Cellulose crystallinity – a key predictor of the enzymatic hydrolysis rate Me ´ lanie Hall, Prabuddha Bansal, Jay H. Lee, Matthew J. Realff and Andreas S. Bommarius School of Chemical and Biomolecular Engineering, Georgia Institute of Technology, Atlanta, GA, USA Introduction The enzymatic hydrolysis of cellulose to glucose has received increased interest over the last 10 years, and growing demand for economically sustainable biofuels indicates an urgent need for reducing the costs associated with their production. Cellulose, a polysaccharide made by most plants, is one of the most abundant organic compounds on Earth and represents a major potential feedstock for the biofuels industry. However, the current enzymatic degradation of cellulose faces major issues that prevent its wide utilization in the production of economically competi- tive biofuels [1–4]. Cellulose is hydrolyzed to glucose via the synergistic action of several enzymes. Endoglucanases (EC 3.2.1.4) break down cellulose chains at random positions within the chains, whereas exoglucanases (i.e. cello- biohydrolases, EC 3.2.1.91) cleave off cellobiose speci- fically from the chain ends in a processive manner [5–10]. Cellobiose is subsequently converted into glucose by b-glucosidase (EC 3.2.1.21) [7,11–14]. The exo-endo synergism is easily expained by the fact that endo- glucanases provide more chain ends for cellobiohydro- lases to act upon [15–19]. The hydrolysis of insoluble, solid cellulose is a heterogeneous reaction, which does not match the assumptions of kinetic models based on Michaelis–Menten kinetics [13,14,20]. After an initial phase of adsorption of cellulases on cellulose, which is fast compared to hydrolysis [16,21–26], the enzymes Keywords Cel7A; cellulases; cellulose crystallinity; hydrolysis; Trichoderma reesei Correspondence A. Bommarius, School of Chemical and Biomolecular Engineering, Georgia Institute of Technology, 311 Ferst Drive, Atlanta, GA 30332-0100, USA Fax: +1 404 894 2291 Tel: +1 404 385 1334 E-mail: andreas.bommarius@chbe.gatech.edu (Received 13 December 2009, revised 16 January 2010, accepted 18 January 2010) doi:10.1111/j.1742-4658.2010.07585.x The enzymatic hydrolysis of cellulose encounters various limitations that are both substrate- and enzyme-related. Although the crystallinity of pure cellulosic Avicel plays a major role in determining the rate of hydrolysis by cellulases from Trichoderma reesei, we show that it stays constant during enzymatic conversion. The mode of action of cellulases was investigated by studying their kinetics on cellulose samples. A convenient method for reaching intermediate degrees of crystallinity with Avicel was therefore developed and the initial rate of the cellulase-catalyzed hydrolysis of cellu- lose was demonstrated to be linearly proportional to the crystallinity index of Avicel. Despite correlation with the adsorption capacity of cellulases onto cellulose, at a given enzyme loading, the initial enzymatic rate contin- ued to increase with a decreasing crystallinity index, even though the bound enzyme concentration stayed constant. This finding supports the determinant role of crystallinity rather than adsorption on the enzymatic rate. Thus, the cellulase activity and initial rate data obtained from various samples may provide valuable information about the details of the mecha- nistic action of cellulase and the hydrolysable ⁄ reactive fractions of cellulose chains. X-ray diffraction provides insight into the mode of action of Cel7A from T. reesei. In the conversion of cellulose, the (021) face of the cellulose crystal was shown to be preferentially attacked by Cel7A from T. reesei. Abbreviations CP ⁄ MAS, cross polarization ⁄ magic angle spinning; CrI, crystallinity index; DNS, dinitrosalicylic acid; PASC, phosphoric acid swollen cellulose. FEBS Journal 277 (2010) 1571–1582 ª 2010 The Authors Journal compilation ª 2010 FEBS 1571 cleave off cellobiose and move along the same chain, hydrolyzing glycosidic bonds until an event occurs that terminates cleavage. As the reaction proceeds to inter- mediate degrees of conversion, the rate of the reaction decreases dramatically, and the final part of cellulose hydrolysis requires an inordinate fraction of the overall total reaction time [27,28]. Several factors, both substrate- and enzyme-related, are suggested to be responsible for this slowdown of the reaction rate but, so far, no mechanistic explanation of the slowdown has been validated. The substrate characteristics often implied in the slowdown of the reaction rate include surface area, porosity, the degree of poly- merization, crystallinity, and the overall composition (complex substrates such as lignocellulosics versus pure cellulose). For enzyme-related features, deactiva- tion, inhibition, jamming, clogging and imperfect processivity are often cited as causes of the slowdown [14,29,30]. One of the most controversial theories concerns the influence of crystallinity and the change of the degree of crystallinity during enzymatic hydrolysis. It is accepted that the initial degree of crystallinity of cellu- lose plays a major role as a rate determinant in the hydrolysis reaction. A completely amorphous sample is hydrolyzed much faster than a partially crystalline cellulose [14,31–33], which has led to the idea that amorphous domains in a partially crystalline cellulose sample are hydrolyzed first, leaving crystalline parts to be hydrolyzed at the end, thus resulting in an increased crystallinity index (CrI) and explaining the dramatic drop in rate at higher degrees of conversion [34]. Studies to (dis)prove this phenomenon have differed in the analytical methods employed (X-ray diffraction versus solid state 13 C-NMR), the nature of the sub- strate used (complex lignocellulosics versus pure cellu- lose) and the source of the hydrolytic enzymes (mostly from Trichoderma reesei and other fungal strains) [35]. Several reviews have stated that it is difficult to con- clude that crystallinity is a key determinant of the rate of enzymatic hydrolysis [13,14,29]. Although a correla- tion between crystallinity and enzymatic hydrolysis rate has already been demonstrated, controversy remains [29]. Usually, different types of cellulose with different degrees of crystallinity are employed in these studies, such as cotton, cotton linter, Avicel, filter paper or bacterial cellulose [15,17,36,37]. Their cellu- lase-catalyzed degradation lead to hydrolysis rates that were directly related to the CrI of the cellulose sample [17,31,37–39]. To correctly relate the CrI with hydro- lysis rate, it is of prime importance to study samples that have the same basic composition and provenance. For this reason, pure cellulose may be preferable to complex substrates because the presence of lignin or hemicellulose may interfere with the action of cellulase and reduce accessibility, and therefore the hydrolysis rates [29,40,41]. Another important criterion related to hydrolysis rate involves the adsorption capacity of cellulases onto cellulose. The rate of hydrolysis was shown to be pro- portional to the amount of adsorbed enzymes [22,25,42–44]. Additionally, the difference in reactivity between a crystalline and an amorphous cellulose was found to be related to the adsorption capacity of endo- glucanases on both types of substrate [45]. Further- more, the degree of crystallinity of cellulose influences adsorption at a given protein loading and the maxi- mum adsorption constant was shown to be greatly enhanced at low crystallinity indices [46]. The same study concluded that the effective binding was the lim- iting parameter with respect to the hydrolysis rate in the case of cellulose with low degrees of crystallinity, despite a high adsorption constant. Amorphous cellulose has been widely used to inves- tigate cellulase activity [35,47–51]. Treatment with 85% phosphoric acid to produce phosphoric acid swol- len cellulose (PASC) results in complete dissolution of the sample [52] and such treatment was shown to have no impact on the reducing-end concentration of the cellulose sample (i.e. its degree of polymerization) [53,54]. However, the effect of various phosphoric acid concentrations has only been investigated across a nar- row range of acid concentrations or mainly at low con- centrations [46,55–57]. Recently, Zhang et al. [52] demonstrated that the concentration of phosphoric acid used to generate swollen cellulose relates to the rate of enzymatic hydrolysis by controlling the state of cellulose solubilization. Hydrolysis rates were one order of magnitude lower for microcrystalline cellulose compared to amorphous cellulose. This reflects the composition of highly crystalline and amorphous cellu- lose at acid concentrations of 0% and above 81%, respectively. The changes in hydrolysis rate, with varia- tions in the degree of crystallinity as a result of treat- ment with various phosphoric acid concentrated solutions, are therefore of significant interest. The present study aimed to determine the role of crystallinity and adsorption in the susceptibility of cellulose to enzymatic degradation. Both 13 C-NMR solid-state spectroscopy and X-ray crystallography were applied to investigate the crystallinity of pure cellulose (Avicel) at different degrees of conversion by cellulases from T. reesei, the most commonly studied cellulase-producing organism. Complementarily, we generated cellulose (Avicel) with controlled degrees of crystallinity using phosphoric acid solutions of precisely Cellulose crystallinity M. Hall et al. 1572 FEBS Journal 277 (2010) 1571–1582 ª 2010 The Authors Journal compilation ª 2010 FEBS calibrated concentration. These pretreated cellulose samples were employed to investigate and elucidate the relationship between the degree of crystallinity, adsorp- tion and the enzymatic hydrolysis rates. Results Cellulase hydrolysis rate and cellulose crystallinity Various types of (ligno)cellulosic substrate are employed in current enzymatic hydrolysis studies and thus are a source of discrepancies in the results obtained and the potential confusion regarding the challenging problem of understanding the mode of action of cellu- lase [35]. The presence of hemicellulose, and especially lignin, a strong adsorbent on cellulase, in lignocellulo- sics, interferes with the enzymatic activity of cellulases on cellulose [14,29,41]. To avoid such interference, we used Avicel, a commonly used, commercially and reproducibly obtainable pure cellulose substrate with a well-characterized structure and an average degree of crystallinity of 60% (measured via solid state 13 C-NMR). Phosphoric acid pretreatment First, to validate the efficiency of the phosphoric acid pretreatment, acid-pretreated samples were hydrolyzed with cellulases and an excess of b-glucosidase to remove product inhibition and fully convert cellobiose to glu- cose, and the initial hydrolysis rates were calculated in terms of the production of glucose after a 2 min reac- tion time. As expected, the more concentrated the phos- phoric acid solution, the higher the sugar production (Fig. 1A), so that the pretreatment procedure was con- sidered to be efficient. Samples treated with pure phos- phoric acid solution (maximum 85%) resulted in amorphous cellulose as demonstrated by X-ray diffrac- tion analysis [58] (Fig. 2). Furthermore, a high amount of glucose (4.75 gÆL )1 Æmin )1 ) was produced from the cellulose sample pretreated with the highest concentra- tion of acid (85%), and all of the Avicel was converted within 2.5 h compared to the 96 h that was necessary for untreated Avicel (data not shown). Phosphoric acid pretreatment has been used to create cellulose samples of various surface areas and this parameter was found to be related to the enzymatic rate [51]. A recent study using phosphoric acid to increase cellulose accessibility in lignocellulosics suggested the presence of a critical point in the phosphoric acid con- centration below which enzymatic hydrolysis was slow, and above which cellulose was easily dissolved [59]. The results obtained in the present study (Fig. 1) confirm that there is a steep change in reactivity (i.e. glucose A B C Fig. 1. Effect of phosphoric acid concentration on: (A) initial rate of Avicel enzymatic hydrolysis (glucose produced in the first 2 min of the reaction with cellulases); (B) CrI obtained from X-ray diffraction data and multivariate statistical analysis; (C) moisture content of cellulose samples after treatment with phosphoric acid (measure- ment performed after tightly controlled filtration and subsequent drying at 60 °C). The results shown are the average of at least triplicates (duplicates for crystallinity). M. Hall et al. Cellulose crystallinity FEBS Journal 277 (2010) 1571–1582 ª 2010 The Authors Journal compilation ª 2010 FEBS 1573 production) from 1 to 4.75 gÆL )1 Æmin )1 (Fig. 1A) over a narrow range of phosphoric acid content (75–80%), and not as a step change but as a steep continuum. No further increase was observed in the range 80–85%, which is the maximum possible phosphoric acid con- centration, close to the 81% obtained by Moxley et al. [59] for maximum glucan digestibility. Below 75%, the glucose production rate tends to level off, with a mini- mum being obtained with untreated Avicel (0.6 gÆL )1 Æ min )1 glucose at 0% phosphoric acid). There are several ways to measure cellulose CrI. One of the most commonly employed techniques is X-ray diffraction where the peak height is used to calculate the CrI [60] (Fig. 2). However, the major drawback of this analytical method stems from the formula itself (see Materials and methods) because it implies that amorphous cellulose gives a main reflec- tion at 2h =18°, which, upon our analysis, is defi- nitely not the case for the Avicel used in the present study (rather, it is shifted to higher angle, $19.5°). Also, the absolute values thus obtained are extremely high (> 90% for Avicel), which does not appear to represent the structure of Avicel well, and deviates substantially from the NMR analysis (60% for Avicel). In addition, the literature contains a wide range of reported values for Avicel using X-ray diffraction, in the range 62–87.6% using the peak height method [61–63], and from 39 to 75.3% using various other methods [61,64,65]. It should be noted, however, that different drying methods are often being employed, which also may add to the reported variations in absolute crystallinity values. Under our conditions, no satisfactory resolution of the C4 carbon signals in NMR analysis could be obtained below a certain degree of crystallinity and within a reasonable acquisi- tion time, so that X-ray diffraction was used as an alternative to map the full crystallinity spectrum. Given the drawbacks of the peak intensity method [60,66,67], we have developed a new method to obtain consistent CrI values using multivariate statistical analysis applied to X-ray diffraction spectra [58]. Figure 1B shows that the CrI closely tracks the breakthrough behavior of reactivity (Fig. 1A) when employing the same amount of phosphoric acid that was used to pretreat the cellulose sample: the degree of crystallinity remains fairly unchanged at approximately 55–60% over a wide range of phosphoric acid concen- trations but decreases linearly to almost 0% in a con- centration range of 75–80% phosphoric acid. Thus, the phosphoric acid effect is clearly evident: not only is it related to dissolution capacity [59], but also it dis- rupts the crystalline structure of cellulose and can turn partially crystalline cellulose amorphous. Avicel, a mi- crocrystalline type of cellulose, has a mixed composi- tion (amorphous and crystalline) and the results obtained in the present study suggest that the more concentrated the acid solution, the more crystalline regions are turned amorphous. The capacity of cellu- lose samples to retain water relative to the proportion of amorphous parts has been postulated [68,69], and was verified with the acid-treated samples. Figure 1C shows the tight relationship between moisture content and acid concentration, supporting the conclusion with respect to structural changes derived from crystallinity measurement occurring in the 75–80% acid concentra- tion range. Upon treatment at higher acid concentra- tions, cellulose samples have a higher capacity to retain water, owing to the higher number of hydroxyl groups that are available to bind to (and adsorb) water molecules because these hydroxyl groups are no longer hydrogen bonded to other glucose moieties. A cellulose sample with 85% moisture content can theoretically accommodate 49 water molecules per glucose unit, whereas, at a 60% moisture content, this ratio is reduced to 13 (based on the observation that 1 g of Avicel yields 1.15 g of glucose at 100% conversion). Cellulose enzymatic hydrolysis There have been numerous, and sometimes controver- sial, studies on the change of cellulose crystallinity I am 20 30 40 I 002 Fig. 2. X-ray diffraction pattern of microcrystalline cellulose Avicel (multiple peaks) and amorphous Avicel (single smooth peak) gener- ated with 85% phosphoric acid (reflection around 20° is attributed to amorphous parts and gives a CrI of 0% based on peak intensity method) [60]. x-axis: Bragg angle (2h). I 002 represents the maximum intensity at 2h = 22.5°, I am shows the minimum intensity at 2h =18° used to calculate crystallinity in the peak height method, and the straight line represents the background (see Materials and methods). Cellulose crystallinity M. Hall et al. 1574 FEBS Journal 277 (2010) 1571–1582 ª 2010 The Authors Journal compilation ª 2010 FEBS upon enzymatic hydrolysis. Both trends (i.e. increased degree of crystallinity over conversion and no change over conversion) were observed at different levels of intensity [14,31,33,70,71]. As mentioned above, the dif- ferent types of substrate as well as the analytical meth- ods employed contributed to the absence of a clear understanding of the mechanistic action of cellulase on partially crystalline cellulose. Furthermore, in situ measurements of cellulose structure under reacting conditions (i.e. in aqueous buffers) are difficult to perform because all current methods require the prior isolation of cellulose and drying [29]. The CrI of Avicel was monitored via X-ray diffrac- tion during its hydrolysis by a commercial mixture of cellulases from T. reesei and an excess of b-glucosidase to prevent cellobiose inhibition. The X-ray diffraction data obtained gave an artificially high degree of crys- tallinity for untreated Avicel (92%) using the method of Segal et al. [60]. Small variations at such high values are challenging to monitor; therefore, cross polariza- tion ⁄ magic angle spinning (CP ⁄ MAS) 13 C-NMR spec- troscopy was employed as an alternate method. The CrI of untreated Avicel (calculated as described previ- ously) [28] averaged 61% and was found to be constant over the course of hydrolysis, until approximately 90% conversion (Fig. 3). Similarly, using purified Cel7A from T. reesei (see Materials and methods) instead of a mixture of cellulases, no change in crystallinity was observed; however, variations in relative peak intensity in X-ray diffraction patterns showed that Cel7A attacked preferentially the (021) plane of the crystal because the peak corresponding to this face (centered around 21°) disappeared after 20% conversion (Fig. 4). Overall, peak intensity ratios for the other peaks were conserved [planes (101), (10  1), (002) and (040) at 15, 16, 22.5 and 35°, respectively]. The same trend was observed with the commercial cellulase mixture, implying no competition for this plane from the other enzymes (endoglucanases, Cel6A and b-glucosidase) or any dominant behavior from Cel7A. The implications of this preferential attack need to be investigated further because this may provide options for engineering Cel7A and thus enable overall faster hydrolysis. Adsorption Adsorption studies were conducted using cellulose samples generated with various amounts of phosphoric acid and thus displaying intermediate degrees of crys- tallinity (Fig. 1B). Adsorption experiments were car- ried out at 4 °C to prevent the hydrolysis of cellulose and the resulting loss of adsorbent material that would ultimately bias the results. Furthermore, the adsorp- tion profile at 4 °C was found to be similar to that at 50 °C after 30 min [46]. The adsorption step has been shown to be rapid, with half of the maximally adsorbed enzyme being bound with 1–2 min and the adsorption equilibrium being reached after 30 min [22]. Adsorption experiments were first performed using the same degree of loading as employed during a com- mon enzymatic hydrolysis run (175 lgÆmg )1 cellulose; Figs 1–3). Surprisingly, a maximum value of adsorbed enzyme concentration (150 lgÆmg )1 cellulose) was reached for the cellulose samples with a CrI below a threshold value of approximately 45% (Fig. 6A, open triangles), whereas the amount of adsorbed enzyme Fig. 3. CrI of Avicel monitored during hydrolysis with cellulases via CP ⁄ MAS 13 C-NMR [reactions were run at 50 °C in sodium acetate buffer (50 m M,pH5)at20gÆL )1 Avicel with the addition of b-gluco- sidase (15 kUÆL )1 ) and cellulases (24 mLÆL )1 , 3.4 gÆL )1 total protein)]. The results shown are the average of duplicates. Fig. 4. X-ray diffraction patterns of untreated Avicel and partially converted cellulose in the range 10–40° (2h). x-axis: Bragg angle (2h). The reflection of face (021) of the crystal (centered around 21°) is visible only for untreated Avicel. M. Hall et al. Cellulose crystallinity FEBS Journal 277 (2010) 1571–1582 ª 2010 The Authors Journal compilation ª 2010 FEBS 1575 appeared to increase inversely and linearly with the CrI at higher crystallinity values (i.e. > 45%). A con- stant amount of adsorbed enzymes ($ 150 lgÆmg )1 cellulose) led to faster hydrolysis at lower degrees of crystallinity (i.e. < 45% CrI; Fig. 6B), whereas, at crystallinity indices above 45%, the adsorption capac- ity increased and was linearly proportional to the initial rate. At higher enzyme loading (seven-fold greater than the original loading; i.e. 1230 lgÆmg )1 cellulose), the initial rates were found to be generally higher (Fig. 6C, filled circles), confirming the findings obtained in previous studies [22,25,42–44], although this trend was especially true at lower degrees of crys- tallinity. By contrast, untreated Avicel (CrI = 60%) displayed similar rates at both enzyme concentrations, and little difference in rate for the two enzyme con- centrations was observed up to a CrI of 50%. Also at high enzyme loading, the profile of adsorbed enzyme versus the degree of crystallinity ⁄ initial rate was similar to that at low enzyme loading, except that constant adsorption was observed only for CrI in the range 0–35%. Discussion Cellulase hydrolysis rate and cellulose crystallinity The correlation between the CrI and the initial hydro- lysis rate (Fig. 5) shows a continuous decrease in rate as crystallinity increases. At higher degrees of crystallinity, cellulose samples are less amenable to enzymatic hydrolysis, less reactive and less accessible. A B C Fig. 6. Adsorption, CrI and initial rates at two cellulases loadings: D, 175 lgÆmg )1 cellulose; •, 1230 lgÆmg )1 cellulose. Initial rates correspond to the amount of glucose produced over a 2 min reac- tion (20 mgÆmL )1 cellulose, cellulases at 175 resp. 1230 lgÆmg )1 cellulose and an excess of b-glucosidase, 50 °C). Adsorption stud- ies were conducted at 4 °C over 30 min. (A) Adsorption versus CrI; (B) initial rate versus adsorption; (C) initial rate versus CrI, where the grey shaded area represents the importance and role of adsorp- tion on enzymatic rate. Dotted lines are added for clarity to help identify trends. The results shown are the average of quadrupli- cates. Fig. 5. Effect of crystallinity (obtained from X-ray diffraction data and multivariate statistical analysis) on the initial rate in Avicel enzymatic hydrolysis (glucose produced in the first 2 min of the reaction with cellulases). The results shown are the average of quadruplicates. Cellulose crystallinity M. Hall et al. 1576 FEBS Journal 277 (2010) 1571–1582 ª 2010 The Authors Journal compilation ª 2010 FEBS The latter is supported by the data obtained from moisture content measurement (Fig. 1C). Most aque- ous reagents can only penetrate the amorphous parts of cellulose; therefore, these domains are also termed the accessible regions of cellulose, and crystallinity and accessibility are closely related [68]. It is likely that crystallinity and accessibility are related; however, moisture content (i.e. the capacity to retain water) by itself is not directly related to enzyme accessibility because water molecules are three orders of magnitude smaller than cellulases [72]. A highly crystalline cellu- lose sample has a tight structure with cellulose chains closely bound to each other, leaving too little space for enzymes to initiate the hydrolysis process anywhere within the cellulose crystal. Overall, the hydrolysis rate versus the phosphoric acid concentration profile resembles a very steep and sharp sigmoid curve (Fig. 1A), which led to an evalua- tion of the concentration range corresponding to the sigmoid region. In their review, Zhang et al. [35] stressed that the CrI of cellulose was not strongly asso- ciated with hydrolysis rates. By contrast, the results obtained in the present study show a very close and lin- ear relationship between the CrI and initial hydrolysis rate for samples of same origin obtained after pretreat- ment with phosphoric acid (R 2 = 0.96; Fig. 5), demon- strating that crystallinity is a good predictor of the hydrolysis rate. More precisely, in a phosphoric acid concentration range of 75–80%, the hydrolysis rate, crystallinity and phosphoric acid concentration are mutually dependent parameters resulting from the structural changes that take place upon acid pretreat- ment of cellulose and are also linearly related. The degree of phosphoric acid addition enables the tight control of the overall structure of cellulose in the Avicel sample. This convenient method for reaching intermedi- ate degrees of crystallinity allows the exclusion of addi- tional parameters that might influence the enzymatic action on cellulose, such as the type and source of cellu- lose or mixed components, and yields an explicit proof of the tight relationship between initial cellulose crystal- linity and the rate of degradation by cellulases from T. reesei. The use of this method could support kinetics studies where the estimation of intrinsic parameters for cellulose is needed. Furthermore, because the interpre- tation of crystallinity data is not trivial, looking at initial hydrolysis rates may be an elegant alternative to estimating the degree of crystallinity of pure cellulose. No significant change was observed in the degree of crystallinity during the enzymatic hydrolysis of Avicel up to 90% conversion (Fig. 3). Despite their ability to distinguish different degrees of crystallinity, cellulases are not efficient at reducing ⁄ disrupting overall cellulose crystallinity, most likely because cellulose chains are hydrolyzed as soon as their interactions with the crys- tal are disrupted, therefore leaving an overall unchanged crystallinity but a structure that is reduced in size. Thus, the belief that mixed cellulose samples have their amorphous components hydrolyzed first is not consistent with the results obtained in the present study. The change in crystallinity cannot account for the sharp decrease in reaction rate observed, and thus another explanation is required for the slowdown. A number of studies reporting an increasing crystal- linity along enzymatic hydrolysis have attributed the slowdown in the rate to this crystallinity change [25,33,34,39,71]. However, the changes reported were often modest. Figure 3 shows that a 10% increase in CrI at high CrI values leads to a 40% decrease in ini- tial rate; therefore, it does not appear physically possi- ble that a change in CrI by some percentage points results in such dramatic drops in the rate. Constant crystallinity and decreased rates indicate surface changes on cellulose that start rapidly after the begin- ning of hydrolysis. Factors other than crystallinity impeding enzymatic action (both enzyme- and sub- strate-related) require closer attention. Adsorption There are multiple substrate-related factors that can influence the reaction rate in the enzymatic hydrolysis of cellulose (see Introduction). From the results obtained in the present study with respect to determin- ing the role of crystallinity in enzymatic activity, it is logical to ask whether crystallinity might not be mask- ing another phenomenon, specifically adsorption. A constant adsorption profile at different enzyme concentrations was found to relate to increasing hydro- lysis rates at decreasing degrees of crystallinity (Fig. 6) and supports our previous conclusion. This is in contrast to studies stating that increased hydrolysis rates were likely the result of an increasing adsorptive capacity rather than substrate reactivity [14]. The observed phenomenon is most likely the result of a difference in the amount of productively bound enzyme and the percentage of surface coverage. Indeed, at low degrees of crystallinity, adsorbed enzymes are more active at the same overall concentration (i.e. initial rates are higher; Fig. 6C), most likely because of a more open cellulose structure that prevents enzyme molecules residing on neighboring chains from hindering one another [73]. At a very low CrI and constant adsorbed enzyme concentration, the percentage of surface cover- age is smaller because the surface area is larger at M. Hall et al. Cellulose crystallinity FEBS Journal 277 (2010) 1571–1582 ª 2010 The Authors Journal compilation ª 2010 FEBS 1577 lower crystallinity indices [14]. Exoglucanases may also locate a chain end faster on an open structure and thus be able to start hydrolysis immediately after binding (initial rates were determined after only a 2 min reac- tion time). Accessibility was suggested to be an impor- tant factor that affects enzymatic hydrolysis rates [72] and its increase at lower degrees of crystallinity was proposed as a reason for enhanced digestibility [59]. It has also been suggested that rendering the substrate more amorphous increases access to the reducing ends of cellulose, thus enhancing reaction rates [53]. These data support these hypotheses only partially and, importantly, demonstrate that the effect of improved access on the hydrolysis rate is limited to higher degrees of crystallinity, whereas, at low degrees of crystallinity, rate enhancement is strictly the result of a dynamic cause that is independent of the adsorption phase (favored enzymatic motion as result of the larger free space available at lower degrees of crystallinity), and is also directly related to the enzyme concentra- tion. This can be related to recent work demonstrating that the overcrowding of enzymes on the cellulose surface lowers their activity [74]. Surface area may also play a role in the various rate profiles observed. Some studies have focused on the relationships between surface area and crystallinity [75]; overall, a reduction in crystallinity would relate to a higher surface area. In the present study, this would easily explain the higher adsorption capacity observed at lower degrees of crystallinity but not why the adsorption reaches a plateau (in an undersaturated regime) below a certain CrI and the rates keep increasing. Also, the internal surface of highly crystalline cellulose is poorly acces- sible to enzymes, leading to such low adsorption, pos- sibly in contrast to more amorphous samples. An accessible surface area has been the subject of numer- ous studies [40] but, in view of the results obtained in the present study, this does not appear to be the only critical parameter with respect to controlling hydro- lysis rates. Avicel hydrolysis rates were not significantly chan- ged upon the addition of a much higher enzyme con- centration for samples displaying a degree of crystallinity in the range 60–50% (Fig. 6C), demon- strating that all hydrolysable fractions of cellulose were already covered by enzymes at lower loading, despite an increase in the amount of adsorbed cellulase at higher loading. High enzyme loading (1230 lgÆmg )1 cellulose) resulted in saturation of the Avicel surface, whereas low enzyme loading (175 lgÆmg )1 cellulose) led to less than full but more than half-saturation (adsorption isotherms not shown). In other words, a higher cellulose surface coverage (in an undersaturated regime) does not necessarily lead to higher rates because it might simply result in unproductive binding once all of the hydrolysable fractions are covered. The role of adsorption for a given cellulose sample appears to be more important to the enzymatic rate at lower degrees of crystallinity (Fig. 6C). At higher enzyme loading, crystallinity appears to play a minor role (Fig. 6A). At degrees of crystallinity in the range 60–35%, the amount of adsorbed enzyme increases linearly, whereas adsorption is constant below a breakpoint that can be estimated at approximately 35% CrI (compared to 45% at lower enzyme loading). Below 35% CrI, a maximum of absorbed cellulases was reached ($ 600 lgÆmg )1 cellulose), whereas the initial rates were still increasing (Fig. 6). The breakpoint below which crystallinity is the only determining factor for the reaction rate is expected to decrease as enzyme loading increases because it becomes comparatively harder to attain the maximum adsorption capacity (saturation) at low degrees of crystallinity (open cellulose structure) as well as the maximum coverage of hydrolysable fractions (investigations underway). Examining various enzyme concentrations and hydrolysis rate ⁄ adsorption profiles on substrates with different degrees of crystallinity may thus provide an effective way of quantifying cellulose hydrolyzability. Finally, future trends for the application of cellu- lases in biofuel technology should focus on efficient ways of disrupting cellulose crystallinity and thus render the overall process economically more viable by reducing the time required to reach full conversion. Materials and methods Materials All chemicals and reagents were purchased from Sigma (St Louis, MO, USA) unless otherwise stated. Avicel PH-101, cellulases from T. reesei (159 FPUÆmL )1 ) and b-glucosidase (from almonds, 5.2 UÆmg )1 ) were obtained from Sigma and phosphoric acid (85%) was obtained from EMD (Gibbs- town, NJ, USA). Trichoderma reesei QM9414 strain was obtained from ATCC (#26921; American Type Culture Collection, Manassas, VA, USA). The BCA protein assay kit was obtained from Thermo Fischer Scientific (Rockford, IL, USA). Phosphoric acid pretreatment One gram of slightly moistened Avicel was added to 30 mL of an ice-cold aqueous phosphoric acid solution (concentration range 42–85% weight) and allowed to react over 40 min with occasional stirring. After the addition of 20 mL of ice-cold Cellulose crystallinity M. Hall et al. 1578 FEBS Journal 277 (2010) 1571–1582 ª 2010 The Authors Journal compilation ª 2010 FEBS acetone and subsequent stirring, the resulting slurry was filtered over a fritted filtered-funnel and washed three times with 20 mL of ice-cold acetone, and four times with 100 mL of water. The resulting cellulose obtained after the last filtration was used as such in the enzymatic hydrolysis experiments, and the moisture content was estimated upon oven-drying at 60 °C overnight. Samples were freeze-dried prior to X-ray diffraction measurement. Enzymatic hydrolysis of cellulose A suspension of Avicel (20 gÆL )1 ) in sodium acetate buffer (1 mL, 50 mm, pH 5) was hydrated for 1 h with stirring at 50 °C. b-Glucosidase (15 kUÆL )1 ) and cellulases (24 mLÆL )1 , 3.4 gÆL )1 total protein) were added and the mixture was stirred at 50 °C. At the desired time points, samples were centrifuged, and glucose content in the super- natant was measured via the dinitrosalicylic acid (DNS) assay. For crystallinity measurements at various conversion levels using CP ⁄ MAS 13 C- NMR [and the corresponding Eqn (2)], reactions were run on a 15 mL scale (one reaction tube per time point, ranging from 10 min to 92 h) and, after centrifugation and washing with buffer and water, recovered cellulose was either freeze-dried, oven-dried (60 °C) or air-dried. When Cel7A was used as single cellu- lase component, 92 lg of purified enzyme per mg of Avicel were added to the reaction mixture. Determination of glucose content Glucose released from cellulose was measured using the DNS assay, as described previously [28]. The calibration curve was generated with pure glucose standards. DNS assay was compared with HPLC analysis and found to yield identical conversion results. Determination of the degree of crystallinity of cellulose X-ray diffraction X-ray diffraction patterns of cellulose samples obtained after freeze-drying were recorded with an X’Pert PRO X-ray diffractometer (PANanalytical BV, Almelo, the Netherlands) at room temperature from 10 to 60 °C, using Cu ⁄ Ka 1 irradiation (1.54 A ˚ ) at 45 kV and 40 mA. The scan speed was 0.021425°Æs )1 with a step size of 0.0167°. CrI was calculated using the peak intensity method [60]: CrI ¼ðI 002 À I am Þ=I 002  100 ð1Þ where I 002 is the intensity of the peak at 2h = 22.5° and I am is the minimum intensity corresponding to the amor- phous content at 2h =18°. Freeze-drying showed no impact on the crystallinity of untreated Avicel. Solid state 13 C-NMR The solid-state CP ⁄ MAS 13 C-NMR experiments were per- formed on a Bruker Avance ⁄ DSX-400 spectrometer (Bruker Instuments, Inc., Bellerica, MA, USA) operating at frequencies of 100.55 MHz for 13 C. All the experiments were carried out at ambient temperature using a Bruker 4-mm MAS probe. The samples ($35% moisture content) were packed in 4 mm zirconium dioxide rotors and spun at 10 kHz. Acquisition was carried out with a CP pulse sequence using a 5 ls pulse and a 2.0 ms contact pulse over 4 h. CrI was calculated according to standard methods [28]: CrI ¼ A 86À92 p:p:m: =ðA 79À86 p:p:m: þ A 86À92 p:p:m: ÞÂ100 ð2Þ where A 86–92 p.p.m. and A 79–86 p.p.m. are the areas of the crys- talline and amorphous C4 carbon signal of cellulose, respectively. Oven-drying (60 °C) showed no impact on the crystallin- ity of untreated Avicel. Multivariate statistical analysis of X-ray data The CrI of cellulose samples was also calculated by quanti- fying the contribution of amorphous cellulose (PASC) and Avicel to its (normalized) X-ray diffraction spectra [58]: I j ð2hÞ¼f j I p ð2hÞþð1 À f j ÞI c ð2hÞþe ð3Þ where I j (2h) is the intensity of the j th sample at diffraction angle 2h, I p (2h) is the intensity of PASC at diffraction angle 2h, I C (2h) is the intensity of untreated Avicel at dif- fraction angle 2h, f j is the contribution of PASC to the spectrum and e is the random error. ^ f j ,the least square estimate of f j , was used to estimate the crystallinity by multiplying the contribution of Avicel ð1 À ^ f j Þ by its crystallinity (calculated by CP ⁄ MAS 13 C- NMR as 60%): CrI j ¼ð1 À f j ÞÂCrI c ð4Þ where CrI j is the crystallinity (in percentage) of the j th sample of Avicel AND CrI c is the crystallinity of Avicel (calculated by CP ⁄ MAS 13 C-NMR as 60%). Cel7A purification Trichoderma reesei QM9414 was grown on potato dextrose agar plate under light illumination. Spores were harvested and used to inoculate the liquid medium (minimal medium: (NH 4 ) 2 SO 4 5gÆL )1 , CaCl 2 0.6 gÆL )1 , MgSO 4 0.6 gÆL )1 , KH 2 PO 4 15 gÆL )1 , MnSO 4 .H 2 O 1.5 mgÆL )1 , FeSO 4 .7H 2 O 5mgÆL )1 , COCl 2 2mgÆL )1 , ZnSO 4 1.5 mgÆL )1 ) supple- mented with glucose (2%). After 3 days at 28 °C and 150 r.p.m., the fungus was grown on lactose (2%) in mini- M. Hall et al. Cellulose crystallinity FEBS Journal 277 (2010) 1571–1582 ª 2010 The Authors Journal compilation ª 2010 FEBS 1579 mal medium for up to 12 days at 28 °C and 150 r.p.m. After filtration over glass-microfiber filter (1.6 lmGF⁄ A; Whatman, Maidstone, UK), the filtrate was diafiltered by repeated concentration and dilution with sodium acetate buffer (50 mm, pH 5.5) using a polyethersulfone membrane (molecular weight cut-off of 10 kDa). The concentrate was purified by means of anion-exchange chromatography using a Q-Sepharose Fast Flow with a 10–500 mm sodium acetate gradient (pH 5.5). Cel7A was eluted in the last peak, and purity was confirmed by SDS-PAGE, where only one single protein band was observable ($ 67 kDa). Enzyme concen- trations were estimated by the Bradford assay, using BSA as standard. Adsorption study Cellulose samples (20 mgÆmL )1 ) in NaOAc buffer (50 mm, pH 5) were incubated at 50 °C for 1 h at 900 r.p.m., and then were cooled down to 4 °C. Cellulases were added in various amounts and the mixture was further agitated for 30 min. After centrifugation, the supernatant was collected and protein content analysis was performed using the BCA protein assay kit (Thermo Fischer Scientific). Acknowledgements Chevron Corporation is acknowledged for their fund- ing. Dr J. Leisen and Dr J. I. Hong are thanked for their technical assistance with the crystallinity measure- ments. References 1 Himmel ME, Ding SY, Johnson DK, Adney WS, Nimlos MR, Brady JW & Foust TD (2007) Biomass recalcitrance: engineering plants and enzymes for biofuels production. Science 315, 804–807. 2 Himmel ME, Ruth MF & Wyman CE (1999) Cellulase for commodity products from cellulosic biomass. Curr Opin Biotechnol 10, 358–364. 3 Wyman CE (2007) What is (and is not) vital to advanc- ing cellulosic ethanol. Trends Biotechnol 25, 153–157. 4 Lynd LR, Laser MS, Brandsby D, Dale BE, Davison B, Hamilton R, Himmel M, Keller M, McMillan JD, Sheehan J et al. (2008) How biotech can transform biofuels. Nat Biotechnol 26, 169–172. 5 Divne C, Sta ˚ hlberg J, Reinikainen T, Ruohonen L, Pettersson G, Knowles JKC, Teeri TT & Jones TA (1994) The 3-dimensional crystal-structure of the catalytic core of cellobiohydrolase-i from Trichoderma reesei. Science 265, 524–528. 6 Divne C, Sta ˚ hlberg J, Teeri TT & Jones TA (1998) High-resolution crystal structures reveal how a cellulose chain is bound in the 50 angstrom long tunnel of cello- biohydrolase I from Trichoderma reesei. J Mol Biol 275, 309–325. 7 Teeri TT (1997) Crystalline cellulose degradation: new insight into the function of cellobiohydrolases. Trends in Biotechnology 15, 160–167. 8 Kipper K, Va ¨ ljama ¨ e P & Johansson G (2005) Processive action of cellobiohydrolase Cel7A from Trichoderma reesei is revealed as ‘burst’ kinetics on fluorescent poly- meric model substrates. Biochem J 385, 527–535. 9 Li YC, Irwin DC & Wilson DB (2007) Processivity, substrate binding, and mechanism of cellulose hydroly- sis by Thermobifida fusca Ce19A. Appl Environ Microbiol 73, 3165–3172. 10 Davies G & Henrissat B (1995) Structures and mecha- nisms of glycosyl hydrolases. Structure 3 , 853–859. 11 Henrissat B (1994) Cellulases and their interaction with cellulose. Cellulose 1, 169–196. 12 Knowles J, Lehtovaara P & Teeri T (1987) Cellulase families and their genes. Trends in Biotechnology 5, 255–261. 13 Lynd LR, Weimer PJ, van Zyl WH & Pretorius IS (2002) Microbial cellulose utilization: fundamentals and biotechnology. Microbiol Mol Biol Rev 66 , 506–577. 14 Zhang YHP & Lynd LR (2004) Toward an aggregated understanding of enzymatic hydrolysis of cellulose: noncomplexed cellulase systems. Biotechnol Bioeng 88, 797–824. 15 Va ¨ ljama ¨ e P, Sild V, Nutt A, Pettersson G & Johansson G (1999) Acid hydrolosis of bacterial cellulose reveals different modes of synergistic action between cellobio- hydrolase I and endoglucanase I. Eur J Biochem 266, 327–334. 16 Medve J, Karlsson J, Lee D & Tjerneld F (1998) Hydrolysis of microcrystalline cellulose by cellobiohy- drolase I and endoglucanase II from Trichoderma reesei: adsorption, sugar production pattern, and synergism of the enzymes. Biotechnol Bioeng 59, 621–634. 17 Hoshino E, Shiroishi M, Amano Y, Nomura M & Kanda T (1997) Synergistic actions of exo-type cellulas- es in the hydrolysis of cellulose with different crystallini- ties. J Ferment Bioeng 84, 300–306. 18 Nidetzky B, Hayn M, Macarron R & Steiner W (1993) Synergism of Trichoderma reesei cellulases while degrad- ing different celluloses. Biotechnol Lett 15, 71–76. 19 Henrissat B, Driguez H, Viet C & Schulein M (1985) Synergism of cellulases from Trichoderma reesei in the degradation of cellulose. Biotechnology 3, 722–726. 20 Bansal P, Hall M, Realff MJ, Lee JH & Bommarius AS (2009) Modeling cellulase kinetics on lignocellulosic substrates. Biotechnol Adv 27, 833–848. 21 Sta ˚ hlberg J, Johansson G & Pettersson G (1991) A new model for enzymatic-hydrolysis of cellulose based on the 2-domain structure of cellobiohydrolase-I. Biotech- nology 9, 286–290. Cellulose crystallinity M. Hall et al. 1580 FEBS Journal 277 (2010) 1571–1582 ª 2010 The Authors Journal compilation ª 2010 FEBS [...]... confined) enzymatic catalysis: contributions from the fractal and jamming (overcrowding) effects Appl Catal A- Gen 317, 7 0–8 1 31 Fan LT, Lee YH & Beardmore DH (1980) Mechanism of the enzymatic- hydrolysis of cellulose – effects of major structural features of cellulose on enzymatichydrolysis Biotechnol Bioeng 22, 17 7–1 99 32 Fan LT, Lee YH & Beardmore DR (1981) The influence of major structural features of cellulose. .. concentration on the rate of the hydrolysis of cellulose Biotechnol Bioeng 33, 122 1–1 234 43 Lee YH & Fan LT (1982) Kinetic-studies of enzymatichydrolysis of insoluble cellulose – analysis of the initial rates Biotechnol Bioeng 24, 238 3–2 406 44 Ryu DDY & Lee SB (1986) Enzymatic- hydrolysis of cellulose – determination of kinetic-parameters Chem Eng Commun 45, 11 9–1 34 45 Klyosov AA, Mitkevich OV & Sinitsyn AP... kinetics of hydrolysis by cellobiohydrolases I and II is consistent with a cellulose surfaceerosion model Eur J Biochem 253, 46 9–4 75 Igarashi K, Wada M, Hori R & Samejima M (2006) Surface density of cellobiohydrolase on crystalline celluloses – a critical parameter to evaluate enzymatic kinetics at a solid-liquid interface FEBS Journal 273, 286 9–2 878 Walker LP & Wilson DB (1991) Enzymatic- hydrolysis of cellulose. .. Szijarto N, Siika-aho M, Tenkanen M, Alapuranen M, Vehmaanpera J, Reczeya K & Viikari L (2008) Hydrolysis of amorphous and crystalline cellulose by heterologously produced cellulases of Melanocarpus albomyces J Biotechnol 136, 14 0–1 47 51 Stone JE, Scallan AM, Donefer E & Ahlgren E (1969) Digestibility as a simple function of a molecule of similar size to a cellulase enzyme Adv Chem Ser 95, 21 9–2 41... lignocellulose fractionation technology and enzymatic cellulose hydrolysis J Agric Food Chem 56, 788 5–7 890 Segal L, Creely JJ, Martin AE & Conrad CM (1959) An empirical method for estimating the degree of crystallinity of native cellulose using the x-ray diffractometer Text Res J 29, 78 6–7 94 Thygesen A, Oddershede J, Lilholt H, Thomsen AB & Stahl K (2005) On the determination of crystallinity and cellulose. .. (1986) Role of the activity and adsorption of cellulases in the efficiency of the enzymatic- hydrolysis of amorphous and crystalline cellulose Biochemistry 25, 54 0–5 42 46 Lee SB, Shin HS, Ryu DDY & Mandels M (1982) Adsorption of cellulase on cellulose – effect of physicochemical properties of cellulose on adsorption and rate of hydrolysis Biotechnol Bioeng 24, 213 7–2 153 47 Schulein M (1997) Enzymatic properties... Vertova A (1999) Microcrys- 1582 65 66 67 68 69 70 71 72 73 74 75 talline cellulose powders: structure, surface features and water sorption capability Cellulose 6, 5 7–6 9 Soltys J, Lisowski Z & Knapczyk J (1984) X-Ray diffraction study of the crystallinity index and the structure of the microcrystalline cellulose Acta Pharm Technologica 30, 17 4–1 80 Park S, Johnson DK, Ishizawa CI, Parilla PA & Davis... on crystallinity and average chain length for bacterial and microcrystalline celluloses Cellulose 14, 28 3– 293 Hong J, Ye XH & Zhang YHP (2007) Quantitative determination of cellulose accessibility to cellulase based on adsorption of a nonhydrolytic fusion protein containing CBM and GFP with its applications Langmuir 23, 1253 5–1 2540 Valjamae P, Sild V, Pettersson G & Johansson G ¨ ¨ (1998) The initial... screening and selection strategies Biotechnol Adv 24, 45 2–4 81 36 McLean BW, Boraston AB, Brouwer D, Sanaie N, Fyfe CA, Warren RAJ, Kilburn DG & Haynes CA (2002) Carbohydrate-binding modules recognize fine substructures of cellulose J Biol Chem 277, 5024 5–5 0254 Cellulose crystallinity 37 Puri VP (1984) Effect of crystallinity and degree of polymerization of cellulose on enzymatic saccharification Biotechnol... Hatakeya T & Nakano J (1974) Dsc study on recrystallization of amorphous cellulose with water J Appl Polym Sci 18, 306 9–3 076 Ramos LP, Zandona A, Deschamps FC & Saddler JN (1999) The effect of Trichoderma cellulases on the fine structure of a bleached softwood kraft pulp Enzyme Microb Technol 24, 37 1–3 80 Chen Y, Stipanovic AJ, Winter WT, Wilson DB & Kim YJ (2007) Effect of digestion by pure cellulases . Cellulose crystallinity – a key predictor of the enzymatic hydrolysis rate Me ´ lanie Hall, Prabuddha Bansal, Jay H. Lee, Matthew J. Realff and Andreas. Avicel. Multivariate statistical analysis of X-ray data The CrI of cellulose samples was also calculated by quanti- fying the contribution of amorphous cellulose (PASC)