Understanding the structure and composition of recalcitrant oligosaccharides in hydrolysate using high-throughput biotin-based glycome profiling and mass spectrometry

Novel Immunological and Mass Spectrometry Methods for Comprehensive Analysis of Recalcitrant Oligosaccharides in AFEX Pretreated Corn Stover. Lignocellulosic biomass is a sustainable alternative to fossil fuel and is extensively used for developing bio-based technologies to produce products such as food, feed, fuel, and chemicals. The key to these technologies is to develop cost competitive processes to convert complex carbohydrates present in plant cell wall to simple sugars such as glucose, xylose, and arabinose. Since lignocellulosic biomass is highly recalcitrant, it must undergo a combination of thermochemical treatment such as Ammonia Fiber Expansion (AFEX), dilute acid (DA), Ionic Liquid (IL) and biological treatment such as enzyme hydrolysis and microbial fermentation to produce desired products. However, when using commercial fungal enzymes during hydrolysis, only 75–85% of the soluble sugars generated are monomeric sugars, while the remaining 15–25% are soluble recalcitrant oligosaccharides that cannot be easily utilized by microorganisms. Previously, we successfully separated and purified the soluble recalcitrant oligosaccharides using a combination of charcoal and celite-based separation followed by size exclusion chromatography and studies their inhibitory properties on enzymes. We discovered that the oligosaccharides with higher degree of polymerization (DP) containing methylated uronic acid substitutions were more recalcitrant towards commercial enzyme mixtures than lower DP and neutral oligosaccharides. Here, we report the use of several complementary techniques that include glycome profiling using plant biomass glycan specific monoclonal antibodies (mAbs) to characterize sugar linkages in plant cell walls and enzymatic hydrolysate, matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-MS) using structurally-informative diagnostic peaks offered by negative ion post-secondary decay spectra, gas chromatography followed by mass spectrometry (GC–MS) to characterize oligosaccharide sugar linkages with and without derivatization. Since oligosaccharides (DP 4–20) are small, it is challenging to mobilize these molecules for mAbs binding and characterization. To overcome this problem, we have applied a new biotin-coupling based oligosaccharide immobilization method that successfully tagged most of the low DP soluble oligosaccharides on to a micro-plate surface followed by specific linkage analysis using mAbs in a high-throughput system. This new approach will help develop more advanced versions of future high throughput glycome profiling methods that can be used to separate and characterize oligosaccharides present in biomarkers for diagnostic applications.


Results and discussion
Recalcitrant oligosaccharides are a common problem for complete enzymatic hydrolysis of lignocellulose. Our previous study found that recalcitrant oligosaccharides remained unhydrolyzed after pretreatment and enzymatic hydrolysis (Fig. 1). In our previously published work, a solid-phase extraction method employing activated charcoal was developed to separate oligosaccharides from AFEX-pretreated corn stover hydrolysate (ACSH) 8 . After the initial extraction and separation, oligosaccharides were further fractionated by Size Exclusion Chromatography (SEC) and collected in the order of molecular weight. Sugar monomers and oligomers released by different pretreatments were analyzed by sugar composition analysis. By comparing the sugar oligomer content resulting from different pretreatment methods, the existence of recalcitrant oligosaccharides is a common problem in biomass conversion to monomeric sugars and can cause at least 10-15% sugar yield loss, and even up to 18% in the case of ACSH. This method was used to further to produce oligosaccharides ELISA-based mAb screening of biotinylated oligosaccharides. Glycome profiling has been proven to be a useful tool for comprehensive glycan structure analysis on extracts isolated from solid biomass residues. However, water-soluble sugars were underrepresented using this routine method 41 , as the small molecule oligosaccharides were challenging to be immobilize on the ELISA plates and were washed away before antibodies were added. Thus, a one-step biotinylation technology was applied to enable the coating of soluble recalcitrant oligosaccharides on the Avidin-coated ELISA Plates, enabling the binding with antibody and characterization. The method has been tested using our previously produced ACSH and its fractions based on molecule weight (or degree of polymerization, DP). One-step biotinylation was used to enhance the binding affinity of oligosaccharides (Fig. 2), by adding biotin-LC-hydrazide to the reducing ends of carbohydrates 42 . In solution, the hemiacetal group at the reducing end reacted with the hydrazide group of biotin-LC-hydrazide to form a hydrazone linkage. In the presence of the reductant NaCNBH 3 , the hydrazone linkage was reduced to become a stable biotinylated end-product. With the sugar reducing end modified, the binding of low-DP oligosaccharides to ELISA plates was then possible and was performed on Avidin-coated plates with glycan-directed mAbs in our study. The Avidin-coated plates with oligosaccharides binding were then submitted to primary and secondary antibody addition and wash in a light sensitive and time sensitive environment. After the antibody binding was completed, TMB substrate was added to incubate the plates. Finally, the reaction was terminated by sulfuric acid. The incubated plates were analyzed using ELISA reader to determine the binding intensity for each antibody and hence probing against antibody-specific cross-linkages. For experiment details and parameters, please refer to the corresponding "Materials and Methods" section.
Recalcitrant oligosaccharides epitopes being detected using cell wall glycan-directed mAbs. We proved the value of this newly developed technique in a specific application by characterization of soluble oligosaccharides present in ACSH and in crude and purified oligosaccharides fractions separated from lignocellulosic hydrolysates. As shown in Fig. 3, the most abundant epitopes identified in ACSH using the bionylated glycome profiling method are substituted xylan are usually by uronic acid (U) or methyl uronic acid (MeU) and pectic arabinogalactan. Most of these were also identified in our previous glycome-profiling studies on unhydrolyzed solids (UHS) 43 .
These non-cellulosic structures cannot be digested by the most abundant cellulases and hemicellulases in the tested commercial enzyme cocktails, which have included most of the commonly used commercial enzymes. Here, (A) Methodology of oligosaccharides separation where, oligosaccharides were separated from AFEXpretreated corn stover Hydrolysate (ACSH) by activated charcoal and celite packed bed; (B) methodology of oligosaccharides separation. Oligosaccharides were further fractionated by size exclusion chromatography (SEC); (C) sugar monomers and oligomers released by different pretreatments (dilute acid: DA, ionic liquid: IL and AFEX). Enzymatic hydrolysis conditions: High solids loading 25% (w/w) (approximately to 8% glucan loading), 96 h hydrolysis, 20 mg/g commercial enzyme loading (Ctec2: Htec2: MP-2:1:1 ratio) and (D) Sugar monomers and oligomers of glucose, xylose and arabinose released by AFEX-pretreated corn stover (ACS).   www.nature.com/scientificreports/ Their hydrolysis therefore requires novel accessory enzymes. Without the needed non-cellulosic accessory enzymes, these non-cellulosic linkages hinder complete conversion to monomeric sugars, even though their parent sugar polymers were extensively hydrolyzed into shorter fragments and solubilized using commercial enzyme cocktails. Further examining the distribution of signals and their binding intensity, we observe that the binding epitopes were more prominent in high DP sugar fractions (A, B, C, DP as high as 20+) than in low DP fractions (D, E, F, DP as low as dimer) (Fig. 1). The acidic fractions were more abundant in non-cellulosic epitopes than neutral fractions. These phenomena were consistent with the patterns observed in our previous study 8 , in which high DP and acidic fractions were more resistant to enzyme hydrolysis. Thus, the existence of non-cellulosic glycan epitopes and U and MeU substitutions might contribute significantly to oligosaccharide recalcitrance. It should be noted that binding and detection effectivity for low DP oligosaccharides may have issues particularly if epitopes are dimer or trimer oligosaccharides. This could potentially be verified using commercial oligosaccharides of different lengths which all contain only one epitope that bind to specific mAbs.
The use of structure-specific antibodies has thus identified the certain types of recalcitrant linkages. Based on the types of antibodies being employed, the corresponding linkage pattern and their resulting signal intensity (most and least abundant), new enzymes can be identified and added into enzyme cocktails semi-quantitatively for a more complete sugar conversion. Using ACSH oligosaccharides profiling as an example, we can generate a database of glycan linkages for each biomass material. It should be mentioned here that different affinities of the antibodies should be considered, resulting in certain challenges in comparing between different antibody signals if their affinities are unknown. Also, the glycan linkages comparison may work best between samples for a single antibody. These recalcitrant linkages can then relate to the CAZyme data base, from which we can identify enzymes, select candidates, and screen enzymes to break the bonds, or develop microbial system to express these enzymes for use in a biorefinery 44 .

Glycosyl-residue composition analyses of recalcitrant oligosaccharides.
To evaluate how immunological methods can complement alternative characterization methods for low molecular weight oligosaccharides present in lignocellulosic hydrolysates, we performed MALDI (Fig. 4, S1-S8) and GC-MS based TMS derivatized sugar analysis (Fig. 5) on the same sets of oligosaccharide fractions. MALDI was used to compare whether the mass distribution of oligosaccharides molecules fit the hypothesized structure. Figure 4 shows the MS of neutral fraction ACN-A and ACN-B. Analyses of ACN-A confirmed Pentose series from DP 4-8 ( Fig. 4) up to DP 22 (Fig. S1), the mass of which corresponded with MeU-Xylan oligosaccharides. Analyses of ACN-B confirmed pentose and gluco-xylan series with DP 8-15. In supplementary materials, e.g., Fig. S3, the mass distribution profile of acidic fraction FA-C revealed a (Me)U substituted pentose series with DP 8-15, which was consistent the substituted xylan epitope detected in ELISA based mAb screening.
Another interesting finding comes from the LC-MS analysis of the oligosaccharides fractions as shown in Fig. S9 (Methods can be seen in Electronic Supplementary Materials). Fragments of hexose and -OAc groups are repeatedly observed in ACN-B fraction linkages. This finding not only confirms the fragmentation observed in glycome profiling and MALDI-TOF, but also provides new information on potential carbohydrate derivatives in the pretreated lignocellulosic biomass.
We also performed sugar composition analysis of oligosaccharides fractions using TMS sugar derivatization. Using GC-MS, we determined the composition of both neural (non-derivative) and acidic sugars (GluA and GalA) in oligosaccharide fractions (Fig. 5). Glucoronic acid was found in acidic fraction C and D while galacturonic acid was found in acid fraction A and B, all of which were high-DP acidic sugar fractions. These results not only support our ELISA and MALDI data, but also are consistent with our previous study on oligosaccharides accumulation. Thus, we are confident that an advanced immunological method using oligosaccharides biotinylation and subsequent ELISA screening was adequate to identify soluble recalcitrant oligosaccharides in different biological samples.
Using commercial oligosaccharides to determine the detection range of glycan-directed mAbs. As the ELISA-based mAb screening method has been verified by several different methods, we wanted to further explore the potential of this novel method for quantification. Two commercial oligosaccharides, xylohexaose oligosaccharide (XHE) and 2 3 -α-L-arabinofuranosyl-xylotriose (A2XX), were purchased and tested using the novel cell wall glycan-directed mAbs method. Figure 6 shows the linear correlation between the biotinylated binding signal and the logarithm concentration of oligosaccharides concentration, indicating a possible Langmuir adsorption model. Among the mAbs, CCRC-M137, CCRC-M138, CCRC-M147, CCRC-M148 and CCRC-M151 showed correlation against XHE, and CCRC-M108, CCRC-M109 and LM11 showed correlation against A2XX in the range of 1 nm to 100 nm. Due to the limited availability of antibodies at the time of experiment, limited experiments were performed for each concentration of oligosaccharides. It should be noted here that for the same oligosaccharides as substrate, some antibodies respond very differently, clearly because they bind to slightly different epitopes and may have very different binding affinity. When applying the novel mAbs method to a real sample, the mechanism and consequence of what the epitopes recognize exactly will be far more complicated.

Conclusion
Use of glycan-directed monoclonal antibodies (glycome profiling or ELISA-based mAb screening) is a powerful tool for in-depth characterization of most major cell wall glycans constituting plant biomass. Classical glycome profiling, however, only facilitates characterization of larger cell wall-glycans, because most oligosaccharides are not efficiently immobilized on ELISA plates. In this study, AFEX pretreated corn stover was subjected to www.nature.com/scientificreports/ enzymatic hydrolysis at high solid-loading. Glycome profiling was employed to determine the composition of recalcitant cell wall carbohydrates in the hydrolysate. Analysis of smaller oligosaccharides in the hydrolysate using mAbs, however, was underpresented and requiring additional tools to efficiently immobilize oligosaccharides on ELISA plates. We report here a novel efficient method to immobilize oligosaccharides for mAb screening by combining oligosaccharide biotinylation with the subsequent ELISA screening on NeutrAvidin™ coated plates. The immobilized biotinylated oligosaccharides showed sufficient affinity for the antibodies enabling rapid and efficient detection of recalcitrant oligosaccharides. Mass spectrometry based composition analysis of these recalcitrant oligosaccharides confirmed the results from this novel immuno-screening method. These studies thus demonstrate that combining oligosaccharide biotinylation and ELISA screening with glycan-directed mAbs is useful in identifying the cross-linkages in oligosaccharides and might be widely applied to other biochemical studies characterizing oligosaccharide structures.
This biotin-based glycome profiling method is the first report to enable probing recalcitrant carbohydrate linkages of soluble oligomeric sugars in plant biomass. This has help to understand why certain portion of biomass is highly recalcitrant while producing biofuels. This method fills an important gap in glycome profiling www.nature.com/scientificreports/ technology and extends its application to a broader spectrum of substrates other than plant oligosaccharides.
In the future, we can use robot to do biotynylation and analyze the samples using ELISA in a high-throughput fashion using the methods we developed.

Biomass. Corn Stover (CS) grown from Pioneer hybrid seed variety 33A14 was harvested in 2010 from
Kramer farm in Wray, CO. Permission was obtained from the land owner to use this biomass for research. The samples were stored dry < 6% moisture in zip-lock bags in room temperature. The study complies with local and national guidelines. Composition analysis was performed using the NREL protocol 45 46 . It took 30 min. for the reactor to reach 140 °C and the ammonia was rapidly released, which quickly brought the biomass to room temperature. The AFEX-pretreated corn stover (ACS) composition was similar to untreated corn stover (UT-CS).
High solids loading enzymatic hydrolysis. High solids loading 25% (w/w) (approximately to 8% glucan loading) ACSH was prepared as starting material for the large-scale production of oligosaccharides. Enzymatic hydrolysis of ACS was performed using a commercial enzymes mixture including Cellic® Ctec2 10 mg  Oligosaccharide analysis. Oligomeric sugar analysis was conducted on the hydrolysate liquid streams using an autoclave-based acid hydrolysis method at a 2 mL scale. Hydrolysate samples were mixed with 69.7 μL of 72% sulfuric acid in 10 mL screw-cap culture tubes and incubated in a 121 °C bench-top hot plate for 1 h, cooled on ice and filtered into High Performance Liquid Chromatography (HPLC) vials. The concentration of oligomeric sugar was determined by subtracting the monomeric sugar concentration of the non-hydrolyzed samples from the total sugar concentration of the acid hydrolyzed samples.
Analytical method. Glucose, xylose and arabinose concentrations in acid hydrolyzed biomass were analyzed using a Shimadzu HPLC system equipped with a Bio-Rad Aminex HPX-87H column equipped with automatic sampler, column heater, isocratic pump, and refractive index detector. The column was maintained at 50 °C and eluted with 5 mM H 2 SO 4 in water at 0.6 mL/min. flowrate.
Liquid and solid composition analysis. The hydrolysate supernatants were diluted and analyzed for monomeric and oligomeric sugar contents. Monomeric sugars produced after enzyme hydrolysis were analyzed using an HPLC equipped with a Bio-Rad (Hercules, CA) Aminex HPX-87P column and de-ashing guard column. Column temperature was held at 80 °C and water was used as the mobile phase flowing at 0.6 mL/min. Oligomeric sugars were determined via dilute acid hydrolysis at 121 °C according to the method reported in the literature 41,48,49 . Glycome profiling. Glycome profiling of untreated, AFEX-pretreated and all unhydrolyzed biomass residues (involving preparation of sequential cell wall extracts and their mAb screenings) were carried out using the procedures previously described 27,43,50,51 . To conduct glycome profiling, alcohol insoluble residue of plant cell wall materials were prepared from biomass residues and were subjected to sequential extractions with increasingly harsh reagents such as ammonium oxalate (50 mM), sodium carbonate (50 mM with sodium borodeutiride at 0.5% w/v), KOH (1 M and 4 M, both containing sodium borodeutiride at 1% w/v) and acidic chlorite as described previously 52,53 . The extracts were then subjected to ELISA against a comprehensive suite of cell wall glycan-directed mAbs 50 and the mAb binding responses were represented as heat maps. Plant cell wall glycandirected mAbs were purchased from laboratory stocks (CCRC, JIM and MAC series).

ELISA-based mAb screening of biotinylated oligosaccharides.
1. One-step biotinylation of oligosaccharides. The coupling of carbohydrates to biotin-LC-hydrazide was carried using the following procedure. Biotin-LC-hydrazide (4.6 mg/12 μmol) was dissolved in dimethyl sulfoxide (DMSO, 70 μl) by vigorous mixing and heating at 65 °C for 1 min. Glacial acetic acid (30 μl) was added, and the mixture was poured onto sodium cyanoborohydride (6.4 mg/100 μmol), which dissolved completely after heating at 65 °C for approximately 1 min. Then 5 to 8 μl of the reaction mixture was added to the dried oligosaccharides (1-100 nmol) to obtain a tenfold or greater molar excess of label over reducing ends. The reaction was carried out at 65 °C for 2 h, after which the samples were purified immediately.

TMS sample preparation.
A 100-500 μg of sample was added into a screw cap tube on analytical balance.
Record the amount added. It is best to have the samples dissolved in solvent at certain concentration, added into the tube as liquid aliquots. A 20 μL of 1 mg/mL myo-inositol was used as internal standard into each sample tube. The amount of internal standard added to sample must be the same as that added into the standard tube.

Silylation.
A 200 μL Tri-Sil was added and the capped tubes were heated for 20 min. at 80 °C and then cooled to room temperature. The samples were further dried using drying manifold to a volume of ~ 50 μL. It is important to note that we did not allow the sample to completely dry.

Gas chromatography (GC) samples preparation and injection. A 2 mL hexane was added and mix
well by vortexing. Pasteur pipette tip (5-8 mm) was pack a bit of glass wool by inserting the glass wool from the top of 5-3/4 -inch pipette. The samples were centrifuged at 3000×g for 2 min. to precipitate any insoluble residue. The samples were dried until it reached 100-150 μL. About 1 µl volume was injected into the GC-MS with initial temp 80 °C and initial time 2.0 min ( Table 2).
Calculation. If running standards for first time, perform TMS derivatization of each sugar separately and run them individually to procure the profiles of retention times, spectra, and distributions of peaks for each monosaccharide. These profiles will be used to identify the monosaccharide in standards mixture and samples. From the GC-MS, obtain the peak area and retention time of each sugar in standards and samples. Assign the corresponding sugar identities to the unknown peaks in samples according to the standards retention time.
Because each sugar can have as many as 4 peaks, in total all the area from peaks corresponding to 1 sugar.  www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.