Accumulation of recombinant cellobiohydrolase and endoglucanase in the leaves of mature transgenic sugar cane

A major strategic goal in making ethanol from lignocellulosic biomass a cost-competitive liquid transport fuel is to reduce the cost of production of cellulolytic enzymes that hydrolyse lignocellulosic substrates to fermentable sugars. Current production systems for these enzymes, namely microbes, are not economic. One way to substantially reduce production costs is to express cellulolytic enzymes in plants at levels that are high enough to hydrolyse lignocellulosic biomass. Sugar cane fibre (bagasse) is the most promising lignocellulosic feedstock for conversion to ethanol in the tropics and subtropics. Cellulolytic enzyme production in sugar cane will have a substantial impact on the economics of lignocellulosic ethanol production from bagasse. We therefore generated transgenic sugar cane accumulating three cellulolytic enzymes, fungal cellobiohydrolase I (CBH I), CBH II and bacterial endoglucanase (EG), in leaves using the maize PepC promoter as an alternative to maize Ubi1 for controlling transgene expression. Different subcellular targeting signals were shown to have a substantial impact on the accumulation of these enzymes; the CBHs and EG accumulated to higher levels when fused to a vacuolar-sorting determinant than to an endoplasmic reticulum-retention signal, while EG was produced in the largest amounts when fused to a chloroplast-targeting signal. These results are the first demonstration of the expression and accumulation of recombinant CBH I, CBH II and EG in sugar cane and represent a significant first step towards the optimization of cellulolytic enzyme expression in sugar cane for the economic production of lignocellulosic ethanol.     

里氏木霉纤维素酶的分离纯化及酶学性质

通过(NH4)2SO4分级沉淀、HiPrep 26/10 Desalting凝胶色谱脱盐、Source 15 Q阴离子交换色谱技术,里氏木霉(Rut C-30)纤维素酶主要组分得以初步分开,再经过Source 15 S阳离子交换色谱、HiPrep Sephacryl S-100 HR凝胶过滤色谱、Superdex 75 PrepGrade凝胶过滤色谱进一步分离纯化,得到2个纯化的内切葡聚糖酶组分EGⅡ、EGⅠ和一个外切葡聚糖酶组分CBHⅠ;经过SDS-PAGE电泳鉴定为电泳纯,测得相对分子质量分别为5.22×104,5.62×104和6.90×104。EGⅡ的最适反应pH是5.6,最适反应温度为65℃;EGⅠ的最适反应pH是4.4,最适反应温度为55℃;以羧甲基纤维素(CMC)为底物时,EGⅠ、EGⅡ的米氏常数(Km)分别为2.20 mg/mL、3.38 mg/mL。CBHⅠ的最适反应pH是5.8,最适反应温度为60℃,以对硝基苯基-β-D-纤维二糖苷(PNPC)为底物时,米氏常数(Km)为0.12 mg/mL。     

Quantification and identification of the main components of the Trichoderma cellulase complex with monoclonal antibodies using an enzyme-linked immunosorbent assay (ELISA)

Summary An enzyme-linked immunosorbent assay (ELISA) using monoclonal antibodies has been developed to measure the concentration of three main cellulase components from Trichoderma reesei, cellobiohydrolase I (CBH I), cellobiohydrolase II (CBH II) and I (EG I), in both commercial enzyme preparations as well as in samples from laboratory fermentations. The sensitivity of the assay is 1–10 ng protein, depending on the type of cellulase. The coefficient of variability is between 10% and 20%. By a combination of two different domain-specific monoclonals against CBH I or II it is also possible to quantify the concentration of intact and truncated forms of these two enzymes, respectively. The use of the ELISA to quantify the formation of the three cellulase components under different cultivation conditions is described. Offprint requests to: C. P. Kubicek     

Binding characteristics of Trichoderma reesei cellulases on untreated, ammonia fiber expansion (AFEX), and dilute-acid pretreated lignocellulosic biomass

Studying the binding properties of cellulases to lignocellulosic substrates is critical to achieving a fundamental understanding of plant cell wall saccharification. Lignin auto-fluorescence and degradation products formed during pretreatment impede accurate quantification of individual glycosyl hydrolases (GH) binding to pretreated cell walls. A high-throughput fast protein liquid chromatography (HT-FPLC)-based method has been developed to quantify cellobiohydrolase I (CBH I or Cel7A), cellobiohydrolase II (CBH II or Cel6A), and endoglucanase I (EG I or Cel7B) present in hydrolyzates of untreated, ammonia fiber expansion (AFEX), and dilute-acid pretreated corn stover (CS). This method can accurately quantify individual enzymes present in complex binary and ternary protein mixtures without interference from plant cell wall-derived components. The binding isotherms for CBH I, CBH II, and EG I were obtained after incubation for 2 h at 4 °C. Both AFEX and dilute acid pretreatment resulted in increased cellulase binding compared with untreated CS. Cooperative binding of CBH I and/or CBH II in the presence of EG I was observed only for AFEX treated CS. Competitive binding between enzymes was found for certain other enzyme-substrate combinations over the protein loading range tested (i.e., 25-450 mg/g glucan). Langmuir single-site adsorption model was fitted to the binding isotherm data to estimate total available binding sites E(bm) (mg/g glucan) and association constant K(a) (L/mg). Our results clearly demonstrate that the characteristics of cellulase binding depend not only on the enzyme GH family but also on the type of pretreatment method employed.     

Visualization of Trichoderma reesei Cellobiohydrolase I and Endoglucanase I on Aspen Cellulose by Using Monoclonal Antibody-Colloidal Gold Conjugates

Monoclonal antibodies (MAbs) specific for cellobiohydrolase I (CBH I) and endoglucanase I (EG I) were conjugated to 10- and 15-nm colloidal gold particles, respectively. The binding of CBH I and EG I was visualized by utilizing the MAb-colloidal gold probes. The visualization procedure involved immobilization of cellulose microfibrils on copper electron microscopy grids, incubation of the cellulose-coated grids with cellulase(s), binding of MAb-colloidal gold conjugates to cellulase(s), and visualization via transmission electron microscopy. CBH I was seen bound to apparent crystalline cellulose as well as apparent amorphous cellulose. EG I was seen bound extensively to apparent amorphous cellulose with minimal binding to crystalline cellulose.     

Functional display of fungal cellulases from <Emphasis Type="Italic">Trichoderma </Emphasis><Emphasis Type="Italic">reesei</Emphasis> on phage M13

To test whether the phage display technology could be applied in cellulase engineering, phagemids harboring the genes encoding the mature forms of cellobiohydrolase I (CBH I) and endoglucanase I (EG I) from filamentous fungus Trichoderma reesei were constructed, respectively. CBH I and EG I fused to the phage coat protein encoded by the g3 gene were expressed and displayed on phage M13. The phage-bound cellulases retained their activities as determined by hydrolysis of the corresponding substrates, Also, their binding abilities to insoluble cellulose substrate were confirmed by an ELISA method. Overall, these results demonstrate that cellulases can be displayed on phage surface while maintaining their biological function, thus providing an alternative for directed evolution and high-throughput screening for improved cellulases.     

Strategy for Identification of Novel Fungal and Bacterial Glycosyl Hydrolase Hybrid Mixtures that can Efficiently Saccharify Pretreated Lignocellulosic Biomass

A rational four-step strategy to identify novel bacterial glycosyl hydrolases (GH), in combination with various fungal enzymes, was applied in order to develop tailored enzyme cocktails to efficiently hydrolyze pretreated lignocellulosic biomass. The fungal cellulases include cellobiohydrolase I (CBH I; GH family 7A), cellobiohydrolase II (CBH II; GH family 6A), endoglucanase I (EG I; GH family 7B), and β-glucosidase (βG; GH family 3). Bacterial endocellulases (LC1 and LC2; GH family 5), β-glucosidase (LβG; GH family 1), endoxylanases (LX1 and LX2; GH family 10), and β-xylosidase (LβX; GH family 52) from multiple sources were cloned, expressed, and purified. Enzymatic hydrolysis for varying enzyme combinations was carried out on ammonia fiber expansion (AFEX)-treated corn stover at three total protein loadings (i.e., 33, 16.5, and 11 mg enzyme/g glucan). The optimal mass ratio of enzymes necessary to maximize both glucan and xylan yields was determined using a suitable design of experiments. The optimal hybrid enzyme mixtures contained fungal cellulases (78% of total protein loading), which included CBH I (loading ranging between 9-51% of total enzyme), CBH II (9-51%), EG I (10-50%), and bacterial hemicellulases (22% of total protein loading) comprising of LX1 (13%) and LβX (9%). The hybrid mixture was effective at 50°C, pH 4.5 to maximize saccharification of AFEX-treated corn stover resulting in 95% glucan and 65% xylan conversion. This strategy of screening novel enzyme mixtures on pretreated lignocellulose would ultimately lead to the development of tailored enzyme cocktails that can hydrolyze plant cell walls efficiently and economically to produce cellulosic ethanol.     

New Effective Method for Analysis of the Component Composition of Enzyme Complexes from <Emphasis Type="Italic">Trichoderma reesei</Emphasis>

A method for analysis of the component composition of multienzyme complexes secreted by the filamentous fungus Trichoderma reesei was developed. The method is based on chromatofocusing followed by further identification of protein fractions according to their substrate specificity and molecular characteristics of the proteins. The method allows identifying practically all known cellulases and hemicellulases of T. reesei: endoglucanase I (EG I), EG II, EG III, cellobiohydrolase I (CBH I), CBH II, xylanase I (XYL I), XYL II, beta-xylosidase, alpha-L-arabinofuranosidase, acetyl xylan esterase, mannanase, alpha-galactosidase, xyloglucanase, polygalacturonase, and exo-beta-1,3-glucosidase. The component composition of several laboratory and commercial T. reesei preparations was studied and the content of the individual enzymes in these preparations was quantified. The influence of fermentation conditions on the component composition of secreted enzyme complexes was revealed. The characteristic features of enzyme preparations obtained in "cellulase" and "xylanase" fermentation conditions are shown.     

Mode of action of cellulases on dyed cotton with a reactive dye

Cotton woven fabrics which were previously dyed with a reactive dye were treated with a commercial cellulase preparation. Dyeing with a reactive dye for cotton apparently inhibited the weight loss activity and saccharification activity of cellulase. In addition, dyed cotton was treated with highly purified cellulases which were exo-type cellulases (Cellobiohydrolase I (CBH I) and Cellobiohydrolase II (CBH II)) and endo-type cellulase (Endoglucanase II (EG II)). Exo-type cellulases were inhibited more than endo-type cellulase by dyeing in the case of saccharification activity. CBH I was severely inhibited by dyeing as compared with CBH II or EG II from the viewpoint of morphological changes in the fiber surface. Dyes on the cellulose substrates severely influenced CBH I in spite of the rare modification, because CBH I hydrolyzed cellulose with true-processive action. The change in the activity of each cellulase component on dyed cotton can affect the synergistic action of cellulases.     

The conserved terminal region of Trichoderma reesei cellulases forms a strong antigenic epitope for polyclonal antibodies.

The specificity of polyclonal antibodies (Pab) raised against Trichoderma reesei cellulases has been studied. cDNAs lacking regions coding for certain functional domains were produced by preparing series of 3'-end deletions from the cDNAs for two cellobiohydrolases, CBH I and CBH II, and an endoglucanase, EG I. The proteins coded by the full length cDNAs and the truncated proteins coded by the deleted cDNAs were expressed in yeast Saccharomyces cerevisiae, under the control of the ADC1 promoter. Each polyclonal antiserum showed cross-reactivity with other cellulases. Pabs for CBH I and CBH II both recognized EG I. Pab for EG I strongly recognized both CBH I and CBH II. By analyzing the truncated proteins, we found that these antibodies were almost entirely directed against the conserved tail of the cellulase enzymes.     

Lactosylamidine-based affinity purification for cellulolytic enzymes EG I and CBH I from Hypocrea jecorina and their properties

Selective adsorption and separation of β-glucosidase, endo-acting endo-β-(1→4)-glucanase I (EG I), and exo-acting cellobiohydrolase I (CBH I) were achieved by affinity chromatography with β-lactosylamidine as ligand. A crude cellulase preparation from Hypocrea jecorina served as the source of enzyme. When crude cellulase was applied to the lactosylamidine-based affinity column, β-glucosidase appeared in the unbound fraction. By contrast, EG I and CBH I were retained on the column and then separated from each other by appropriately adjusting the elution conditions. The relative affinities of the enzymes, based on their column elution conditions, were strongly dependent on the ligand. The highly purified EG I and CBH I, obtained by affinity chromatography, were further purified by Mono P and DEAE chromatography, respectively. EG I and CBH I cleave only at the phenolic bond in p-nitrophenyl glycosides with lactose and N-acetyllactosamine (LacNAc). By contrast, both scissile bonds in p-nitrophenyl glycosides with cellobiose were subject to hydrolysis although with important differences in their kinetic parameters.     

嗜热毛壳菌纤维素酶（CBHⅡ）cDNA的克隆及在毕赤酵母中的表达

嗜热毛壳菌Chaetomium thermophilum CT2是一种土壤腐生菌,可产生具有重要工业生产价值的纤维素酶类。RACE-PCR获得嗜热毛壳菌纤维二糖水解酶Ⅱ（CBHⅡ）的编码基因(cbh2)。DNA序列分析表明cbh2的开放阅读框由1428个碱基组成,编码476个氨基酸。推断的氨基酸序列包含一个典型真菌纤维素酶的糖结合域（CBD）、催化域（CD）以及二者之间富含脯氨酸和羟基氨基酸的连接桥。根据氨基酸序列推算该酶分子量为53kD,属于糖苷水解酶第六家族,具有该家族催化保守区的典型特征。PCR扩增cbh2的成熟蛋白编码基因,利用基因重组的方法构建可在毕赤酵母分泌表达系统中表达纤维二糖水解酶蛋白的重组表达载体,并转化毕赤酵母得到重组子。在毕赤酵母醇氧化酶AOX1基因启动子的作用下,重组蛋白得到高效表达,小规模发酵量达1.2 mg/mL。经硫酸铵沉淀、DEAESepharose Fast flow阴离子层析等步骤纯化了该重组表达蛋白。SDS-PAGE得到重组蛋白分子量为67kD,与从嗜热毛壳菌中纯化的该酶分子量一致。该重组纤维二糖水解酶作用的最适合温度50℃,最适pH4.0,在70℃的半衰期为30min,具有较好的热稳定性。     

Cellobiohydrolase II is the main conidial-bound cellulase in Trichoderma reesei and other Trichoderma strains

Monoclonal antibodies have been used to determine the presence of cellobiohydrolases I and II (CBH I and II), and endoglucanase I (EG I) on the surface of conidia from Trichoderma reesei QM 9414 and RUT C-30, and 8 other Trichoderma species. For this purpose, proteins were released from the conidial surface by treatment with a non-ionic detergent (Triton X-100 and -octylglucoside), followed by SDS-PAGE/Western blotting and immunostaining. Both CBH I and II were clearly present, but — unlike in extracellular culture fluids from Trichoderma — CBH II was the predominant cellulase. In T. reesei EG I could not be detected. The higher producer strain T. reesei RUT C-30 exhibited a higher conidial level of CBH II than T. reesei QM 9414. In order to assess the importance of the conidial CBH II level for cellulase induction by cellulose, multiple copies of the chb2 gene were introduced into the T. reesei genome by cotransformation using PyrG as a marker. Stable multicopy transformants secreted the 2- to 4-fold level of CBH II into the culture medium when grown on lactose as a carbon source, but their CBH I secretion was unaltered. Upon growth on cellulose, both CBH I and CBH II secretion was enhanced. Those strain showing highest cellulase activity on cellulose also appeared to contain the highest level of conidial bound CBH II. CBH II was also the predominant conidial cellulase in various other Trichoderma sp. However, roughly the same amount of conidial bound CBH II was detected in all strains, although their cellulase production differed considerably.     

Intrinsic Fluorescence in Endoglucanase and Cellobiohydrolase from Trichoderma pseudokiningii S-38: Effects of pH, Quenching Agents, and Ligand Binding

To gain further insight into the difference in substrate specificity between endoglucanase and cellobiohydrolase, the intrinsic fluorescence properties of cellobiohydrolase I (CBH I) and endoglucanase I (EG I) from Trichoderma pseudokiningii S-38 were investigated. The results for the spectral characteristics, ligand binding and fluorescence quenching suggest that the fluorescence of two enzymes comes from tryptophan residues, and that tryptophan residue(s) may be involved in the function of the two enzymes. The results also suggest that the binding tryptophan in EG I may be more exposed to solvent than that in CBH I. This interpretation is supported by the observations that the effects of pH upon the fluorescence of EG I are greater than that of CBH I; spectral shifts are different in EG I and CBH I under various conditions, and fluorescence lifetime changes caused by cellobiose binding are larger for EG I than for CBH I.     

Trichoderma reesei cellobiohydrolase II is associated with the outer membrane when overexpressed in Escherichia coli

Cellulose degradation is essential for the future production of many advanced biofuels. Cellulases from the filamentous fungus Trichoderma reesei are among the most efficient enzymes for the hydrolysis of cellulosic materials. One of the cellulases from T. reesei, cellobiohydrolase II (CBH2), was studied because of its industrial relevance and proven enzymatic activity. Using both crude and rigorous membrane fractionation methods we show that full length T. reesei CBH2 is exclusively localized to the outer membrane when expressed recombinantly in Escherichia coli. Even fusing signal sequence-free maltose-binding protein to the N-terminus of CBH2, which has been shown to increase solubility of other proteins, did not prevent the outer membrane localization of CBH2. These results highlight the difficulties in producing fungal cellulases in bacterial hosts and provide a stepping stone for future cellulase engineering efforts.     

Intrinsic Fluorescence in Endoglucanase and Cellobiohydrolase from Trichoderma pseudokiningii S-38: Effects of pH, Quenching Agents, and Ligand Binding

To gain further insight into the difference in substrate specificity between endoglucanase and cellobiohydrolase, the intrinsic fluorescence properties of cellobiohydrolase I (CBH I) and endoglucanase I (EG I) from Trichoderma pseudokiningii S-38 were investigated. The results for the spectral characteristics, ligand binding and fluorescence quenching suggest that the fluorescence of two enzymes comes from tryptophan residues, and that tryptophan residue(s) may be involved in the function of the two enzymes. The results also suggest that the binding tryptophan in EG I may be more exposed to solvent than that in CBH I. This interpretation is supported by the observations that the effects of pH upon the fluorescence of EG I are greater than that of CBH I; spectral shifts are different in EG I and CBH I under various conditions, and fluorescence lifetime changes caused by cellobiose binding are larger for EG I than for CBH I.     

Hydrolysis of microcrystalline cellulose by cellobiohydrolase I and endoglucanase II from Trichoderma reesei: adsorption, sugar production pattern, and synergism of the enzymes

Microcrystalline cellulose (10 g/L Avicel) was hydrolysed by two major cellulases, cellobiohydrolase I (CBH I) and endoglucanase II (EG II), of Trichoderma reesei. Two types of experiments were performed, and in both cases the enzymes were added alone and together, in equimolar mixtures. In time course studies the reaction time was varied between 3 min and 48 h at constant temperature (40 degrees C) and enzyme loading (0.16 micromol/g Avicel). In isotherm studies the enzyme loading was varied in the range of 0.08-2.56 micromol/g at 4 degrees C and 90 min. Adsorption of the enzymes and production of soluble sugars were followed by FPLC and HPLC, respectively. Adsorption started quickly (50% of maximum achieved after 3 min) but was not completed before 60-90 min. For CBH I a linear relationship was observed between the production of soluble sugars and adsorption, showing that the average activity of the bound CBH I molecules does not change with increasing saturation. For EG II the corresponding curve levelled off which is explained by initial hydrolysis of loose ends on Avicel. The enzymes competed for binding sites, binding of EG II was considerably affected by CBH I, especially at high concentration. CBH I produced more soluble sugars than EG II, except at conversions below 1%. At 40 degrees C when the enzymes were added together they produced 27-45% more soluble sugars than the sum of what they produced alone, i.e. synergistic action was observed (the final conversion after 48 h of hydrolysis was 3, 6, and 13% for EG II, CBH I, and their mixture, respectively). At 4 degrees C, on the other hand, when the conversion was below 2.5%, almost no synergism could be observed. Molar proportions of the produced sugars were rather stable for CBH I (11-15%, 82-89%, and <6% for glucose, cellobiose, and cellotriose, respectively), while it varied considerably with both time and enzyme concentration for EG II. The observed stable but high glucose to cellobiose ratio for CBH I indicates that the processivity for this enzyme is not perfect. EG II produced significant amounts of glucose, cellobiose, and cellotriose, which are not the expected products of a typical endoglucanase activity on a solid substrate. We explain this by hypothesizing that EG II may show processivity due to its extended substrate binding site and the presence of its cellulose binding domain.     

Pure enzyme cocktails tailored for the saccharification of sugarcane bagasse pretreated by using different methods

The compositions and physical properties of pretreated lignocellulose vary depending on pretreatment methods; therefore, enzyme cocktails specific to pretreatments are desired for efficient saccharification of lignocellulose. Here, enzyme cocktails consisting of three pure lignocellulolytic enzymes endoglucanase (EG), cellobiohydrolase (CBH) and endoxylanase (XN) with a fixed amount of β-glucosidase were tailored for acid- and alkali-pretreated sugarcane bagasse (ACID and ALKALI, respectively). Based on a mixture design, the optimal mass ratios of EG, CBH, and XN were determined to be 61.25:38.73:0.02 and 53.99:34.60:11.41 for ACID and ALKALI, respectively. The optimized enzyme cocktail yielded a higher or comparable amount of reducing sugars from the hydrolysis of ACID and ALKALI when compared to that obtained using commercial cellulase mixtures. Using the commercial and easily available pure enzymes, this simple method for the in-house preparation of an enzyme cocktail specific to pretreated lignocellulose consisting of only four enzymes with a high level of hydrolysis will be helpful for achieving enzymatic saccharification in the lignocellulose-based biorefinery.     

Identification of glycan structure and glycosylation sites in cellobiohydrolase II and endoglucanases I and II from Trichoderma reesei

Mass spectrometric techniques combined with enzymatic digestions were applied to determine the glycosylation profiles of cellobiohydrolase (CBH II) and endoglucanases (EG I, II) purified from filamentous fungus Trichoderma reesei. Electrospray mass spectrometry (ESMS) analyses of the intact cellulases revealed the microheterogeneity in glycosylation where glycoforms were spaced by hexose units. These analyses indicated that glycosylation accounted for 12-24% of the molecular mass and that microheterogeneity in both N- and O-linked glycans was observed for each glycoprotein. The identification of N-linked attachment sites was carried out by MALDI-TOF and capillary liquid chromatography-ESMS analyses of tryptic digests from each purified cellulase component with and without PNGase F incubation. Potential tryptic glycopeptide candidates were first detected by stepped orifice-voltage scanning and the glycan structure and attachment site were confirmed by tandem mass spectrometry. For purified CBH II, 74% of glycans found on Asn310 were high mannose, predominantly Hex(7-9)GlcNAc(2), whereas the remaining amount was single GlcNAc; Asn289 had 18% single GlcNAc occupancy, and Asn14 remained unoccupied. EG I presented N-linked glycans at two out of the six potential sites. The Asn56 contained a single GlcNAc residue, and Asn182 showed primarily a high-mannose glycan Hex(8)GlcNAc(2) with only 8% being occupied with a single GlcNAc. Finally, EG II presented a single GlcNAc residue at Asn103. It is noteworthy that the presence of a single GlcNAc in all cellulase enzymes investigated and the variability in site occupancy suggest the secretion of an endogenous endo H enzyme in cultures of T. reesei.     

Colloidal Gold Cytochemistry of Endo-1,4-beta-Glucanase, 1,4-beta-D-Glucan Cellobiohydrolase, and Endo-1,4-beta-Xylanase: Ultrastructure of Sound and Decayed Birch Wood

Colloidal gold coupled to endo-1,4-beta-glucanase II (EG II) and 1,4-beta-D-glucan cellobiohydrolase I (CBH I), isolated from Trichoderma reesei (QM9414), and endo-1,4-beta-xylanase from Aureobasium pullulans (NRRLY-2311-1) was used successfully to determine the ultrastructural localization of cellulose and xylan in sound birch wood. In addition, these enzyme-gold complexes demonstrated the distribution of cellulose and xylan after decay by three white rot fungi, Phanerochaete chrysosporium, Phellinus pini, and Trametes versicolor, and one brown rot fungus, Fomitopis pinicola. Transverse sections of sound wood showed that EG II was localized primarily on the S(1) layer of the secondary wall, whereas CBH I labeled all layers of the secondary wall. Oblique sections showed a high concentration of gold labeling, using EG II or CBH I. Preference for the sides of the microfibrillar structure was observed for both EG II and CBH I, whereas only CBH I had a specificity for the cut ends of microfibrils. Labeling with the xylanase-gold complex occurred primarily in the inner regions of the S(2) layer, S(1), and the middle lamella. In contrast, little labeling occurred in the middle lamella with EG II or CBH I. Intercellular regions within the cell corners of the middle lamella were less electron dense and labeled positively when EG II- and xylanase-gold were used. Wood decayed by P. chrysosporium or P. pini was delignified, and extensive degradation of the middle lamella was evident. The remaining secondary walls labeled with EG II and CBH I, but little labeling was found with the xylanase-gold complex. Wood decayed by T. versicolor was nonselective, and erosion of all cell wall layers was apparent. Remaining wall layers near sites of erosion labeled with both EG II and CBH I. Erosion troughs that reached the S(1) layer or the middle lamella had less xylanase-gold labeling in the adjacent cell wall that remained. Brown-rotted wood had very low levels of gold particles present in sections treated with EG II or xylanase. Labeling with CBH I had the lowest concentrations in the S(2) layer near cell lumina and corresponded to sites with the most extensive degradation.