Research Article Cite This: ACS Sustainable Chem. Eng. 2019, 7, 1113−1119
pubs.acs.org/journal/ascecg
Use of Carbohydrate Binding Modules To Elucidate the Relationship between Fibrillation, Hydrolyzability, and Accessibility of Cellulosic Substrates Kevin Aïssa,† Vera Novy,† Fredrik Nielsen, and Jack Saddler* Forest Products Biotechnology and Bioenergy Group, Department of Wood Science, Faculty of Forestry, The University of British Columbia, 2424 Main Mall, Vancouver, British Columbia V6T 1Z4, Canada
ACS Sustainable Chem. Eng. 2019.7:1113-1119. Downloaded from pubs.acs.org by IOWA STATE UNIV on 01/15/19. For personal use only.
S Supporting Information *
ABSTRACT: It is widely acknowledged that the rate limiting step in the enzyme-mediated deconstruction of the biomass process is the restricted ability of the enzymes to access the cellulosic substrate. An ongoing challenge has been to find reproducible and quantifiable methods for measuring enzyme accessibility to cellulose. Type A (crystalline cellulose) and type B (paracrystalline) cellulose binding modules (CBMs) were used in parallel with microscopy, fiber analysis (aspect ratio), and water retention values (WRV) to determine if the observed and anticipated changes in differentially prepared microfibrillated cellulose (MFC) substrates were similar. It was apparent that with increasing refining there was a corresponding increase in fibrillation (SEM and WRV), as well as a decrease in aspect ratio. Although the initial degree and rate of enzymatic hydrolysis increased with prolonged refining, above 1000 kWh ton−1 little improvement in either was observed. However, when cellulose accessibility was assessed by the CBM method, the observed trend followed the hydrolysis profile. Although the other methods (WRV, SEM, and aspect ratio) suggested increased refining should result in greater accessibility and a corresponding improvement in hydrolysis, the CBM method more accurately predicted enzyme accessibility, implying that refining did not significantly improve enzyme accessibility at the microfibril level of the cellulosic substrate. KEYWORDS: Carbohydrate binding modules (CBMs), Cellulose accessibility, Microscopy, Enzyme mediated hydrolysis, Fibrillation
■
with the refining energy, resulting in increased fiber swelling due to internal fiber fibrillation and delamination.10,11 Carbohydrate-binding modules (CBMs) have recently been advocated as a promising tool to characterize cellulosic substrates and measure cellulose accessibility,12 because it can overcome many of the shortcomings of the other methods. The diversity of CBMs and their specificity, stability, and affinity make them ideal probes to investigate the microfibril structure of cellulosic substrates. In nature, CBMs occur mainly as the substrate recognition domain of carbohydrate active enzymes produced by bacteria and fungi. It has been shown that type A CBMs preferentially bind to crystalline cellulose, whereas type B CBMs predominantly bind to glucan chains and, therefore, substrate regions associated with a paracrystalline organization; type C CBMs bind predominantly to smaller oligosaccharides.13 A summary of CBMs used as probes is given in Table 1.
INTRODUCTION One of the major impediments limiting the commercialization of enzyme-mediated processes for conversion of biomass to sugars, fuels, and chemicals is the relatively poor efficacy of the enzymes and proteins that are required for the hydrolysis of the cellulosic fraction. Although the efficiency of hydrolysis is often measured by the amount of sugar released, it is generally acknowledged that it is the restricted enzyme accessibility to the cellulosic substrate that is the rate-limiting step.1 In-depth analysis of cellulose accessibility to enzymes (CAE), however, is challenged by the lack of quantifiable, reproducible, and relatively easy methods that allow quantification of CAE.2 Previous work has used methods such as water retention,3 microscopy,4−6 and fiber characterization (e.g., aspect ratio)7 as ways of trying to predict cellulose accessibility and, consequently, effectiveness of enzyme-mediated hydrolysis. However, these methods cannot elucidate the possible changes at the microfibril level of cellulose, which directly affect enzyme behavior.8 The water retention value (WRV) measures the ability of a cellulose fiber to retain water, and it is commonly used by the pulp and paper industry to characterize the extent of fiber fibrillation.9 The WRV of a fiber typically increases © 2018 American Chemical Society
Received: September 18, 2018 Revised: November 14, 2018 Published: December 12, 2018 1113
DOI: 10.1021/acssuschemeng.8b04780 ACS Sustainable Chem. Eng. 2019, 7, 1113−1119
Research Article
ACS Sustainable Chemistry & Engineering
Table 1. Summary of Previous Work That Has Used Cellulose Binding Module (CBM) Adsorption To Try To Better Elucidate Enzyme-Mediated Changes in Cellulosea CBM CBM2a(H6) (C. f imi) CBM3 (C. thermocellum) CBM4−1 (C. f imi) CBM15 (C. japonicus) CBM17 (C. cellulovorans) CBM28 (C. cellulovorans) CBM44 (C. thermocellum) CBM27 (C. cellulovorans)
tag
substrates
primary goal of work
GFP
Avicel, PASC, cotton fiber, CNC, cellulose II and III Avicel, Kraft pulp, cellulose II
b mOrange2
cotton fiber Kraft pulp
influence on hydrolysis, tracking changes in cellulose accessibility to enzymes during fiber swelling tracking of changes in cellulose accessibility to enzymes during hydrolysis tracking paracrystalline structures monitoring the xylan
Non, CFP mCherry CFP
Avicel, Kraft pulp
monitoring transition structure and paracrystalline cellulose
Organosolv pretreated hardwood, cellulose II Avicel, PASC, cotton fiber, CNC, cellulose II and III Kraft pulp
monitoring transition structure and paracrystalline cellulose
NA
NA CFP
tracking of changes in cellulose accessibility to enzymes during fiber swelling monitoring the mannan
ref 18−20
8, 19, 22, 23 17, 21
8, 15, 19, 21 19 22
16, 20 22
NA: not applicable, CFP: cyan fluorescent protein, GFP: green fluorescent protein. bSome CBMs have been implemented with His-tags and have been analyzed using immunology-based methods.19 a
at the microfibril level, as well as on CAE and hydrolyzability, is poorly understood. The aim of this study was to analyze the effect of mechanical refining on a simplified but relevant substrate (kraft pulp) and assess if observed changes in fibrillation relate to CAE and substrate hydrolyzability. For this purpose, a northern bleached softwood Kraft (NBSK) pulp was refined to various extents (0 to 1500 kWh ton−1), resulting in microfibrillated cellulose (MFC) substrates with negligible differences in their chemical composition. This minimizes the potential influence of lignin and hemicellulose on both hydrolysis and the methods used to predict accessibility in general and CAE in particular. The MFC substrates were never dried, thus keeping the fibers swollen and preventing the pores from collapsing.26 The latter effect has been shown to result in hornification that can negatively affect accessibility and hydrolyzability.24 Two CBMs with different binding preferences toward specific structures within the cellulosic substrate were used herein, CBM2aH6 (type A CBM, crystalline cellulose) and CBM17 (type B, paracrystalline). These CBMs were chosen because their structures and binding mechanisms are well documented.27−29
Thus, CBMs have been used to carry out quantitative and qualitative characterization of cellulosic substrates to try to better understand enzymatic hydrolysis.14,15 Other researchers used CBM adsorption on Avicel to indicate that most of the paracrystalline cellulose is located inside the Avicel particles and, thus, cannot be readily accessed by the cellulases during the initial stages of hydrolysis.8 CBMs have also been used to track changes in cellulose surface morphology during the swelling of cellulose when investigating the potential contribution of Swollenin during enzymatic hydrolysis.16 In work related to the current study, CBM adsorption was shown to correlate well with the observed hydrolysis rates of a range of hydrothermally and thermochemically pretreated substrates.17 As summarized above, much of the previous work that has looked at enzyme accessibility to cellulose has used model substrates such as cotton fiber and Avicel.8,24 Although the lignin−carbohydrate complex imposes a multitude of challenges on the hydrolyzability of realistic substrates, the fiber, fibril, and microfibril structure limits cellulose accessibility. To highlight this, Figure 1 shows one of the unresolved issues,
■
MATERIALS AND METHODS
Substrates Used and Their Preparation. Avicel was purchased from Sigma-Aldrich (St. Louis, MO, U.S.A.). Phosphoric acid swollen cellulose (PASC) was produced from Avicel, as described previously.30 In brief, 4 g of Avicel was suspended in 100 mL of phosphoric acid. After stirring for 1 h at 4 °C, the suspension was diluted with 1900 mL of cold water. After another incubation step (1 h at 4 °C, constant stirring), the extracted swollen cellulose was collected by filtration. The swollen cellulose was washed four times with ultrapure water, two times with 1% NaHCO3 for neutralization, and then three more times with ultrapure water. The MFC substrates were produced by treating NBSK pulp with different refining energies, using an Aikawa single disc 14 in. refiner (Advanced Fiber Technologies Inc., Petaluma, CA, U.S.A.). Refining was performed with a starting consistency of 3.5% (w/v). During refining at net power levels of 20 kW for about 6 h total, the actual consistency of each sample dropped to about 3% (w/v) because of the added cooling water. Each pass through the refiner added a refining energy of about 25 kWh ton−1. The MFC0 sample was collected after 10 passes, corresponding to a refining energy of 250 kWh ton−1. The MFC2, MFC4, and MFC6 substrates were additionally treated 10, 30, and 50 times, resulting in refining energies of 500 kWh ton−1, 1000 kWh ton−1, and 1500 kWh ton−1, respectively.
Figure 1. Diagrammatic representation of relative size of a typical pulp fiber, macrofibril, microfibril, and elemental fibril (glucan chain) in comparison to Cellobiohydrolase 1 (CBHI/Cel7A).
which is how a typical cellulase, 5 nm in size, in this case a diagrammatic representation of Cellobiohydrolase I (Cel7A), is able to access and hydrolyze the closely packed elemental fibrils within the microfibril/macrofibril/fiber structure.7 Mechanical refining has been proposed as a pretreatment method due to its capacity to open up the strongly integrated biomass structure by external fibrillation and internal delamination.25 However, its impact on the cellulose structure 1114
DOI: 10.1021/acssuschemeng.8b04780 ACS Sustainable Chem. Eng. 2019, 7, 1113−1119
Research Article
ACS Sustainable Chemistry & Engineering Analysis and Characterization of the MFCs. Compositional Analysis. The chemical compositions of the MFCs were determined following the two-step sulfuric acid hydrolysis method.31 Glucose, xylose, galactose, arabinose, and mannose were quantified by a high performance liquid chromatography (HPLC; ICS-3000, Dionex, Sunnyvale, CA, USA) equipped with a CarboPac PA1 anion exchange column (Dionex). Nano pure water was used as isocratic eluent at a flow rate of 1 mL min−1. All analyses were performed in triplicates. SEM Imaging of the MFCs. The oven-dried MFC substrates were mounted on aluminum stubs using double-sided tape. After sputtercoating with 10 nm Au/Pd (Gold/Palladium, 80:20 mix), the MFCs were imaged using scanning electron microscopy (SEM, Hitachi S2600 VP-SEM, Tokyo, Japan). Measurement of the Aspect Ratio. The aspect ratio represents the ratio of fiber width versus fiber length. The aspect ratio was determined using an Optest Hi-Resolution fiber quality analyzer (LDA02-series, OpTest Equipment Inc., Hawkesbury, Ontario, Canada). About 5000 fibers were collected to calculate the average length over the range of 0.07 and 10.0 mm. For fibers between 0.5 and 10 mm length, the average width was calculated over the range of 7 to 60 μm. Measurement of Water Retention Value (WRV). The Water Retention Value (WRV) represents the amount of water retained by fibers after centrifugation relative to the dry weight of the substrate. It was determined and calculated using a slightly modified version of TAPPI UM 256 (Water retention value (WRV), Useful Method UM 256 (2015)). In brief, 0.1 g of MFC was incubated in water for 2 h, then centrifuged at 900g for 30 min, and finally oven-dried at 105 °C for 3 days to reach constant weight. Analysis of the Hydrolyzability. The hydrolyzability of MFC0, MFC2, MFC4, and MFC6 was assessed using the commercial enzyme mixture Cellic Ctec3 (Novozymes, Bagsværd, Denmark). The substrate and enzyme loading used was 1% (w/v) dry mass and 5 Filter paper units per g−1 of substrate dry mass, respectively. Reactions were performed in sodium acetate buffer (20 mM, pH 4.8) in 250 mL screw cap Erlenmeyer flasks. The total reaction weight was 50 g. The substrate suspensions were autoclaved, and the enzyme solution was added aseptically. Incubation was at 50 °C, 200 rpm, for 24 h in an orbital incubator shaker (IST-4075, Thermo Fisher Scientific, North Hampton, NH, U.S.A.). During hydrolysis samples were taken regularly. Immediate sample workup included boiling (100 °C, 10 min) and centrifugation (17 000g, 5 min, accuSpin Micro 17, Thermo Fisher). The supernatant was stored at −20 °C prior to quantification by HPLC, as described above. Hydrolysis efficiency was either evaluated by the amount of glucose released (in g L−1) or as glucan conversion calculated by the following equation:
Conversion [%] =
Glucose released [g L−1] 1.1 × cellulose loaded [g L−1]
× 100
shaker. Cells were then transferred to 50 mL of TB-Kan-5 medium in a 250 mL baffled shaken flask. The starting OD600 was 0.05. Incubation was at 37 °C and 150 rpm until an OD600 of 5 was reached. Bioreactor CultivationBatch Phase. For the bioreactor cultivations an Applikon bioreactor system with an ADI-1025 BioConsole and ADI-1010 controller (Applikon Biotechnology, Delft, The Netherlands) with a 2 L working volume was used. The cultivation was run in two phases, a batch followed by a fed-batch phase. For the batch phase, 1 L of TB-Kan medium with 20 g L−1 glycerol (TB-Kan-20) was prepared and inoculated with cells from the starter culture to an OD600 of 0.05. Initially the cultivation conditions were 37 °C, pH 7, and 30% dissolved oxygen (pO2), regulated by a cascaded controller manipulating agitation and aeration with pressurized air. When an OD600 of 0.7 was reached, protein expression was induced with 0.3 mM isopropyl-β-D-thiogalactopyranosid (Corning, New York, NY, U.S.A.). At the same time the temperature was reduced to 30 °C and the pO2 increased to 60%. During the batch phase samples were taken regularly to measure the growth rate as determined by the increase in OD600. The batch phase was continued until glycerol was depleted. This was calculated assuming a specific glycerol update rate (qGlycerol) of 2.67 g gcell dry weight−1 h−1 and a cell dry weight to OD600 ratio of 0.52.32 Bioreactor CultivationFed-Batch Phase. Using a fed-batch strategy, a carbon-source limiting feeding scheme was applied to control the growth rate (μ) at 0.12 h−1. The μ is described to be below μcrit (∼0.2 h−1) at which the onset of acetic acid formation was based on overflow metabolism.32 The feed rate was modeled using the following equation.32 ij μ yz 1 + mzzzz × VtF × X tF × × e μSet × (t − tF) F(t ) = jjjj Y S X / S F k {
(2)
where F(t) is the volumetric flow [g L−1], μ is the uncontrolled growth rate measured in the batch phase [h−1], YX/S is the biomass yield [g g−1], m is the specific maintenance coefficient [g g−1 h−1], VtF is the volume when the feeding starts, XtF is the biomass when the feeding starts [g L−1], SF is the concentration of glycerol in the substrate feed [g L−1], μSet is the controlled feed rate during the fedbatch phase, and tF is the time point when the feeding starts [h]. The μ was calculated during the batch phase as an increase in OD600. The m and YX/S were assumed to be 0.03 g g−1 h−1 and 0.45 g g−1, respectively. The XtF was determined assuming a cell dry weight to OD600 ratio of 0.52 g L−1. The SF was 100 g L−1 glycerol, and μSet was 0.12 h−1. After modeling, the calculated feeding trajectory was used to program the peristaltic feed pump. A 10-fold concentrated TB-Kan medium containing 100 g L−1 glycerol was used as feed. The fedbatch phase was run at 30 °C, pH 7, and 60% pO2. To keep the latter stable throughout the fed-batch phase, the air inflow was fortified with pure O2. Cell Disruption and CBM Purification. Cell disruption on the small scale (
