Bioresource Technology 101 (2010) 5444–5448

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier .com/locate /bior tech

An attempt towards simultaneous biobased solvent based extraction of proteinsand enzymatic saccharification of cellulosic materials from distiller’s grainsand solubles q

Saurav Datta a, B.D. Bals b, Yupo J. Lin a, M.C. Negri a, R. Datta c, L. Pasieta c, Sabeen F. Ahmad a,Akash A. Moradia a,d, B.E. Dale b, Seth W. Snyder a,*

a Process Technology Research, Energy Systems Division, Argonne National Laboratory, Argonne, IL 60439, USAb Biomass Conversion Research Laboratory, Department of Chemical Engineering and Materials Science, Michigan State University, East Lansing, MI, USAc Vertec BioSolvents Inc, Downers Grove, IL, USAd Illinois Mathematics and Science Academy, Aurora, IL, USA

a r t i c l e i n f o

Article history:Received 10 July 2009Received in revised form 22 January 2010Accepted 8 February 2010Available online 4 March 2010

Keywords:Distiller’s grains and solubles (DGS)Biobased solventsProtein extractionEnzymatic hydrolysisAFEX

0960-8524/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.biortech.2010.02.039

q The submitted manuscript has been createdOperator of Argonne National Laboratory (‘‘Argonne”of Energy Office of Science laboratory, is operated u06CH11357. The US Government retains for itself, andpaid-up nonexclusive, irrevocable worldwide licenseprepare derivative works, distribute copies to the pubdisplay publicly, by or on behalf of the Government.

* Corresponding author. Tel.: +1 630 252 7939; faxE-mail address: seth@anl.gov (S.W. Snyder).

a b s t r a c t

Distiller’s grains and solubles (DGS) is the major co-product of corn dry mill ethanol production, and iscomposed of 30% protein and 30–40% polysaccharides. We report a strategy for simultaneous extractionof protein with food-grade biobased solvents (ethyl lactate, D-limonene, and distilled methyl esters) andenzymatic saccharification of glucan in DGS. This approach would produce a high-value animal feedwhile simultaneously producing additional sugars for ethanol production. Preliminary experiments onprotein extraction resulted in recovery of 15–45% of the protein, with hydrophobic biobased solventsobtaining the best results. The integrated hydrolysis and extraction experiments showed that biobasedsolvent addition did not inhibit hydrolysis of the cellulose. However, only 25–33% of the total proteinwas extracted from DGS, and the extracted protein largely resided in the aqueous phase, not the solventphase. We hypothesize that the hydrophobic solvent could not access the proteins surrounded by theaqueous phase inside the fibrous structure of DGS due to poor mass transfer. Further process improve-ments are needed to overcome this obstacle.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Volatility in fuel and corn prices has adversely affected the eth-anol industry in recent years. In order to improve the economics ofcorn to ethanol production, enhancing the value of the co-productsis necessary. The US produces over 11 billion gallons of ethanol peryear, predominantly in corn dry mills. The main co-product of thecorn dry milling process, distiller’s grains (DG) or distiller’s grainsand solubles (DGS), contains significant amount of protein, poly-saccharide and fat (Kim et al., 2008a), and is therefore, primarilysold as an animal feed replacement for whole corn. A bushel ofcorn (56 lb) yields around 16 lb of DGS, or 5.4 lb of bone driedDGS (DDGS) per gallon of undenatured ethanol (Arora et al.,

ll rights reserved.

by UChicago Argonne, LLC,). Argonne, a US Departmentnder Contract No. DE-AC02-others acting on its behalf, ain said article to reproduce,

lic, and perform publicly and

: +1 630 252 1342.

2008). The annual production of DGS was 22 million metric tonin 2008 (www.ethanolrfa.org).

Both DG and DGS have been studied as an additional ethanolsource by enzymatic hydrolysis of the residual polysaccharides(Bals et al., 2006; Kim et al., 2008b; Tucker et al., 2004). Generally,DGS or any other biomass is subjected to a pre-treatment step forthe improvement of the efficiency of the hydrolysis (Wyman et al.,2005). Various pretreatment techniques, such as hot water (Kimet al., 2008b), dilute acid (Lloyd and Wyman, 2005), lime (Kimand Holtzapple, 2005), ammonia fiber expansion (AFEX) (Teymouriet al., 2005), ammonia recycle percolation (Kim and Lee, 2005)have been investigated thoroughly. Among them, the AFEX tech-nique is of particular interest to us. AFEX is a dry-to-dry process,producing no liquid streams. This allows the entire lignocellulosicstructure to be hydrolyzed in one reactor and at very high solidloadings. Several studies have reported improved efficiency ofenzymatic hydrolysis with AFEX on various substrates (Alizadehet al., 2005; Murnen et al., 2007; Teymouri et al., 2005). In partic-ular, AFEX pretreatment is associated with significant modifica-tions in the physical and chemical structure of DG (Kim et al.,2008b), resulting in a more open fiber structure than the untreatedDG.

S. Datta et al. / Bioresource Technology 101 (2010) 5444–5448 5445

DGS has also drawn the attention of the research community asa source of valuable proteins. Extraction of proteins from DGSusing ethanol (Wu et al., 1981; Wolf and Lawton, 1997), alkali (Balset al., 2009), and alkali–ethanol mixtures (Rosentrater et al., 2006)have been reported. A recent study has been carried out on extrac-tion of proteins from DG (oil removed) using aqueous ethanol,alkaline-ethanol and aqueous enzyme (protease) solution (Cook-man and Glatz, 2009). They have also studied the effect of particlesize reduction of DG (milling) and disulfide bond reduction (addi-tion of sodium bisulfite) on protein extraction. Without any pre-treatment (milling and sodium sulfite addition), the aqueousenzyme solution was able to extract around 90% of the proteinsfrom DG, whereas the alkaline-ethanol solution required millingto extract 90% protein from DG. The extracted protein was mainlyzein with the alkaline-ethanol extraction and hydrolyzed proteinswith the enzyme solution.

We report on a strategy for simultaneous extraction of proteinand enzymatic conversion of the polysaccharides from DGS. Thegoal is to extract a high-value animal feed with low fiber contentand produce additional sugar for ethanol production. DGS ismainly used as an animal feed for ruminants such as cattle. DGSsuffers from two disadvantages. (i) Due to high fiber content, it issuitable for monogastrics, such as hogs and poultry, only in blends,and (ii) due to high moisture content (50–60%), it must be dried ifthere is no immediate and local market. Drying DGS to producedried distiller’s grains and solubles (DDGS) is energy intensiveand requires about 1/3 of the energy consumed in ethanol drymills. Typically, the whole process of drying DGS to DDGS con-sumes 4.77 MJ of energy in the form of natural gas and 0.5 MJ ofenergy in the form of electricity per kg of DDGS (McAloon et al.,2000).

To address these challenges, we developed a method to simul-taneously extract the proteins from DGS using biobased solventsand to hydrolyze the unrecovered cellulosic materials to sugarsusing enzymes. During enzymatic hydrolysis the rigid fibrousstructure of the cellulosic materials present in DGS is degraded.The reason for selecting simultaneous protein extraction and enzy-matic saccharification is the potential that the biobased solventmight be able to penetrate the degraded DGS structure and morereadily access the residual protein. Moreover, the AFEX-treatedDGS has a more open fiber structure than the regular DGS andtherefore, the reaction mixture (bio-solvent and enzymes) has en-hanced accessibility to the target components (protein, cellulosicmaterials) present in AFEX-DGS. The protein-rich biobased solventphase could be used as a high-value animal feed. Additional bene-fits of this method are that the biobased solvents are produced asfood grade materials, which avoid the need for rigorous removal ofsolvents from the protein stream. The biobased solvents them-selves could increase nutritional value (fatty acids) of the animalfeed. This will help to increase the market for renewable biobasedsolvents. A concentrated protein stream would reduce the energyrequired for drying DGS. The hydrolyzed sugar could be used toproduce additional ethanol. Overall, the process economics of thecorn to ethanol production would be improved due to the addi-tional biofuel production, decreased energy inputs, and increasevalue of the protein co-products.

2. Methods

2.1. Equipments and materials

The initial protein extraction experiments were carried out in15 ml centrifuge tubes, rotated in a tube rotator for mixing. Thesimultaneous protein extraction and enzymatic saccharificationexperiments were conducted in 250 ml sealed glass conical flasks

kept inside an incubator/shaker with proper temperature andspeed controller.

DGS (moisture content approximately 50% total weight basis)was kindly donated by the former US Bioenergy ethanol facilityin Woodbury, MI. The DGS was kept frozen at �20 �C prior touse. All biobased solvents were formulated by Vertec BioSolvents(www.vertecbiosolvents.com). The solvents used in this studyare: Citrus-derived D-Limonene (DL), Soybean-derived distilledmethyl esters (DME) and Corn syrup-derived Ethyl Lactate (EL).The enzymes, Accelerase 1000 (cellulase) and Stargen (amylase)were generously donated by Genencor, a Danisco Division.

2.2. Experimental methods

2.2.1. Protein extractionSome preliminary experiments were conducted for protein

extraction only. The goal was to establish that the biobased sol-vents are capable of extracting protein from DGS. The additionalgoals were to select the ideal solvent(s) for protein extractionand to study the effect of temperature and extraction time on yield.In all experiments, 10 ml of biobased solvent was added to 1 g ofwet DGS in a 15 ml centrifuge tube and the tube was rotated in atube rotator. The time of extraction and the temperature of extrac-tion were varied in between 20–60 min and 30–50 �C, respectively.EL was used as the model hydrophilic biobased solvent and DL andDME were used as model hydrophobic biobased solvents. Afterextraction, the mixture was centrifuged at 4000 rpm to separatethe treated DGS from the solvent followed by washing with deion-ized water. The wet, treated DGS was weighed and analyzed formoisture and protein. Most experiments were replicated threetimes.

2.2.2. PretreatmentAmmonia fiber expansion (AFEX) was used to pretreat DGS. DGS

was loaded into a 1.5 L Parr reactor (Parr Instrument Company,Moline, IL, USA) at its native moisture content (approximately50% total weight basis) at 150 g dry DGS. The vessel was boltedshut and the ambient pressure reduced to below 0.5 atm using avacuum pump. Concurrently, 120 g anhydrous ammonia (0.8 g/gdry DGS) was loaded in a separate vessel and heated to 400 psi.The ammonia was then charged into the Parr reactor, rapidly heat-ing the biomass to its desired temperature. DGS reached its maxi-mum temperature within 1 min and slowly cooled (approximately10 �C) over the next 15 min, after which the pressure was explo-sively released. Temperatures between 80 and 120 �C were usedfor these experiments (Bals et al., 2006). After releasing the pres-sure, the pretreated DGS was allowed to dry in a fume hood toapproximately 15% moisture (total weight basis). Multiple batchesof pretreated DGS were combined and thoroughly mixed priorextraction and hydrolysis.

2.2.3. Simultaneous protein extraction and enzymatic saccharificationThe preliminary experiments for only protein extraction have

established the superiority of hydrophobic solvents over hydro-philic solvents in extracting the proteins from DGS (will be dis-cussed in Section 4). Hence, for simultaneous protein extractionand enzymatic saccharification experiments only the hydrophobicbiobased solvents were used. For all of these experiments, AFEX-treated DGS was used. 20 ml of 0.3 M Na-citrate/citric acid buffer(pH 5.2) and 10 ml of biosolvents were added to 10 g of AFEX-DGS. Tetracycline and cycloheximide were added to suppress thegrowth of microorganisms. The mixture was then pre-heated in-side an incubator/shaker (300 rpm) to 50 �C followed by the addi-tion of enzyme or enzyme mixtures. DL and DME were used assolvents. The time of reaction was varied between 48 and 120 h.In order to study the efficacy of two different enzyme mixtures

5446 S. Datta et al. / Bioresource Technology 101 (2010) 5444–5448

on the hydrolysis of cellulosic materials in the presence of thehydrophobic bio-solvents, a set of experiments were carried outwith cellulase only and another set with a mixture of cellulaseand amylase. Enzyme loadings were 1 mg of Stargen (amylase)and 10 mg of Accelerase (cellulase) for each gram of dry AFEX-DGS. Experiments with only the aqueous buffer phase (blank withrespect to bio-solvent) were also conducted. After reaction, themixture was centrifuged at 4000 rpm. After centrifugation, thethree layers of aqueous phase, bio-solvent phase and solid phasewere separated. The solid phase was then washed with 50 ml ofbuffer and re-centrifuged to recover the entrained aqueous phase.The aqueous phase (including the recovered entrained aqueousphase) was analyzed for protein and sugar, the solid phase wasanalyzed for moisture and protein, and the solvent phase was ana-lyzed for protein. Most experiments were replicated three times.

2.3. Analytical techniques

DGS composition was determined by the method described byKim et al. (2008a). The protein contents in the solid and solventphases were analyzed by nitrogen combustion method (AOAC990.03). The proteins present in the aqueous phase were separatedfrom the other components of the hydrolysate by diafiltration (2diavolumes) using a 1000 molecular weight cut-off regeneratedcellulose ultrafiltration membrane assisted centrifuge device andthen analyzed by Bradford protein assay (Bradford, 1976) tech-nique. The glucose content in the aqueous phase was determinedby HPLC using a Bio-Rad (Richmond, CA) Aminex HPX-87C column.Degassed HPLC water with a flow rate of 0.6 mL/min was used asthe mobile phase, while the temperature in the column was keptconstant at 85 �C.

3. Results and discussion

3.1. Protein extraction

The compositional analysis showed the presence of 30% crudeprotein and 18% glucan along with 10% xylan and 6% arabinan inthe untreated DGS used in this study. One gram of dry DGS(�300 mg of protein) was used in all of the experiments anddepending on the experimental conditions 15–45% of the protein(45–135 mg) was extracted with the biobased solvents. Given thetraditionally low yields from DGS protein extraction, 45% extrac-

0

5

10

15

20

25

30

35

40

45

50

Pro

tein

Rec

over

y (%

)

DL 30 CEL 30 CDL 40 CEL 40 CDL 50 CEL 50 C

20 min 30 min 60 min

DL 300

C

EL 300

C

DL 300

C

DL 300

C

EL 300

C

EL 300

C

DL 400

C

EL 400

C

DL 500

C

EL 500

C

Fig. 1. Comparison of hydrophilic (Ethyl Lactate, EL) and hydrophobic (D-Limonene,DL) bio-based solvents for extraction of proteins from DGS. For both DL and EL, oneset of experiments were conducted for 20, 30 and 60 min at 30 �C and another set ofexperiments were conducted at 30 �C, 40 �C and 50 �C for 30 min. Protein recoverywas calculated based on the initial protein content in DGS, i.e., 30% (dry basis). Thedata points represent the mean experimental value and the error bars represent95% confidence interval, wherever applicable.

tion is a promising result for further study. The amount of proteinextracted was a strong function of the types of solvent used. Fig. 1represents the amount of protein extracted as a function of timeand temperature for a hydrophilic (EL) and a hydrophobic solvent(DL). Although specific correlations between the amount of proteinextracted and process parameters (time and temperature) have notbeen established in Fig. 1, it clearly shows the superiority of hydro-phobic solvents over hydrophilic solvent for extracting proteinfrom DGS under the existing conditions. The 95% confidence inter-val test indicated a significance difference between the proteinrecovery obtained for DL and EL at a temperature of 30 �C and pro-cessing times of 20 min and 30 min as shown in Fig. 1. The p valuesfor these data points are <0.05 establishing the statistically signif-icant difference between the recovery of DL and EL. The p value forthe 60 min processing time data is 0.1, which shows a marginal dif-ference in protein recovery for DL and EL at this particular datapoint. However, the protein recovery data for higher temperatures(40 �C and 50 �C) reconfirm the superiority of hydrophobic solventover hydrophilic solvent as shown in Fig. 1. The same types of re-sults have also been observed for DME (the other hydrophobic sol-vent) compared to EL, but the data is not shown here. Fig. 1 alsodemonstrates that even at a higher temperature of 50 �C the bio-based solvents are effective in extracting significant fraction of pro-teins from DGS. This temperature is necessary for simultaneousprotein extraction and enzymatic saccharification as the enzymes(cellulase and amylase) are active near 50 �C. The superiority ofhydrophobic solvent over hydrophilic solvent could be attributedto the fact that the major amount of the extractable proteins inDGS is zein (Wolf and Lawton, 1997), which is a hydrophobic pro-tein (Shukla and Cheryan, 2001). Thus, from the preliminary exper-iments it has been concluded that the hydrophobic solvents arebetter for protein extraction from DGS and a temperature of50 �C has no adverse effect on protein extraction. Therefore, furtherexperiments for simultaneous enzymatic saccharification and pro-tein extraction were conducted with hydrophobic biobased sol-vents (DL and DME) only.

3.2. Simultaneous protein extraction and enzymatic saccharification

Designing of these experiments was not trivial as it involvedtwo different processes, aqueous phase enzymatic hydrolysis andbiosolvent phase protein extraction, running simultaneously intwo different phases. The minimum amount of water and solventrequired for these two processes was determined with the goalof minimizing downstream recovery costs in an industrial process.A phase behavior study was conducted for AFEX-DGS in water andit was observed that a minimum of 2:1 aqueous phase:solid ratiowas necessary to completely hydrate AFEX-DGS. Furthermore, itwas determined that the minimum amount of biobased solvent re-quired to properly mix the 33% solid cake was a 1:1 solid:solventratio. It was also observed that the DGS essentially retains all themoisture in the presence of the biobased solvent. Therefore, allthe future experiments with simultaneous saccharification andbio-solvent based protein extraction were performed using a1:2:1 ratio of AFEX-DGS (dry weight, g):aqueous phase (volume,ml): bio-solvent (volume, ml).

Two different enzyme formulations were used here; Acceleraseonly and a mixture of Accelerase and Stargen. Fig. 2 shows thatboth of the combinations of enzyme mixtures were efficient inhydrolyzing the cellulosic materials present in AFEX-DGS. The glu-cose content of the hydrolysates has been measured. As can beseen from Fig. 2, 50–80 mg glucose/g of dry AFEX-DGS was ob-tained after 120 h of processing time by enzymatic hydrolysis atdifferent process conditions. Literature values for hydrolysis ofDG and DGS range from 100 to 180 mg/g (Kim et al., 2008b; Tuckeret al., 2004; Bals et al., 2009; Noureddini et al., 2009). While the

0

20

40

60

80

100

knalBEMDLD

mg

gluc

ose/

gm d

ry A

FE

X-D

GS

Cellulase and AmylaseCellulase

Fig. 2. Effect of the enzyme amylase on the enzymatic saccharification of cellulosicmaterials of DGS present in different solutions. The solvents used are D-Limonene(DL), distilled methyl esters (DME) and buffer (blank). Time = 120 h, tempera-ture = 50 �C, AFEX-DGS:buffer:bio-solvent = 1:2:1. The data points represent themean experimental value and the error bars represent 95% confidence interval.

0

20

40

60

80

100

knalBEMDLD

mg

gluc

ose/

gm d

ry A

FE

X-D

GS

120 hr72 hr48 hr

Fig. 3. Effect of reaction time on the enzymatic saccharification of cellulosicmaterials of DGS present in different solutions. The solvents used are D-Limonene(DL), distilled methyl esters (DME) and buffer (blank). Temperature = 50 �C, AFEX-DGS:buffer:bio-solvent = 1:2:1. The data points represent the mean experimentalvalue and the error bars represent 95% confidence interval, wherever applicable.

Table 1Distribution of protein in different phases after simultaneous enzymatic saccharifi-cation (using both Accelerase and Stargen) and protein extraction for 120 h usingdifferent solvents.

Solvent Protein (mg)

Solubilized from solid In aqueous phase In biosolvent phase

DME 1300 1160 100DL 1200 1075 75Blank 1200 1200 –

S. Datta et al. / Bioresource Technology 101 (2010) 5444–5448 5447

results in the current work are lower than previous values, thereare three potential reasons for this discrepancy. First, the literaturevalues were performed at solid loadings ranging from 10% to 15%solids. Neither the carbohydrates nor the enzymes are soluble inthe solvents used in this study, and so the effective solid loadingfor hydrolysis in this study is 33% solids. Second, previous workin our lab suggests that different sources of DGS can vary greatlyin response to AFEX treatment and enzymatic hydrolysis, likelydue to different conditions present within various corn ethanolrefineries. Finally, recent studies of AFEX treated material at highsolid loadings suggest that most of remaining carbohydrates arein oligomeric form (Lau and Dale, 2009). Thus, a different enzymemixture or fermenting organism capable of breaking down theseoligomeric sugars might increase the monomeric sugar releaseand subsequent ethanol yield. Considering these issues, theamount of glucose obtained in this study was encouraging. Fig. 2also shows the advantage of addition of amylase (in presence ofcellulase) on enzymatic hydrolysis for DL, DME and blank solventcases. The 95% confidence interval test indicated a significantlyhigher amount of glucose obtained with amylase for all the threesolvents. The p values for all the data points (<0.05) confirm thestatistically significant difference. Approximately 1/3 of the glucanin DGS is starch, indicating the need for amylase as well as cellu-lase (Kim et al., 2008a). Hence, further experiments were con-ducted with a mixture of cellulase (Accelerase) and amylase and(Stargen).

Fig. 3 represents the amount of glucose formed for DL, DME andblank solvent with varying reaction time. It can be observed that theamount of glucose formed was significantly higher (as shown by the95% confidence interval and p values < 0.05) with increasing time ofreaction for all three solvents. Glucose (40–80 mg) per gm of dryAFEX-DGS was obtained depending on the time of reaction. It canalso be observed from Fig. 3 that the amount of glucose formedfor the same reaction time is comparable between the blank andthe biobased solvents. This implies that the biobased solvents didnot adversely affect enzymatic hydrolysis. This is an importantobservation and could open a new pathway for aqueous phase enzy-matic reactions in the presence of the biobased solvents.

Protein analysis on treated AFEX-DGS reveals that around 25–33% (1–1.3 g protein) of the total protein present in 10 g ofAFEX-DGS (4 g protein) was solubilized in the process. However,the protein analysis on the biobased solvent phase shows that only5–8% (50–100 mg) of the solubilized protein was extracted into the

biobased solvent phase. Protein analysis on the aqueous phase con-firms that the rest of the soluble protein remains in the aqueousphase (1.1–1.2 g). Table 1 shows the protein distribution in solid,aqueous and biosolvent phases after simultaneous saccharificationand protein extraction for 120 h. As can be seen from the Table 1,only 33% and 30% of the total proteins in AFEX-DGS (4 g) were sol-ubilized from the solid phase and only 8% and 6% of the solubilizedproteins were in the biobased solvent phase for DME and DL,respectively.

The loss of 25–33% of protein from treated biomass is typical forenzymatic hydrolysis experiments, even without the solvent phase(Bals et al., 2006). Thus, the biobased solvent was not able to en-hance the protein recovery from DGS. This can be attributed to themass transfer (accessibility) limitations of the biobased solvents to-wards the proteins in DGS. It can be hypothesized that the biomasswas completely surrounded by the aqueous phase, thereby shield-ing it from the hydrophobic biobased solvent. Hence, the bio-solventwas unable to access the additional proteins and extract them.Although the amount of water used during hydrolysis was mini-mized in this experiment, DGS rapidly breaks down and solubilizesduring enzymatic hydrolysis, producing a separate aqueous layerbetween the residual proteins in the solid phase and the biobasedsolvent. Since the cellulase enzymes require an aqueous media tobreak down the polysaccharides, it is not possible to completelyeliminate this layer. The fact that the transfer of protein from the so-lid DGS phase to the hydrophobic biobased solvent phase was notequilibrium (solubility) limited was established by few additionalexperiments with higher amount (20 ml instead of the usual10 ml) of hydrophobic biobased solvents. Even the higher amountof solvent was not able to extract any additional amount of protein.

On the other hand, the fact, that only 5–8% of the solubilizedprotein was in the biobased solvent phase, has suggested the

5448 S. Datta et al. / Bioresource Technology 101 (2010) 5444–5448

involvement of mass transfer and/or solubility limitations. It ispossible that during the enzymatic hydrolysis initially the proteingets transferred to the aqueous phase. After that, significant masstransfer resistance hinders the protein mobility between the twophases and it stays mainly in the aqueous phase. Another possibil-ity is that the extracted proteins are more soluble in water (hydro-philic), and therefore, prefer aqueous phase over the hydrophobicsolvent phase. Whereas, the hydrophobic proteins are still en-trapped in DGS. Determination of the partition coefficient of theextracted proteins between the aqueous and the hydrophobic bio-based solvent phase will explain the situation better, but is out ofscope of this research work.

The protein extracted in the aqueous phase was separated fromthe other components of the hydrolysate by diafiltration and recon-stituted in water. If needed, this extracted protein could be reconsti-tuted in the biobased solvents and used as high-value animal feed.Depending on the nature of the extracted proteins, hydrophobic,hydrophilic or a combination of solvents could be used. However,that process would not enable a single step biosolvent based proteinextraction and enzymatic hydrolysis. To be successful, improvedoperating conditions (size reduction, addition of other chemicals)and process engineering (mixing conditions) will be required.

4. Conclusion

Simultaneous enzymatic saccharification of cellulosic materialsand bio-solvent based extraction of proteins has been attemptedwith DGS. Preliminary study (without enzymatic saccharification)revealed that hydrophobic biosolvent extraction was more effec-tive in solubilizing protein from DGS than the hydrophilic biosol-vent. The biobased solvents were able to extract proteins fromDGS at as high a temperature as 50 �C. For the combined study,enzymatic saccharification was able to produce 40–80 mg of glu-cose per gram of dry AFEX-DGS at an effective solid loading of33% solid. While the absolute amount of glucose released is low,this is due to the nature of the DGS itself and the high solid loadingrather than any adverse effect from the biobased solvent. Only 25–33% of the total protein was solubilized from the DGS and largelyresided in the aqueous phase rather than in the bio-solvent phase.Removal of protein from DGS (25–33%) is common for enzymaticsaccharification without any solvent. Thus, biobased solvents wereunable to recover additional protein from DGS with the currentconditions. We believe that a significant fraction of total proteinsin DGS is entrapped inside the rigid fibrous structure and difficultto recover under operating conditions used in this study. Althoughonly partially successful, this study gives an insight on the behaviorof the components of DGS in presence of biobased solvents andcalls for improved process engineering and operating conditions.With the modifications, this technique has the potential to simul-taneously produce high-value animal feed and additional sugarsfrom DGS in a single step process.

Acknowledgements

The authors would like to acknowledge the US Dept. of Agricul-ture – CSREES Grant # 68-3A75-6-505 for financial support of this

research. Argonne is operated by The UChicago-Argonne LLC forthe US Dept. of Energy under contract DE-AC02-06CH11357. Wewould also like to thank the Woodbury, MI ethanol facility for-merly owned by US Bioenergy for donating the DGS, and Genencor,a Danisco Division, for donating the enzymes used in this study.

References

Alizadeh, H., Teymouri, F., Gilbert, T.I., Dale, B.E., 2005. Pretreatment of switchgrassby ammonia fiber explosion (AFEX). Appl. Biochem. Biotechnol. 121–124, 1133–1141.

Arora, S., Wu, M., Wang, M., 2008. Update of Distillers Grains Displacement Ratiosfor Corn Ethanol Life-Cycle Analysis. Argonne National Laboratory (http://www.transportation.anl.gov/pdfs/AF/527.pdf).

Bals, B., Dale, B., Balan, V., 2006. Enzymatic hydrolysis of distiller’s grain and soluble(DDGS) using ammonia fiber expansion treatment. Energy Fuels 20, 2732–2736.

Bals, B., Balan, V., Dale, B., 2009. Integrating alkaline extraction of proteins withenzymatic hydrolysis of cellulose from wet distiller’s grains and solubles.Bioresour. Technol. 100, 5876–5883.

Bradford, M.M., 1976. A rapid and sensitive method for the quantitation ofmicrogram quantities of protein utilizing the principle of protein–dye binding.Anal. Biochem. 72, 248–254.

Cookman, D.J., Glatz, C.E., 2009. Extraction of protein from distiller’s grain.Bioresour. Technol. 100, 2012–2017.

Kim, S., Holtzapple, M.T., 2005. Lime pretreatment and enzymatic hydrolysis of cornstover. Bioresour. Technol. 96, 1993–2006.

Kim, T.H., Lee, Y.Y., 2005. Pretreatment and fractionation of corn stover by ammoniarecycle percolation (ARP) process. Bioresour. Technol. 96, 2007–2013.

Kim, Y., Mosier, N.S., Hendrickson, R., Ezeji, T., Blaschek, H., Dien, B., Cotta, M., Dale,B., Ladisch, M., 2008a. Composition of corn dry-grind ethanol by-production:DDGS, wet cake, and thin stillage. Bioresour. Technol. 99, 5165–5176.

Kim, Y., Hendrickson, R., Mosier, N.S., Ladisch, M.R., Bals, B., Balan, V., Dale, B.E.,2008b. Enzyme hydrolysis and ethanol fermentation of lquid hot water andAFEX pretreated distiller’s grains at high-solids loadings. Bioresour. Technol. 99,5206–5215.

Lau, M.W., Dale, B.E., 2009. Cellulosic ethanol production from AFEX-treated cornstover using Saccharomyces cerevisiae 424A(LNH-ST). Proc. Nat. Acad. Sci. 106,1369–1373.

Lloyd, T.A., Wyman, C.E., 2005. Combined sugar yields for dilute sulfuric acidpretreatment of corn stover followed by enzymatic hydrolysis of the remainingsolids. Bioresour. Technol. 96, 1967–1977.

McAloon, A., Taylor, F., Yee, W., Ibsen, K., Wooley, R., 2000. Determining the cost ofproducing ethanol from corn starch and lignocellulosic feedstocks. NREL/TP-580-28893. National Renewable Energy Laboratory. Colorado.

Murnen, H.K., Balan, V., Chundawat, S.P.S., Bals, B., Sousa, L., Dale, B.E., 2007.Optimization of ammonia fiber expansion (AFEX) pretreatment and enzymatichydrolysis of Miscanthus � giganteus to fermentable sugars. Biotechnol. Prog.23, 846–850.

Noureddini, H., Byun, J., Yu, T., 2009. Stagewise dilute-acid pretreatment andenzyme hydrolysis of distillers’ grains and corn fiber. Appl. Biochem.Biotechnol.. doi:10.1007/s12010-009-8544-9.

Rosentrater, K., Subramanian, D., Padu, K., 2006. Fractionation techniques toconcentrate nutrient streams in distiller’s grains. ASABE Annual InternationalMeeting, Portland, OR, USA.

Shukla, R., Cheryan, M., 2001. Zein: the industrial protein from corn. Ind. Crop Prod.13, 171–192.

Teymouri, F., Laureano-Perez, L., Alizadeh, H., Dale, B.E., 2005. Optimization of theAFEX treatment parameters for enzymatic hydrolysis of corn stover. Bioresour.Technol. 96, 2014–2018.

Tucker, M., Nagle, N., Jennings, E., Ibsen, K., Aden, A., Nguyen, Q., Kim, K., Noll, S.,2004. Conversion of distiller’s grain into fuel alcohol and a higher-value animalfeed by dilute-acid pretreatment. Appl. Biochem. Biotechnol. 115, 1139–1159.

Wolf, W.J., Lawton, J.W., 1997. Isolation and characterization of zein from corndistiller’s grains and related fractions. Cereal Chem. 74, 530–536.

Wu, Y.V., Sexson, K.R., Wall, J.S., 1981. Protein-rich residue from corn alcoholdistillation: fractionation and characterization. Cereal Chem. 58, 343–347.

Wyman, C.E., Dale, B.E., Elander, R.T., Holtzapple, M., Ladisch, M.R., Lee, Y.Y., 2005.Coordinated development of leading biomass pretreatment technologies.Bioresour. Technol. 96, 1959–1966.