In Situ FT-IR Microscopic Study on Enzymatic Treatment of Poplar

Jul 18, 2008 - The feasibility of Fourier transform infrared (FT-IR) microscopy to monitor in situ the enzymatic degradation of wood was investigated...
0 downloads 0 Views 4MB Size
2194

Biomacromolecules 2008, 9, 2194–2201

In Situ FT-IR Microscopic Study on Enzymatic Treatment of Poplar Wood Cross-Sections Notburga Gierlinger,*,† Luna Goswami,† Martin Schmidt,† Ingo Burgert,† Catherine Coutand,‡ Tilmann Rogge,§ and Manfred Schwanninger| Department of Biomaterials, Max-Planck Institute of Colloids and Interfaces, Potsdam, Germany, Institut National de la Recherche Agronomique (INRA), umr Physiologie Inte´grative de l’Arbre Fruitier et Forestier (PIAF), 234 av. du Bre´zet, 63100 Clermont-Ferrand, France, Forschungszentrum Karlsruhe GmbH, Institut fu¨r Mikrostrukturtechnik, Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany, and Department of Chemistry, Boku - University of Natural Resources and Applied Life Sciences, Vienna, Muthgasse 18, A-1190 Vienna, Austria Received March 23, 2008; Revised Manuscript Received June 9, 2008

The feasibility of Fourier transform infrared (FT-IR) microscopy to monitor in situ the enzymatic degradation of wood was investigated. Cross-sections of poplar wood were treated with cellulase Onozuka RS within a custombuilt fluidic cell. Light-optical micrographs and FT-IR spectra were acquired in situ from normal and tension wood fibers. Light-optical micrographs showed almost complete removal of the gelatinous (G) layer in tension wood. No structural and spectral changes were observed in the lignified cell walls. The accessibility of cellulose within the lignified cell wall was found to be the main limiting factor, whereas the depletion of the enzyme due to lignin adsorption could be ruled out. The fast, selective hydrolysis of the crystalline cellulose in the G-layer, even at room temperature, might be explained by the gel-like structure and the highly porous surface. Young plantation grown hardwood trees with a high proportion of G-fibers thus represent an interesting resource for bioconversion to fermentable sugars in the process to bioethanol.

Introduction Cellulose is the major polymeric component of plant matter and is the most abundant polysaccharide on Earth. Cellulosic components of biomass are likely to become key resources in the transition of many major national economies to greater reliance on renewable resources within the next decades.1 Although native cellulose, termed cellulose I, has been the most studied subject in polymer science, the biosynthesis and ultrastructure of cellulose are not yet completely resolved.2–5 The deceptive simplicity of the repeating β-1-4 linking cellobiose unit is not indicative of the complex arrangement of amorphous and crystalline regions at the fibril and fiber level. Furthermore, within the plant cell wall, cellulose is embedded in a matrix of other polysaccharides and lignin and, therefore, difficult to access. Enzymes that are capable of degrading or modifying cellulose are known since the beginning of the last century,6 and research has been ongoing over decades.7–9 Besides understanding plant cell wall structure and enzymatic degradation mechanism, the industrial potential in using enzymatic modification has been a driving force, recently, especially the potential for sustainable production of biofuels.10–14 Because of the structural complexity and rigidity of the cellulosic substrates, nature has evolved a diversity of degradative enzymes: the cellulases.15 These cellulases are typically modular, often consisting of a carbohydrate binding module (CBM) that gives the cellulase strong affinity for cellulose, a linker, and a catalytic core module.16 The CBM brings the * To whom correspondence should be addressed. E-mail: gierlinger@ mpikg.mpg.de. † Max-Planck Institute of Colloids and Interfaces. ‡ Institut National de la Recherche Agronomique. § Forschungszentrum Karlsruhe GmbH. | BOKU.

catalytic core module into close and prolonged contact with the carbohydrate substrate.17 Efficient saccharification of cellulose requires a mixture of several functional types of cellulases: (i) endoglucanase (β-1-4-endoglucanase) that randomly cleaves internal glycosidic bonds within an unbroken glucan chain, (ii) cellobiohydrolase (β-1-4-exoglucanase), which hydrolyses cellobiose dimers from the reducing and nonreducing end, and finally (iii) cellobiase (β-glucosidase), which splits cellobiose into glucose monomers.18 The rate and efficiency of enzymatic cellulose hydrolysis is affected by many factors, such as the enzyme system (synergism, adsorption), the substrate characteristics (degree of polymerization, crystallinity, pore size), and association with hemicelluloses and lignin.18–22 Considering the native plant cell wall, differences in substrate characteristics and polymer composition and association are expected in different tissue types within one plant and also within the different layers of the plant cell wall, like middle lamella, primary wall, and secondary cell wall. Within reaction wood of angiosperms, formed often in young trees of plantations, an additional gelatinous layer (G-layer) consisting of highly crystalline cellulose is laid down on a small lignified secondary cell wall. This so-called tension wood contains 40-50% more cellulose than normal wood and is, thus, of interest in terms of effective cellulose hydrolysis of biomass.23,24 To monitor the cellulose degradation of different tissue types at a molecular level the feasibility of Fourier transform infrared (FT-IR) microscopy was investigated. Infrared (IR) spectroscopy in general has so far proven to be a powerful method for the investigation of cellulose,25–31 plant tissues,32,33 and wood.34–37 Custom-built hydration and fluidic cells for FT-IR microscopy have been developed to allow analysis of biological samples in the wet state.38,39 For in situ enzymatic treatment experiments

10.1021/bm800300b CCC: $40.75  2008 American Chemical Society Published on Web 07/18/2008

Enzymatic Treatment of Poplar Wood Cross-Sections

Figure 1. Design of the custom-built fluidic cell for FT-IR microscopy: Within a steel mount (I) a polyimide structured CaF2 window (IIa) is compressed to a CaF2 window with fluidic connections (IIb), together with a heating element (III) and rubber rings on the upper and lower side (IVa,b).

a heating device to enhance enzyme activity40 and a constant flow to exclude feedback inhibition by endproducts20 had to also be implemented. To ensure uniform accessibility of the enzyme to different cell wall layers of the wood fibers, thin cross-sections were used. The overall aim was to achieve enzymatic hydrolysis of native plant cell walls within the fluidic cell and to follow in situ during hydrolysis the microscopic and spectral changes at positions with different cell wall composition. Besides, the influence of environmental conditions (e.g., temperature) on the enzymatic hydrolysis was aimed to be monitored in context with anatomical structure and chemical composition. Cell wall areas, hydrolyzed fast and, thus, of interest for the production of bioethanol, will be identified together with regions needing different and prolonged treatment. In future studies, the developed approach should furthermore allow to optimize treatment protocols for the different cell wall areas.

Biomacromolecules, Vol. 9, No. 8, 2008

2195

that is placed in a meandering groove of the anodized aluminum ring, sealed with silicone, and covered by a thin polyester sheet. The temperature is controlled using a thermocouple and a proportional integral derivative (PID) controller. During the enzymatic treatment spectra and light-optical micrographs were acquired using a FT-IR microscope (Hyperion 2000, 15× objective) equipped with a liquid nitrogen cooled MCT detector and connected to a Vertex 70 FT-IR spectrometer (Bruker GmbH, Germany). A total of 128 scans were coadded per sample spectrum (wavenumber range 4000-900 cm-1) and apodized, applying the Blackman-Harris three-term function and a zero filling factor of 2. Time-dependent measurements, at intervals down to 5 min in the beginning of each experiment, were always done on the same positions (aperture 70 × 70 µm) on tension wood fibers, normal wood fibers, and of the pure enzyme solution. For spectral analysis (OPUS Vers. 6.5), an atmospheric compensation and a baseline correction (rubberband-method, 32 points) were done before peak picking and intensity calculation. When bands appear as shoulders (e.g., bands close to 1645 cm-1), the correct wavenumber position of the maxima was assessed by the following procedures: calculation of the second derivative, difference spectra, as well as, curve fitting. Spectra related to the degraded part of the sample were derived by subtracting the spectrum acquired at the end of the treatment from the spectra acquired during treatment (difference spectra). To investigate changes of the acetyl groups during treatment second derivatives (according to Savitzky and Golay,41 13-point smoothing) were calculated, in which the shoulder visible in the original spectra becomes separated from the dominant water band. The progression of enzymatic degradation was visualized by plotting the absolute and the relative (percental) change in intensity of the bands in the fingerprint region with treatment time. The region below 1000 cm-1 was not available due to the absorbance of the CaF2 windows. The higher wavenumber range of the OH- and CH- stretching vibrations (3800-2800 cm-1) of the acquired spectra was not analyzed because (i) transmission in the OH region was low due to almost total absorbance of water and (ii) the very broad water band influences the CH region making it not suitable for the analysis of the wet wood samples.

Experimental Section Samples were taken from a three-year-old hybrid poplar (Populus nigra × P. deltoids i14551, artificially tilted). Without any further sample preparation, 8 µm thick cross-sections were cut on a rotary microtome (LEICA RM2255, Germany) and placed wet in a freezer. For the in situ experiment, the cross-sections were placed on one of the two calcium fluoride windows (Crystech Inc., UV grade, L ) 1 in., thickness ) 2 mm), forming the central part of the custom-built microfluidic cell for FT-IR microscopy (Figure 1). One of the CaF2 windows carries polyimide structures produced by photolithography (Figure 1, IIa). These polyimide structures serve as a spacer, define the sample volume and act as a sealant when compressing both windows. In this way, a low path length (10 µm) is achieved, which minimizes strong liquid or solvent absorption in the mid infrared wavelength range and enables transmission measurements in liquid environment. The second CaF2 window contains lateral fluidic connections using PEEK capillaries (outer L ) 360 µm, inner L ) 150 µm; Figure 1, IIb). These are connected to a high-precision, low-flowrate syringe infusion pump (Harvard Apparatus, 11 Plus) to achieve a constant flow rate. First, background measurements were done in the dry state, followed by a constant water flow until all air bubbles were removed. Then the enzyme solution was delivered with a constant flow rate of 8 µL/h (volume of the cell ∼ 10 µL). The multicomponent enzyme “Cellulase Onozuka RS” from Trichoderma Viride (E.C.3.2.1.4)” (Yakult Pharmaceutical Industry, Co. Ltd., Japan) was used with a concentration of 0.05 g/mL, dissolved in buffer (50 mM Na2HPO4 adjusted with citric acid to pH 5). Enzymatic treatment was done at room temperature (22 °C) and at 50 °C using a heating element (a resistively heated disk-shaped metal ring, Figure 1, III) in contact with the upper CaF2 window. This element is made up of a conducting wire

Results Light-optical micrographs show selected areas of normal and tension wood (Figure 2A-C). The latter one has a pronounced gelatinous (G) layer, which almost completely fills the lumen of the cells (Figure 2B,C). The images taken during the cellulase treatment show no change in cell wall structure of normal wood (Figure 2A) but almost complete removal of the G-layer in the tension wood after 50 h of treatment at room temperature (Figure 2B). Within one cell the G-layer was removed already after 5 h, followed by other single cells during the next 10 h. After 30 h, a clear change is observed and after 50 h beside the secondary cell wall only some remnants of the G-layer are visible (Figure 2B). Treatment at a temperature of 50 °C showed noticeable changes after 5 h and already after 10 h within almost all cells the G-layer is strongly affected (Figure 2C). The fingerprint region of the spectra from the normal and tension wood regions show clear differences (Figure 3). By comparing the spectral characteristics of both cell variants differences in chemistry become clear and spectral contributions of the different wood polymers as well as band assignments can be ascertained (Table 1). In the tension wood spectrum, the bands at 1732 and 1247 cm-1, attributed to acetyl groups in hemicelluloses (xyloglucan), as well as the aromatic stretching vibrations of lignin at 1506 and 1595 cm-1 are smaller compared to the normal wood spectrum. The band at 1462 cm-1 is slightly higher and the one at 1373 cm-1 is lower in the normal wood spectra. Contributions of lignin and xyloglucan are supposed

2196

Biomacromolecules, Vol. 9, No. 8, 2008

Gierlinger et al.

Figure 2. Structural changes of normal wood (A, first row) and tension wood of poplar during a 50 h long enzymatic treatment within the fluidic cell at room temperature (B) and at 50 °C (C; picture size ) measurement area ) 70 × 70 µm).

Figure 3. Comparison of the absolute intensities and wavenumber positions of the bands in the fingerprint region of baseline-corrected FT-IR spectra acquired from normal wood (NW) and tension wood (TW).

to overlay the OH- and CH-deformation vibrations expected from cellulose in this region (Table 1). Bands characteristic and clearly resolved in the tension wood, but not in the normal wood, at 1335 and 1205 cm-1, are attributed to OH in-plane bending vibrations, the latter one is also assigned to the symmetric C-O-C stretching vibration. Within the CO stretching region, the most remarkable differences of band positions and heights are the 1111 cm-1 band is higher in the tension wood and there is a maximum at 1060 cm-1, whereas in the normal wood two maxima are found at 1051 and at 1034 cm-1. At 1033 cm-1, aromatic C-H in-plane deformation assigned to lignin overlays with the C-O stretching band of cellulose at 1035 cm-1. Many band positions are shifted to higher wavenumbers in the tension wood (Figure 3, Table 1). The most pronounced shift (9 cm-1) is observed at 1060 cm-1, assigned to C-O stretching of C3-O3H. Baseline-corrected spectra of the normal wood show almost no changes during the 50 h enzymatic treatment (Figure 4A), proving that the enzyme is not degrading the lignified cell wall to an extent that was observable by FT-IR. In contrast, in the tension wood spectra a decrease of almost all resolved bands is seen, except the band at 1645 cm-1, attributed to the deformation vibration of adsorbed and free water, is increasing (Figure 4B). The most pronounced decrease is observed in the C-O

stretching region accompanied by a clear shift to lower wavenumbers after 30 h. Subtraction of the spectrum taken after 50 h of treatment from spectra taken after different treatment times gives spectra representative for the degraded part (Figure 4C); mainly cellulose bands (Table 1) are observed, which diminish without a change in band position. In the second derivatives of the spectra, the shoulder at 1738 cm-1, attributed to acetyl groups of xyloglucan, is resolved as a single band (Figure 4D). In contrast to the cellulose bands, it declines mainly during the first hours. For visualizing the progression of the enzymatic degradation the change in band intensity was plotted with time for all resolved bands in the fingerprint region as absolute (Figure 5A) and percental (taking into account the different band heights, Figure 5B) change. The intensity of the water band at 1645 cm-1 shows a pronounced increase (16%) during the first 7 h and thereafter a continuous increase of 9% (Figure 5B). The small lignin band found at 1506 cm-1 shows a slight increase in absolute intensity (Figure 5A) and the same percental increase as the water band (Figure 5B). The bands least affected by the treatment are the ones at 1460 and 1428 cm-1, followed by the band at 1247 cm-1 (Figure 5A,B). The first two as shown by comparison to the normal wood spectrum (Figure 3) are strongly influenced by the lignin and the latter by xyloglucan (Table 1). The 1372, 1335, and 1319 cm-1 bands show minor decrease in absolute intensity (Figure 5A), but taking the band height into account, especially the 1335 and 1319 cm-1 bands, reveals about a 40% decrease (Figure 5B). The same is found for the C-O-C stretching vibration at 1162 cm-1 (Figure 5 A-B). The two bands showing the largest absolute decrease are the characteristic tension wood bands at 1111 and 1060 cm-1 (Figure 5A), which are reduced to about 30-40% (Figure 5B). The profiles of the degradation curves are similar for all bands and show an exponential decay with the highest rate during the first 7 h. Remarkably, the enzymatic degradation starts immediately (at least with respect to the time resolution of the measurement), which is not visible in the light-microscopic images. Therefore, the first hours and the penetration of the enzyme were studied in more detail. To verify the assumption that influences on the spectra due to concentration changes of the enzyme solution and the enzyme solution itself can be excluded or at least be neglected, spectra were also collected from a blank region aside the cross-sections of the wood samples. By this it was furthermore possible to

Enzymatic Treatment of Poplar Wood Cross-Sections

Biomacromolecules, Vol. 9, No. 8, 2008

2197

Table 1. Bands Found in Normal Wood (NW), Tension Wood (TW), and G-Layer Spectra and their Assignments According to the Literaturea NW 1732

TW

G-layer

1732 sh

1738

cellulose

43

1645 1595 sh 1505 1462

1645

1635 (adsorbed water)

1506 1460

1453

45

1455 (OH i.p. bend.)

1425

1428

1430

1430 (CH2 bend.)45

1373

1372

1371

1374 (CH bend.)45

1321

1335 1319

1246

1247

1336 1318 1274 1247

1206

1205

1336 (OH i.p. bend.)45 1317 (CH2 wag.)45 1277 - 1282 CH def.48 1235 - 1225 OH i.p. def., also COOH48 1205 (OH i.p. bend.);45 1205 (sym. C-O-C str.)49 1160 C-O-C asym. str. 49 1115 (C-O str., mainly of C2-O2H secondary alcohols)49 1060 (C-O str. mainly of C3-O3H secondary alcohols)49 1060-1015 (C-O valence vibration)48 1035 (C-O str., mainly of C6H2-O6H primary alcohols)49 1060-1015 (C-O valence vibration) 48

1159 1108

1162 1111 1060

1163 1112 1061

1051 1034

1036

1037

lignin (beech)42

xyloglucan 1745-1725 (CdO str. of acetyl or carboxylic acid groups)43,44 1635 (adsorbed water) 1460 (CH2 sym. bend. on pyran ring)43 1425 (carboxylic acid and sym. COO- str.)43 + CH bend.46 1374 (CH sym. def. of CH3 from acetyl groups)44 + 1370 CH2 bend.46,47 1336 -1202(CH and OH def.)47 1240 (asym. str. of C-O-C of acetyl groups)43

1735 CdO str. in unconjugated ketones 1593 1505 1462 -CH2 1422 C-H 1367

(arom. skeletal plus CdO str) (arom. skeletal) (CH-def.; asym. in -CH3 and (arom. skeletal combined with i.p. def) (aliphatic C-H in CH3)

1329 (phenolic OH def.) 1266 (G-ring plus CdO str.) 1227 (C-C plus C-O plus CdO str.)

50

1153 (C-O-C asym.) 1118 (C-O str.)50

1126 (arom. C-H i.p. def. plus CdO str., typical for S-units) 50,51

1078 (C-O and C-C ring)

1041 (C-O and C-C ring)50,51 1034 (sym. str. of C-O-C of acetyl groups, very weak band)44

1033 (arom. C-H i.p. def. plus C-O def. plus CdO str.)

a The spectra and band positions of the hydrolyzed G-layer were derived by subtracting the spectrum at the end of the treatment from the spectrum at the beginning (sh ) shoulder, sym ) symmetric, asym- antisymmetric, str ) stretching, i.p. ) in plane, bend. ) bending, def. ) deformation, arom. ) aromatic, wag. ) wagging, asym ) antisymmetric).

Figure 4. Spectral changes in the fingerprint region of baseline-corrected (bc) FT-IR spectra of normal wood (A) and tension wood (B) during a 50 h cellulase treatment at room temperature. Spectra of the degraded part (G-layer) were calculated by subtracting the spectrum at 50 h treatment from the spectra at different treatment times (C). To visualize changes of the acetyl groups, the second derivatives of the spectra are shown (D).

2198

Biomacromolecules, Vol. 9, No. 8, 2008

Gierlinger et al.

Figure 5. Change in absolute (A) and relative (B) absorbance intensity of all bands during a 50 h cellulase treatment of poplar tension wood at ambient temperature.

follow the penetration of the enzyme solution and thus to determine the starting point of treatment exactly as the enzyme solution shows small additional bands at 1158 and 1077 cm-1 due to the phosphate buffer. The change from water to enzyme solution was observed within about 5 min and there was no further change during the treatment time (Figure 6). By subtracting the water spectrum from the enzyme solution spectrum, the amide I and II bands from the enzymes could be seen at 1645 and 1567 cm-1 (Figure 6, inset). The overall intensity of the pure enzyme solution did not exceed an absorbance of 0.08 and, thus, concentration changes during the treatment can be considered negligible. No degradation products could be identified in the wood and enzyme solution spectra, which is due to the continuous flow of the enzyme solution. After penetration of the enzyme solution the degradation of the G-layer started almost immediately, even at room temperature (Figure 7A). Already within the first 5 h about one-third of the whole degradation took place. Utilizing the heating element for the fluidic cell allowed treatments also at higher temperature. Compared to room temperature, the degradation rate at 50 °C increased dramatically (Figure 7B, note different scale). An immediate, strong degradation especially within the first 15 min was followed by a decrease of up to 38% of the band heights within 5 h. Comparing the decrease in absorbance

Figure 6. Fingerprint region of water and the enzyme solution acquired before and during the treatment, respectively. The contribution of the enzyme and the buffer was derived by subtracting the water spectrum from the enzyme solution spectrum (inset).

of the band at 1111 cm-1 at room temperature and at 50 °C shows clearly the difference in the degradation rate and the respective time needed for “complete” removal of the G-layer (Figure 7C).

Enzymatic Treatment of Poplar Wood Cross-Sections

Biomacromolecules, Vol. 9, No. 8, 2008

2199

Figure 7. Change in relative absorbance intensity during the first 5 h of cellulase treatment of poplar tension wood at ambient temperature (A) and at 50 °C (B), and a comparison of the decrease of the band at 1111 cm-1 during the enzymatic treatment at ambient temperature and at 50 °C (C).

Discussion To get a better understanding of the enzymatic degradation of native plant cell walls, microscopic approaches are needed due to the chemical diversity at the tissue and cell level. The light-microscopic images acquired during the enzymatic treatment showed that the G-layer was removed almost completely

in the tension wood, while the secondary cell wall seemed rather unaffected in tension wood as well as in normal wood (Figure 2). Some remnants of the G-layer were still found after long treatment. The spatial resolution of FT-IR microscopy (∼10 µm) does not allow investigating these remnants or the thin secondary cell wall and the middle lamella separately and a chemical

2200

Biomacromolecules, Vol. 9, No. 8, 2008

imaging approach is therefore not straightforward. By contrast, the high spatial resolution (