Anomeric Selectivity and Product Profile of a Processive Cellulase

Dec 1, 2016 - The moment of the dissolution was taken as zero time for the aging of the G2β preparation. For the preparation of α-cellobiose (G2α),...
0 downloads 0 Views 1MB Size
Subscriber access provided by UNIV OF REGINA

Article

Anomeric selectivity and product profile of a processive cellulase Jeppe Kari, Riin Kont, Kim Borch, Steen Buskov, Johan Pelck Olsen, Nicolaj Cruys-Bagger, Priit Väljamäe, and Peter Westh Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.6b00636 • Publication Date (Web): 01 Dec 2016 Downloaded from http://pubs.acs.org on December 12, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biochemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Anomeric selectivity and product profile of a processive cellulase Jeppe Kari1, Riin Kont2, Kim Borch3, Steen Buskov3, Johan Pelck Olsen1, Nicolaj Cruyz-Bagger3, Priit Väljamäe2 and Peter Westh1* 1

Research Unit for Functional Biomaterials, Roskilde University, Denmark, 2Institute of Molecular and Cell Biology, University of Tartu, Estonia, 3 Novozymes A/S, Krogshøjvej 36, DK-2880, Denmark.

Supporting Information Placeholder ABSTRACT: Cellobiohydrolases (CBHs) make up an important group of enzymes both for natural carbon cycling and industrial deconstruction of lignocellulosic biomass. The consecutive hydrolysis of one cellulose strand relies on an intricate pattern of enzyme-substrate interactions in the long, tunnelshape binding site of the CBH. In the current work, we have investigated the initial complexation mode with cellulose of the most thoroughly studied CBH, Cel7A from Hypocrea jecorina (HjCel7A). We found that HjCel7A predominantly produce glucose when it initiates a processive run on insoluble microcrystalline cellulose, confirming the validity of even and odd product ratio as an estimate of processivity. Moreover, the glucose released from cellulose was predominantly α-glucose. A link between the initial binding mode of the enzyme and the reducing end configuration was investigated by inhibition studies with the two anomers of cellobiose. A clear preference for β-cellobiose in product binding site +2 was observed for HjCel7A, but not the homologous endoglucanase, HjCe7B. Possible relationships between this anomeric preference in the product-site and the prevalence of odd-numbered initial-cut products are discussed and a correlation between processivity and anomer selectivity is proposed.

Introduction Cellulose strands can be arranged in a crystalline structure, which is stabilized by both hydrogen bonding and hydrophobic interactions,1, 2 and this makes the glycosidic bonds that connect pyranose units difficult to access for hydrolytic enzymes. Thus, cellulases must be able to dislodge a cellulose chain from the crystal and keep it detached while performing the catalysis.3 One particularly elegant way to achieve this is seen for processive cellobiohydrolases (CBHs). These enzymes remain attached to the strand after completion of a catalytic reaction, and thus performs consecutive hydrolyses while keeping the strand separated from the crystal.4, 5 This mechanism has proven effective for the breakdown of crystalline cellulose and CBHs make up the major fraction of enzyme in both secretomes of wood decaying fungi and commercial cocktails for deconstruction of lignocellulosic biomass.6 On a structural level, the processive mechanism is linked to a long, tunnel-shaped binding site with numerous subsites for pyranose binding. 7-11 This design appears intuitive for the ability of the enzyme to remain associated with the substrate after completion of a hydrolytic reaction, but details of the underlying interactions and their role for enzyme activity remain elusive. Much progress in this area has been made through structural, 12-18 biochemical 5, 8, 9, 19-37 and computational 38-41 studies. However, molecular level understanding of processivity is still incomplete, and this is illustrated by protein engineering studies, where severe effect on processivity has been observed by simple point mutations in either the catalytic domain 7, 8, 29, 30, 42 or the carbohydrate binding module.30 In rough terms, three molecular steps exist for a processive cellulase: adsorption, processive hydrolysis and dissociation. Both the processive hydroly1

ACS Paragon Plus Environment

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 20

sis 25, 37, 43 and the dissociation8, 19, 27 has been investigated in some detail so in the current study we have focused on the initial hydrolytic event of CBH Cel7A from Hypocrea jerorina (HjCel7A) on cellulose. We developed an experimental technique based on 18O labeling to examine which type of product HjCel7A made when it initiated a processive run. We found that HjCel7A mainly produced saccharides with an odd number of pyranose rings (cellotriose and particularly glucose) in the initial cut of a processive run. This is the first characterization of the initial cut product pattern on real cellulosic substrate, and the result provide the first experimental validation for using even and odd product ratio as an estimate of processivity on insoluble cellulose. The molecular origin of odd-numbered product initiation was investigated by analyzing enzyme-substrate interactions in the product site of HjCel7A. We found a distinctive preference for the β-anomer of the pyranose bound in subsite +2, and we suggest that this is at least part of the reason that the dominant initial-cut product of HjCel7A is odd-numbered. Experimental procedures Enzyme. The cellobiohydrolase and endoglucanase Cel7 from Hypocrea jecorina (HjCel7A and HjCel7B) were expressed heterologously in Aspergillus oryzae and purified as described elsewhere.44 The purity of the enzyme was assessed from the absence of other bands in SDS-PAGE and specific contamination by β-glucosidase (BG) was confirmed by the lack of activity against cellobiose. Enzyme concentrations were determined by absorbance at 280 nm using molar extinction coefficients of 84810 M-1cm-1 and 71910 M-1cm-1 for HjCel7A and HjCel7B respectively. HjCel6A was purified from the culture filtrate of Hj QM 9414 as described previously32, 45 and a molar extinction coefficients of 71910 M1 cm was used to determine the enzyme concentration from absorbance measurement at 280nm. Inhibition by sugar anomers. Inhibition of HjCel7A (0.5 µM) and HjCel7B (0.5 µM) by cellobiose (G2) and glucose was measured with 5 µM 4-methyl-umbelliferyl-β-D-lactoside (MUL) as substrate. Reactions were conducted in 50 mM sodium acetate buffer pH 5 (supplemented with 0.1 mg/ml BSA) at 25°C for 30 s and stopped by the addition of equal volume of 0.2 M ammonium hydroxide. The activity was determined by measuring the fluorescence of released 4-methyl-umbelliferone. For inhibition, the reactions were supplemented with G2 (initially as β- or α-anomer) preparations with different age. For the preparation of β-cellobiose (G2β) the G2β solids (Fluka 22150) were rapidly dissolved (it took about 30 s to dissolve the solids) in 50 mM sodium acetate buffer pH 5. The moment of the dissolution was taken as zero time for the aging of G2β preparation. For the preparation of α-cellobiose (G2α) the hydrolysis of bacterial microcrystalline cellulose (BMCC) by the inverting CBH, HjCel6A was used. BMCC (0.6 mg/ml) was incubated with HjCel6A (1 µM) in 50 mM sodium acetate buffer pH 5 at 0°C for 40 min. BMCC was separated by centrifugation (2 min, 104 × g) and the supernatant was used as G2α preparation. The moment of the removal of supernatant from BMCC pellet was taken as zero time for the aging of G2α preparation. The MUL hydrolyzing activity of HjCel6A was judged to be negligible. In inhibition with G2 being initially G2β the total concentration of G2 was 20 µM and 7 mM in experiments with HjCel7A and HjCel7B, respectively. In inhibition with G2 being initially G2α the total concentration of G2 (27.5 µM) was found from the inhibition of HjCel7A by G2α preparation that was allowed to equilibrate for mutarotation and using the Ki value of 18 ± 3 µM measured by us before.46 Product-profile – soluble substrates. Cellooligosaccharides (COS) with a degree of polymerization of 4 to 6 (Megazymes, Ireland) were incubated with 0.1 µM HjCel7A in 50 mM sodium acetate buffer pH 5 at 25 °C for 1 h. Experiments were made using COS concentrations of 200 µM, 300 µM, and 400 µM. The reactions were quenched by the addition of ammonium hydroxide to 0.1 M. The enzyme was removed by ultrafiltration using Vivaspin 500 centrifugal device with 3.0 kDa cut-off PES membrane. The filtrate was neutralized with acetic acid and the sugars were concentrated using vacuum evaporator operating at 45 °C. Sugars were separated using HPLC with Aminex HPX-87P column and refractive index detection as described before.47 Product profile – insoluble substrate. The time-dependent concentration of glucose, cellobiose, cellotriose and cellotetraose was measured during 60 min hydrolysis reactions with 500 nM HjCel7A and 50 g/L Avicel in 50 mM acetate buffer, 2 mM CaCl2, pH 5. At selected time points, samples of 500 µl ACS Paragon Plus Environment 2

Page 3 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

were taking from a 50 ml, stirred reaction mixture kept at 25.0°C. Samples were immediately quenched with an equal volume 0.1 M NaOH and following centrifugation the concentrations of glucose, cellobiose, cellotriose and cellotetraose in the supernatants were measured on a Dionex ICS-5000 ion chromatograph (Termo Fisher Scientific Waltham, MA) as described elsewhere.8, 48 Biosensor measurement. The release of β-glucose during hydrolysis Avicel was monitored in realtime by amperometric biosensors based on glucose oxidase from Aspergillus niger (AnGOX). Sensors were based on mediator modified benzoquinone carbon paste with surface immobilized AnGOX, as described in detail previously.49 The reaction was started by addition of HjCel7A to a 5 ml waterjacketed glass cell with a stirred (600 rpm) suspensions of 50 g/L Avicel PH 101 (Sigma Aldrich). The enzyme was injected to a final concentration of 0.5 µM from a Chemyx Fusion 100 syringe pump and concentration of β-glucose was followed in real-time during 40 min hydrolysis. Two separate control experiments with 100 µM α- or β-glucose were done to illustrate the anomeric specificity of the sensor and estimate the rate of mutarotation for glucose. In these controls, the dry glucose was dissolved in 50 mM acetate buffer, 2 mM CaCl2, pH5 and immediately injected into the glass cell, to limit mutarotation during sample handling. All measurements were made in duplicate. 18 O-labelling. All experiments were done using microcrystalline cellulose Avicel PH 101 (Sigma Aldrich) as substrate and 50 mM acetate buffer, 2 mM CaCl2, pH 5. To remove any traces of soluble oligosaccharides the substrate was washed three times with milli-Q water. To remove almost all unlabelled water from the substrate, the washed Avicel was dried at 90°C overnight before use. Two different reactions, one labelled and one control, were then prepared. The former used H218O/buffer (Sigma Aldrich) as solvent while the the latter was made in the standard Milli-Q buffer with normal water. Each reaction was done in triplicate. Before starting the reaction, 0.01 g washed and dried Avicel was transferred to a 1.5 ml Eppendorf tube, together with 192 µL MilliQ water or 97% H218O. Four µL of highly concentrated acetate buffer (2.5 M acetate buffer, 0.1 M CaCl2, pH 5) was added. All samples were then preincubated 5min in a thermomixer at 25°C and 1400 RPM. To start the reaction 4 µL of concentrated HjCel7A stock (25 µM) was added to reach a final enzyme- and substrate concentration of respectively 0.5 µM and 50 g/L. The reaction was allowed to run for 30 min at 25°C and 1400 RPM before it was quenched by filtration (Kinesis KX syringe filter, PTFE 13 mm, 0.45 µm). The supernatant was diluted 1:10 in Acetonitrile with 10% water (v/v) before LCMS analysis. The enzyme to substrate ratio and incubation period was chosen so it mimics the single hit condition found by Kurasin and Valjamae 27 This condition has been shown to minimize the possibility that the same chain is attacked twice during the hydrolysis. Exchange of the anomeric oxygen was tested by incubating 10 ppm glucose, cellobiose or cellotriose with 97% H218O for 1 h under the same condition as the hydrolysis experiment. LCMS analysis. All hydrolysates (controls and 18O labelled) samples were analyzed on a triplequadropole LC-MS instrument from Waters (Xevo TQ-S micro®). The injection volume was 20 µL and the samples were ionized in a turbo electrospray running in negative mode. The separation of glucose, cellobiose and cellotriose were done on a synergy 4u Fusion-RP 80A, 100x2.0 mm column (Phenomenex) held at constant temperature (40C°). The flow rate was 200 µL/min and the eluents were A (95% acetonitrile, 5% milliQ water, 5 mM ammonium acetate) and B (MilliQ water, 5 mM ammonium acetate). The column was equilibrated for 2 min with 95% B followed by a 4 min gradient elution toward 50% B. To ensure all oligosaccharides were eluted from the column, the flow with 50% B was continued isocratically for an additional 2 min. A re-equilibration step of 8 min with 95% B was performed between each injection. All samples were scanned with multiple reaction monitoring (MRM). This was chosen to decrease the general background noise and to decrease the isotope-signal from 13C-atoms. The de-clustering potential (DP), entrance potential (EP), collision cell entrance potential (CEP), collision energy (CE) and collision cell exit potential (CXP) were optimized by direct infusion. A standard curve with glucose, cellobiose and cellotriose was included in all measurement. The optimized parameters from the MS-analysis together with the m/z values of the parent- and product ion for all analyzed oligosaccharides is given in the supplementary material.

ACS Paragon Plus Environment 3

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 20

Results Product profile – Insoluble substrate. The concentrations of glucose, cellobiose, cellotriose and cellotetraose are plotted as a function of time in the upper panel of Fig. 1. It appears that there is a characteristic initial rapid phase (a so-called burst) for the tree main products (glucose, cellobiose and cellotriose), followed by a near-linear progress curve after 5-10 minutes. In the pseudo steady-state regime from 5-60 min the specific rate for the overall product formation was found to be ~ 4min-1 which is close to previously reported values of the maximal specific rate values (Vmax/E0) at 25°C.8, 19, 50 This supports the assumption that HjCel7A was saturated with substrate under the chosen experimental condition. Small amounts of cellotetraose also built up (inset) in the initial phase, but the concentration of this product fell to undetectable levels after about 20 min. As cellotetraose is an excellent substrate for HjCel7A 51, this behavior probably reflected that except during the initial burst, hydrolysis in the supernatant was fast enough to remove this oligomer. Plots of product ratios in the lower panel of the figure 1 revealed that both [cellobiose]/[cellotriose] and [cellobiose]/[glucose] rose gradually over the first 10-15 min and then leveled out at values of respectively 30 (right ordinate) and 6 (left ordinate).

Figure 1. Progress curves and product ratios. The upper panel shows the concentration of glucose (right ordinate) and cellobiose (left ordinate), cellotriose (right ordinate) and cellotetraose (inset) produced by HjCel7A (0.50 µM) during hydrolysis of Avicel (50 g/l). The lower panel shows the cellobiose/cellotriose and cellobiose/glucose concentration ratios calculated from the results in the upper panel. Product profile – soluble substrate. Product concentrations for the hydrolysis of cellooligosaccharides (COSs) with a degree of polymerization (DP) from 4 to 6 are shown in Table 1. Earlier work has demonstrated that Michaelis constants, KM, for HjCel7A acting on COS with a DP over 4, is in the low µM range,51 and as the current measurements were conducted at much higher substrate concentration (200 µM), the enzyme will be largely saturated with substrate throughout these experiments (hydrolysis will progress at the maximal rate, Vmax). This interpretation was confirmed in control experiments with higher substrate concentrations (300 µM and 400 µM), which showed no difference in hydrolytic rate and product profile. Consistent with the processive mode of action of HjCel7A we did not observe release of intermediate products (G4 in the case of G5-substrate, or G4 and G5 in the case of G6substrate). In the current context, the most important result in Tab. 1 is the probability of the product being odd-numbered, Podd. This parameter was calculated both from the current results and an earlier data set for COS hydrolysis by HjCel7A.51 A detailed description of how the observed product profile is converted to Podd is given in supporting information. The resulting values, listed in Tab. 1, show that Podd increases with the DP of the substrate and tends to level off around 0.8 for the longer COSs. ACS Paragon Plus Environment 4

Page 5 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Table 1. Product concentrations and probability of odd-numbered initial product, Podd, for HjCel7A hydrolyzing cellooligosaccharides with different degree of polymerization. Cellooligosaccharides are abbreviated GX, where X is the degree of polymerization. Product (µM) Probability COS G1 G2 G3 Podd 1 G4 19.6 108.8 n.a. 0.27; 0.333 ±1.3 ±7.5 G5 30.7 98.5 36.9 0.63; 0.733 ±1.1 ±5.1 ±7.4 G6 38.1 96.0 67.6 0.74; 0.803 ±3.8 ±1.9 ±3.7 2 2 G7 n.a. n.a. n.a.2 0.853 G8 n.a.2 n.a.2 n.a.2 0.803 1 Because of the large peak of G4 substrate the G3 was not sufficiently separated from G4 in this experiment. 2Not measured in the current work. 3Calculated from literature data51. Initial product profile – 18O labeling. We implemented an 18O labeling procedure to enable distinction of initial- and processive cuts in unmodified solid substrate (Avicel). Thus, if the enzymatic reaction is conducted in H218O, HjCel7A will produce an initial product with an unlabeled (16O) reducing end, while subsequent, processive cuts will make 18O labeled cellobiose. It follows that initial- and processive cuts can be distinguished, even if they both produce cellobiose. This principle is illustrated in Fig. 2. A major advantage of 18O labeling is that it uses an unmodified substrate. However, like other approaches, the methodology has a number of limitations. One concern is that the number of initial attacks may be underestimated if instead of attacking the cellulose chain end, the enzyme initiates a processive run by an endo-hydrolytic mechanism. The same problem occurs if the same chain is attacked twice during the experiment (even if these attacks are exo-hydrolytic). Both scenarios lead to a labeled initial product, which is indistinguishable from processive products. However, both limitations can be mitigated by the application of high substrate to enzyme ratio and/or short contact times. Based on this, we chose experimental conditions (see Materials and Methods) where systematic errors from both endoand double attacks were deemed unimportant. This was documented in control experiments described in the supporting information. The use of high substrate load also diminished secondary hydrolysis of cellotriose in the bulk, which could potentially contaminate the unlabeled and labeled glucose and cellobiose. Another concern of the method was the risk of exchanging the oxygen on the anomeric carbon of the unlabeled product with the heavy 18O-labelled oxygen in H218O. We tested this by incubating unlabeled glucose, cellobiose or cellotriose in 18O-labeled water and found no exchange on a time scale relevant for our measurement (see supporting information). We conclude that systematic errors from these effects are small and that the method appears to provide a reasonable distinction of initial and processive cuts. More details regarding this are discussed in the supporting information. Representative results from the LCMS-measurements are shown in Fig. 3. As expected, it appears (right panel) that hydrolysis in 93% (mol/mol) H218O produces a large excess of labeled cellobiose, but that a measurable amount of unlabeled saccharide is also identified. A very small amount of product with a molecular weight like 18 O-cellobiose was observed in the hydrolysate made in normal water (left panel). This may reflect experimental uncertainties or small amounts of cellobiose with two 13C isotopes. The experimental procedure involved the addition of enzyme stock dissolved in normal-water buffer, and hence a reduction in the isotopic purity of the reaction buffer (in most trials, the final purity was 93% 18O). We corrected the data for this by assuming that 7% of processive cuts would lead to unlabeled cellobiose. Results after this correction are listed in Tab. 2 together with results from a parallel control experiment made with unlabeled water.

ACS Paragon Plus Environment 5

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 20

Figure 2. Illustration of the product labeling technique. Because water is added to the newly formed reducing end, the initial cut product will be unlabeled while processive cut will be 18O-labeled if the hydrolysis takes place in H218O.

Figure 3. Ion-spray mass chromatogram of cellobiose (black line) and 18O-cellobiose (red line) using water (left) as solvent or 93% H218O (right). Enzyme concentration was 0.5 µM HjCel7A and substrate concentration was 50 g/L Avicel. Incubation time was 30 min at 25°C and pH 5. It appears that the total product concentrations were similar for the two experiments (and also similar to the results after 30 min in Fig. 1). This confirms earlier suggestions52 that 18O did not alter enzyme activity. More importantly, Tab. 2 shows that the vast majority (~83%) of initial cuts leads to the formation of an odd-numbered product. This conclusion is reached by considering the proportion of oddunlabeled product in table 2, and it is in line with Podd derived from hydrolysis of COS (Tab.1). It is also interesting to note that glucose is by far the most prevalent unlabeled product, and hence that HjCel7A typically starts a processive run by releasing the monomer. One consequence of this is that the (moderate) occurrence of cellotriose in the hydrolysate (Fig. 1) is not or at least only to a small extent, the result of initial cuts (as the concentration of unlabeled cellotriose is very low). Measuring anomeric selectivity in product binding sites. Anomeric selectivity in product binding sites of HjCel7A was assessed by measuring the inhibition strength of respectively α- and β-cellobiose ([G2α] and [G2β]) during hydrolysis of the soluble substrate analog MUL. Cellobiose binds in the product-site of HjCel7A and inhibit the hydrolysis of MUL with a competitive mechanism,5, 53, 54 which make this an ideal system to investigate the difference in binding strength of α- and β-cellobiose. We used a MUL concentration of 5 µM, which is much lower than the reported KM for MUL53 to ensure that the inhibitor was not outcompeted by the substrate in the experiment; i.e. [MUL] > KiG2β in Eq 1, the KiG2β should equal 2/3 KiG2α/β which is 12 ± 2 µM. The latter value corresponds within error limits with the KiG2β values found from two inhibition experiments (Fig 4A). Collectively these results suggest that binding of G2α to HjCel7A is weak compared to that of G2β and can be neglected in the analysis of G2 inhibition. Next, we tested the anomeric selectivity of G2 inhibition of endoglucanase HjCel7B. Since the KiG2α/β for HjCel7B is high (see supporting information) only the inhibition with G2 being initially G2β can be tested in practice. No change in inhibition strength was observed with aging of the G2β preparation (Fig 4B) indicating no anomeric selectivity in the product binding sites of HjCel7B.

Figure 4. Inhibition of HjCel7A (A) and HjCel7B (B) by cellobiose anomers. Shown is the dependency of the inhibition strength of MUL hydrolysis (vi/v0) on the age of cellobiose preparation being initially α-cellobiose (black symbols) or β-cellobiose (red symbols). Solid lines represent the best-fit according to the Eq 2. Anomeric composition of glucose product. Results in Tab. 2 show that glucose was the dominant first-cut product in hydrolysis of cellulose. To elucidate its anomeric composition, we applied an amperometric biosensor with high sensitivity, time resolution and specificity for b-glucose.49, 59 The ability of the sensor to distinguisha-and b-glucose is illustrated in Fig 5A. It appears the sensor showed a strong response from b-glucose, while essentially no initial signal was recorded for the a anomer. Over time, mutarotation equilibrated both samples to an [b]: [a]ratio around 2 in line with literature values.60 Application of a simple first-order fit suggested a rate constant of about 5 x 10-4 s-1 for the decay towards equilibrium and this is in accord with both the inhibition experiments discussed above and earlier reports.49 Results from biosensor analysis of HjCel7A activity under conditions identical to the 18O labeling experiments (Tab. 2) are shown in Fig 5B. The output from the biosensor (showing the concentration of b-glucose) is compared to chromatographic measurements quantifying all glucose (the sum of a- and b-anomers). It is clear that almost no b-glucose occurs over the first few minutes of hydrolysis, and this is in sharp contrast to total glucose, which shows a distinctive burst phase. After 2 min, for exACS Paragon Plus Environment 8

Page 9 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

ample, we found undetectable amounts of b-glucose (less than 0.3 µM), while the total glucose (and hence a-glucose) reached 5 µM. b-glucose only built up with much delay, and we suggest that the monomeric product made directly by Cel7A on Avicel was almost exclusively a-glucose. This implies that the b-glucose found at later stages of the hydrolysis in Fig. 5B mainly occurs as a result of mutarotation of the originally released a anomer. This interpretation finds some support in the comparable time scales for changes in the two panels of Fig. 5. When this interpretation is combined with the data in Tab. 2 (dominance of glucose over other initial-cut products) we arrive at the somewhat surprising conclusion that by far the most common initial-cut product of HjCel7A on Avicel is a-glucose.

Figure 5. Results from biosensor analyses. A) Control experiment showing the sensitivity of the sensor to respectively α and β glucose and the time-course of mutarotation. The signals were recorded for solutions that were initially either 100 µM α-D-glucose or β-D-glucose. Further details about the sensor can be found in supporting information. B) The amount of total glucose (black symbol measured by chromatography) and β-glucose (red line measured by the biosensor) produced by 0.50 µM HjCel7A acting on 50 g/l Avicel at 25°C, pH 5.0

Discussion Cellobiohydrolases from Gylcoside Hydrolase family 7 are processive enzymes and hence conduct consecutive hydrolytic events on one cellulose strand. They utilize a retaining mechanism and this together with information from crystal structures and molecular dynamic simulation, 14, 43, 61 suggests that consecutive hydrolytic steps exclusively produce b-cellobiose. This interpretation is in line with many activity studies, which have reported a dominance of cellobiose in hydrolysates obtained under a range of experimental conditions.5, 26, 34, 51, 62 Nevertheless, odd-numbered products, particularly glucose and cellotriose, have been repeatedly reported.5, 34, 51 In contrast to cellobiose, the origin of these oddnumbered products is not understood in detail. Most discussions of this question have focused on the 180° flip of successive pyranose moieties (two-fold screw axis) in the cellulose chain and the associated “zigzag pattern” of the connecting glycosidic bonds.5, 12, 23, 26, 34, 51, 52, 63, 64 Since the tight tunnel with numerous interactions to the cellulose chain does not allow the chain to flip 180 degree,14 only every second bond will be productively orientated for hydrolysis. Hence, odd-numbered products may occur either in the initial cut or as the final product, when a cellulose chain is fully hydrolyzed. In the current work, we have conducted a number of complementary experiments for the purpose of elucidating origins of odd-numbered products. Our main motivation is to use this information in discussions of enACS Paragon Plus Environment 9

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 20

zyme-substrate interactions in the product site of HjCel7A, and hence elucidate some aspects of the complex processive mechanism. Before turning to mechanistic issues, we will consider a more methodical aspect of the work, namely how the 18O labeling technique may be used in assessments of the processivity number. The processivity number, N (i.e. the average number of hydrolytic events before dissociation) is often a very useful parameter in discussions cellulase activity, but it has proven difficult to measure. The most widespread experimental approach is to use some ratio of even- and odd-numbered products as an indicator of N.5, 8, 20, 21, 23, 29, 34 This idea is attractive because it allows estimation of N on the basis of standard chromatographic measurements. The approach relies on the assumption that odd-number products are only made from initial cuts, but experimental justification of this premise remains essentially absent and no agreement has been reached on exactly which ratio of glucose, cellobiose and cellotriose should be used.65 Unfortunately, other biochemical methods for estimation of N for cellulases are experimentally demanding and therefore rarely used.65 In the current work, we used 18O labeling to identify initial-cut products and this provides some insight into relationships between the number of odd-numbered products and initiations of processive sweeps. Thus, as illustrated in Fig. 2, 18O labeling of the reducing end is expected for all except the initial-cut product (with the experimental provisos discussed earlier). It follows that N will be the ratio total product/un-labeled product. For the current system (HjCel7A/Avicel) the data in Tab. 2 suggest N=9. A direct implication of this is that 18O labeling combined with LCMS provides a useful method for determination of N for cellulases that attack the reducing end of cellulose. However, like most earlier theoretically sound methods, 65 18O-labeling/LCMS is quite labor intensive. Therefore, it is relevant to use the combined product profile and labeling information in Tab. 2 to identify dominant initial-cut products as this could potentially “calibrate” the more experimentally convenient chromatographic approach discussed above. Interestingly, inspection of Tab. 2 shows that initial-cut products are dominated by glucose. Thus, this product makes up about 80% of the total unlabeled compounds in Tab. 2. Moreover, labeled glucose is quite scarce (