V. Pecota of the sweetpotato breeding program and the staff of the North Carolina Department of Agriculture and Consumer Services Horticultural Crops Research Station, Clinton, NC are acknowledged for their assistance with the production of the sweetpotatoes used in these studies. References Arasaratnam, V., Balasubramaniam, K., 1993. Synergistic action of ␣-amylase and gluco-amylase on raw corn. Starch/Starke 45 (6), 231–233. Azhar, A., Hamdy, M.K., 1981. Alcohol fermentation of sweet-potato. 1. Acidhydrolysis and factors involved. Biotechnol. Bioeng. 23 (4), 879–886. Biliaderis, C.G., 1991. The structure and interactions of starch with food constituents. Can. J. Physiol. Pharmacol. 69 (1), 60–78. Bondaruk, J., Markowski, M., Blaszczak, W., 2007. Effect of drying conditions on the quality of vacuum-microwave dried potato cubes. J. Food Eng. 81 (2), 306–312. Boulton, C., Quain, D., 2001. Brewing Yeast and Fermentation. Blackwell Science Ltd, Iowa. Bridgers, E.N., Chinn, M.S., Truong, V.D., 2010. Extraction of anthocyanins from industrial purple-fleshed sweetpotatoes and enzymatic hydrolysis of residues for fermentable sugars. Ind. Crop Prod. 32 (3), 613–620. Buckow, R., Weiss, U., Heinz, V., Knorr, D., 2007. Stability and catalytic activity of alpha-amylase from barley malt at different pressure–temperature conditions. Biotechnol. Bioeng. 97 (1), 1–11. Chinn, M.S., 2003. Solid substrate cultivation of anaerobic thermophilic bacteria for the production of cellulolytic enzymes. Dissertation. Department of Biosystems and Agricultural Engineering, University of Kentucky, Lexington, KY. Collado, L.S., Mabesa, R.C., Corke, H., 1999. Genetic variation in the physical properties of sweet potato starch. J. Agric. Food Chem. 47 (10), 4195– 4201. Crabb, W.D., Shetty, J.K., 1999. Commodity scale production of sugars from starches. Curr. Opin. Microbiol. 2 (3), 252–256. Dettori-Campus, B.G., Priest, F.G., Stark, J.R., 1992. Hydrolysis of starch granules by the amylase from Bacillus stearothermophilus NCA 26. Process Biochem. 27, 17–21. Donovan, J.W., 1979. Phase-transitions of the starch–water system. Biopolymers 18 (2), 263–275.

Duvernay, W., 2008. Conversion of industrial sweetpotatoes for the production of ethanol. M.S. Thesis. North Carolina State University, Raleigh, NC, USA. Energy Efficiency and Renewable Energy (EERE), 2010. 2009 Renewable Energy Data Book. U.S. Department of Energy Biomass Program, Energy Efficiency and Renewable Energy, Washington, DC, Available from: [cited http://www1.eere.energy.gov/maps data/pdfs/eere databook.pdf on September 8, 2011]. Evans, J.D., Haismann, D.R., 1982. The effects of solutes on the gelatinization temperature of potato starch. Starke 34, 224–231. Farone, W., Cuzens, J., 1996. Method of producing sugars using strong acid hydrolysis of cellulosic and hemicellulosic materials. US Patent 5,562,777. Filed March 26, 1993. Issued October 8, 1996. Food and Agricultural Organization of the United Nations, 2012. Available from: http://faostat.fao.org/site/339/default.aspx [cited on March 06, 2012]. George, N.A., Pecota, K.V., Bowen, B.D., Schultheis, J.R., Yencho, G.C., 2011. Root piece planting in sweetpotato – a synthesis of previous research and directions for the future. HortTechnology 21, 703–711 (Reviews). Hageniman, V., Vezina, L.P., Simard, R.E., 1992. Distribution of amylases within sweet potato (Ipomoea batatas L.) root tissue. J. Agric. Food Chem. 40, 1777– 1783. Hall, M.R., Smittle, D.A., 1983. Industrial-Type Sweet Potatoes: A Renewable Energy Sources for Georgia. The University of Georgia, College of Agriculture Experimental Stations, p. 10. Hill, G.A., Macdonald, D.G., Lang, X., 1997. Alpha-amylase inhibition and inactivation in barley malt during cold starch hydrolysis. Biotechnol. Lett. 19 (11), 1139–1141. Hoover, R., 2001. Composition, molecular structure, and physicochemical properties of tuber and root starches: a review. Carbohydr. Polym. 45 (3), 253– 267. Houng, J.Y., Chiou, J.Y., Chen, K.C., 1992. Production of high maltose syrup using an ultrafiltration reactor. Bioprocess. Eng. 8 (1–2), 85–90. Karakatsanis, A., Liakopoulou-Kyriakides, M., 1998. Comparative study of hydrolysis of various starches by alpha-amylase and glucoamylase in PEG-dextran and PEG-substrate aqueous two phase systems. Starch/Starke 50 (8), 349– 353. Kim, K., Hamdy, M.K., 1985. Acid hydrolysis of sweet-potato for ethanol production. Biotechnol. Bioeng. 27 (3), 316–320. Koehler, P.E., Kays, S.J., 1991. Sweet-potato flavor – quantitative and qualitative assessment of optimum sweetness. J. Food Quality 14 (3), 241–249. Mays, D.A., et al., 1990. Fuel production potential of several agricultural crops. In: Janick, J., Simon, J.E. (Eds.), Advances in New Crops. Timber Press, Portland, OR, pp. 260–263. McArdle, R.N., Bouwkamp, J.C., 1986. Use of heat-treatments for saccharification of sweet-potato mashes. J. Food Sci. 51 (2), 364–366. Mojovic, L., Nikolic, S., Rakin, M., Vukasinovic, M., 2006. Production of bioethnaol from corn meal hydrolyzates. Fuel 85, 1750–1755. Nichols, K., 2007. NC State University Researchers Brewing Energy From Sweet Potatoes. NC State News Room. Available from: http://news.ncsu.edu/news/sweet%20potato%20release.php. Noda, T., Ohtani, T., Shiina, T., Nawa, Y., 1992. Semicontinuous hydrolysis of sweet-potato raw starch by chalara-paradoxa glucoamylase. J. Food Sci. 57 (6), 1348–1352. North Carolina SweetPotato Commission, 2008. National Agricultural Statistics Service. US Department of Agriculture. Benson, NC. Available from: http://www.ncsweetpotatoes.com/content/view/26/40/ [database online] [cited on July 18, 2008]. Ridley, S.C., Kim, M., Heenan, S., Bremer, P., 2005. Evaluation of sweet potato cultivars and heating methods for control of maltose production, viscosity and sensory quality. J. Food Quality 28 (2), 191–204. Robyt, J., 1984. Enzymes in the hydrolysis and synthesis of starch. In: Whistler, R., Beniller, J., Paschell, E. (Eds.), Starch: Chemistry and Technology. Academic Press, New York, NY. Roy, I., Gupta, M.N., 2004. Hydrolysis of starch by a mixture of glucoamylase and pullulanase entrapped individually in calcium alginate beads. Enzyme Microb. Technol. 34 (1), 26–32. Shin, S.I., Byun, J., Park, K.H., Moon, T.W., 2004. Effect of partial acid hydrolysis and heat-moisture treatment on formation of resistant tuber starch. Cereal Chem. 81 (2), 194–198. Srichuwong, S., Sunarti, T.C., Mishima, T., Isono, N., Hisamatsu, M., 2005. Starches from different botanical sources I: contribution of amylopectin fine structure to thermal properties and enzyme digestibility. Carbohydr. Polym. 60 (4), 529– 538. Srichuwong, S., Orikasa, T., Matsuki, J., Shiina, T., Kobayashi, T., Tokuyasu, K., 2012. Sweet potato having a low temperature-gelatinizing starch as a promising feedstock for bioethanol production. Biomass Bioenerg. 39, 120–127. Stevens, D.J., Elton, G.A., 1981. Thermal properties of the starch/water system I. Measurement of heat of gelatinization by differential scanning calorimetry. Starke 23, 8–11. Vanden, T., Biermann, C.J., Marlett, J.A., 1986. Simple sugars, oligosaccharides, and starch concentrations in raw and cooked sweet-potato. J. Agric. Food Chem. 34 (3), 421–425. van der Maarel, M., van der Veen, B., Uitdehaag, J.C.M., Leemhuis, H., Dijkhuizen, L., 2002. Properties and applications of starch-converting enzymes of the alphaamylase family. J. Biotechnol. 94 (2), 137–155. Weselake, R.J., Hill, R.D., 1983. Inhibition of alpha-amylase catalyzed starch granule hydrolysis by cycloheptaamylose. Cereal Chem. 60 (2), 98–101.

W.H. Duvernay et al. / Industrial Crops and Products 42 (2013) 527–537 Wu, X., Zhao, R., Wang, D., BEan, S.R., Seib, P.A., Tuinstra, M.R., Campbell, M., Brien, A.O., 2006. Effects of amylose, corn protein, and corn fiber contents on production of ethanol from starch rich media. Cereal Chem. 83 (5), 569–575. Zhang, T., Oates, C.G., 1999. Relationship between alpha-amylase degradation and physico-chemical properties of sweet potato starches. Food Chem. 65 (2), 157–163.

537

Zhang, L., Chen, Q., Jin, Y., Xue, H., Guan, J., Wang, Z., Zhao, H., 2010. Energy saving direct ethanol production from viscosity reduction mash of sweet potato at very high gravity (VHG). Fuel Process. Technol. 91 (12), 1845–1850. Ziska, L.H., Runion, G.B., Tomecek, M., Prior, S.A., Torbet, H.A., Sicher, R., 2009. An evaluation of cassava, sweetpotato and field corn as potential carbohydrate sources for bioethanol production in Alabama and Maryland. Biomass Bioenerg. 33, 1503–1508.