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: email@example.com
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 cellulose 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 continued 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 mechanistic 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.
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 competitive biofuels [1–4].
Cellulose is hydrolyzed to glucose via the synergistic action of several enzymes. Endoglucanases (EC 18.104.22.168) break down cellulose chains at random positions within the chains, whereas exoglucanases (i.e. cellobiohydrolases, EC 22.214.171.124) cleave off cellobiose specifically from the chain ends in a processive manner [5–10]. Cellobiose is subsequently converted into glucose by β-glucosidase (EC 126.96.36.199) [7,11–14]. The exo-endo synergism is easily expained by the fact that endoglucanases provide more chain ends for cellobiohydrolases 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 cleave off cellobiose and move along the same chain, hydrolyzing glycosidic bonds until an event occurs that terminates cleavage. As the reaction proceeds to intermediate 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 polymerization, crystallinity, and the overall composition (complex substrates such as lignocellulosics versus pure cellulose). For enzyme-related features, deactivation, 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 cellulose 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 .
Studies to (dis)prove this phenomenon have differed in the analytical methods employed (X-ray diffraction versus solid state 13C-NMR), the nature of the substrate used (complex lignocellulosics versus pure cellulose) and the source of the hydrolytic enzymes (mostly from Trichoderma reesei and other fungal strains) . Several reviews have stated that it is difficult to conclude that crystallinity is a key determinant of the rate of enzymatic hydrolysis [13,14,29]. Although a correlation between crystallinity and enzymatic hydrolysis rate has already been demonstrated, controversy remains . 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 cellulase-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 hydrolysis 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 proportional 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 endoglucanases on both types of substrate . Furthermore, the degree of crystallinity of cellulose influences adsorption at a given protein loading and the maximum adsorption constant was shown to be greatly enhanced at low crystallinity indices . The same study concluded that the effective binding was the limiting 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 investigate cellulase activity [35,47–51]. Treatment with 85% phosphoric acid to produce phosphoric acid swollen cellulose (PASC) results in complete dissolution of the sample  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 narrow range of acid concentrations or mainly at low concentrations [46,55–57]. Recently, Zhang et al.  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 cellulose at acid concentrations of 0% and above 81%, respectively. The changes in hydrolysis rate, with variations in the degree of crystallinity as a result of treatment 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 13C-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 calibrated concentration. These pretreated cellulose samples were employed to investigate and elucidate the relationship between the degree of crystallinity, adsorption and the enzymatic hydrolysis rates.
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 cellulase . The presence of hemicellulose, and especially lignin, a strong adsorbent on cellulase, in lignocellulosics, 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 13C-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 β-glucosidase to remove product inhibition and fully convert cellobiose to glucose, and the initial hydrolysis rates were calculated in terms of the production of glucose after a 2 min reaction time. As expected, the more concentrated the phosphoric acid solution, the higher the sugar production (Fig. 1A), so that the pretreatment procedure was considered to be efficient. Samples treated with pure phosphoric acid solution (maximum 85%) resulted in amorphous cellulose as demonstrated by X-ray diffraction analysis  (Fig. 2). Furthermore, a high amount of glucose (4.75 g·L−1·min−1) was produced from the cellulose sample pretreated with the highest concentration 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 . A recent study using phosphoric acid to increase cellulose accessibility in lignocellulosics suggested the presence of a critical point in the phosphoric acid concentration below which enzymatic hydrolysis was slow, and above which cellulose was easily dissolved . The results obtained in the present study (Fig. 1) confirm that there is a steep change in reactivity (i.e. glucose production) from 1 to 4.75 g·L−1·min−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 concentration, close to the 81% obtained by Moxley et al.  for maximum glucan digestibility. Below 75%, the glucose production rate tends to level off, with a minimum being obtained with untreated Avicel (0.6 g·L−1·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  (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 reflection at 2θ = 18°, which, upon our analysis, is definitely 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 acquisition 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 .
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 concentrations but decreases linearly to almost 0% in a concentration range of 75–80% phosphoric acid. Thus, the phosphoric acid effect is clearly evident: not only is it related to dissolution capacity , but also it disrupts the crystalline structure of cellulose and can turn partially crystalline cellulose amorphous. Avicel, a microcrystalline type of cellulose, has a mixed composition (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 cellulose 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 concentration range. Upon treatment at higher acid concentrations, 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 controversial, studies on the change of cellulose crystallinity 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 different types of substrate as well as the analytical methods 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 .
The CrI of Avicel was monitored via X-ray diffraction during its hydrolysis by a commercial mixture of cellulases from T. reesei and an excess of β-glucosidase to prevent cellobiose inhibition. The X-ray diffraction data obtained gave an artificially high degree of crystallinity for untreated Avicel (92%) using the method of Segal et al. . Small variations at such high values are challenging to monitor; therefore, cross polarization/magic angle spinning (CP/MAS) 13C-NMR spectroscopy was employed as an alternate method. The CrI of untreated Avicel (calculated as described previously)  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), (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 β-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 studies were conducted using cellulose samples generated with various amounts of phosphoric acid and thus displaying intermediate degrees of crystallinity (Fig. 1B). Adsorption experiments were carried 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 adsorption profile at 4 °C was found to be similar to that at 50 °C after 30 min . 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 .
Adsorption experiments were first performed using the same degree of loading as employed during a common enzymatic hydrolysis run (175 μg·mg−1 cellulose; Figs 1–3). Surprisingly, a maximum value of adsorbed enzyme concentration (150 μg·mg−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 appeared to increase inversely and linearly with the CrI at higher crystallinity values (i.e. > 45%). A constant amount of adsorbed enzymes (∼ 150 μg·mg−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 capacity increased and was linearly proportional to the initial rate.
At higher enzyme loading (seven-fold greater than the original loading; i.e. 1230 μg·mg−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 crystallinity. By contrast, untreated Avicel (CrI = 60%) displayed similar rates at both enzyme concentrations, and little difference in rate for the two enzyme concentrations 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%.
Cellulase hydrolysis rate and cellulose crystallinity
The correlation between the CrI and the initial hydrolysis 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. The latter is supported by the data obtained from moisture content measurement (Fig. 1C). Most aqueous 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 . 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 . A highly crystalline cellulose 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 evaluation of the concentration range corresponding to the sigmoid region. In their review, Zhang et al.  stressed that the CrI of cellulose was not strongly associated with hydrolysis rates. By contrast, the results obtained in the present study show a very close and linear relationship between the CrI and initial hydrolysis rate for samples of same origin obtained after pretreatment with phosphoric acid (R2 = 0.96; Fig. 5), demonstrating 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 pretreatment 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 intermediate degrees of crystallinity allows the exclusion of additional parameters that might influence the enzymatic action on cellulose, such as the type and source of cellulose or mixed components, and yields an explicit proof of the tight relationship between initial cellulose crystallinity 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 interpretation 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 crystal 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 crystallinity 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 initial rate; therefore, it does not appear physically possible 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 beginning of hydrolysis. Factors other than crystallinity impeding enzymatic action (both enzyme- and substrate-related) require closer attention.
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 determining the role of crystallinity in enzymatic activity, it is logical to ask whether crystallinity might not be masking another phenomenon, specifically adsorption. A constant adsorption profile at different enzyme concentrations was found to relate to increasing hydrolysis 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 . 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 . At a very low CrI and constant adsorbed enzyme concentration, the percentage of surface coverage is smaller because the surface area is larger at lower crystallinity indices . 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 reaction time). Accessibility was suggested to be an important factor that affects enzymatic hydrolysis rates  and its increase at lower degrees of crystallinity was proposed as a reason for enhanced digestibility . It has also been suggested that rendering the substrate more amorphous increases access to the reducing ends of cellulose, thus enhancing reaction rates . 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 concentration. This can be related to recent work demonstrating that the overcrowding of enzymes on the cellulose surface lowers their activity . 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 ; 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 accessible to enzymes, leading to such low adsorption, possibly in contrast to more amorphous samples. An accessible surface area has been the subject of numerous studies  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 hydrolysis rates.
Avicel hydrolysis rates were not significantly changed upon the addition of a much higher enzyme concentration for samples displaying a degree of crystallinity in the range 60–50% (Fig. 6C), demonstrating 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 μg·mg−1 cellulose) resulted in saturation of the Avicel surface, whereas low enzyme loading (175 μg·mg−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 μg·mg−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 cellulases 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
All chemicals and reagents were purchased from Sigma (St Louis, MO, USA) unless otherwise stated. Avicel PH-101, cellulases from T. reesei (159 FPU·mL−1) and β-glucosidase (from almonds, 5.2 U·mg−1) were obtained from Sigma and phosphoric acid (85%) was obtained from EMD (Gibbstown, 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 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 g·L−1) in sodium acetate buffer (1 mL, 50 mm, pH 5) was hydrated for 1 h with stirring at 50 °C. β-Glucosidase (15 kU·L−1) and cellulases (24 mL·L−1, 3.4 g·L−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 supernatant was measured via the dinitrosalicylic acid (DNS) assay. For crystallinity measurements at various conversion levels using CP/MAS 13C- 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 cellulase component, 92 μg 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 . 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 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/Kα1 irradiation (1.54 Å) at 45 kV and 40 mA. The scan speed was 0.021425°·s−1 with a step size of 0.0167°. CrI was calculated using the peak intensity method :
where I002 is the intensity of the peak at 2θ = 22.5° and Iam is the minimum intensity corresponding to the amorphous content at 2θ = 18°.
Freeze-drying showed no impact on the crystallinity of untreated Avicel.
Solid state 13C-NMR
The solid-state CP/MAS 13C-NMR experiments were performed on a Bruker Avance/DSX-400 spectrometer (Bruker Instuments, Inc., Bellerica, MA, USA) operating at frequencies of 100.55 MHz for 13C. 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 μs pulse and a 2.0 ms contact pulse over 4 h. CrI was calculated according to standard methods :
where A86–92 p.p.m. and A79–86 p.p.m. are the areas of the crystalline and amorphous C4 carbon signal of cellulose, respectively.
Oven-drying (60 °C) showed no impact on the crystallinity of untreated Avicel.
Multivariate statistical analysis of X-ray data
The CrI of cellulose samples was also calculated by quantifying the contribution of amorphous cellulose (PASC) and Avicel to its (normalized) X-ray diffraction spectra :
where Ij (2θ) is the intensity of the jth sample at diffraction angle 2θ, Ip (2θ) is the intensity of PASC at diffraction angle 2θ, IC (2θ) is the intensity of untreated Avicel at diffraction angle 2θ, fj is the contribution of PASC to the spectrum and ε is the random error.
, the least square estimate of fj, was used to estimate the crystallinity by multiplying the contribution of Avicel by its crystallinity (calculated by CP/MAS 13C-NMR as 60%):
where CrIj is the crystallinity (in percentage) of the jth sample of Avicel AND CrIc is the crystallinity of Avicel (calculated by CP/MAS 13C-NMR as 60%).
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: (NH4)2SO4 5 g·L−1, CaCl2 0.6 g·L−1, MgSO4 0.6 g·L−1, KH2PO4 15 g·L−1, MnSO4.H2O 1.5 mg·L−1, FeSO4.7H2O 5 mg·L−1, COCl2 2 mg·L−1, ZnSO4 1.5 mg·L−1) supplemented with glucose (2%). After 3 days at 28 °C and 150 r.p.m., the fungus was grown on lactose (2%) in minimal medium for up to 12 days at 28 °C and 150 r.p.m. After filtration over glass-microfiber filter (1.6 μm GF/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 concentrations were estimated by the Bradford assay, using BSA as standard.
Cellulose samples (20 mg·mL−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).
Chevron Corporation is acknowledged for their funding. Dr J. Leisen and Dr J. I. Hong are thanked for their technical assistance with the crystallinity measurements.