Enzyme-Coupled Assay for {beta}-Xylosidase Hydrolysis of Natural Substrates

We describe here a new enzyme-coupled assay for the quantitationof D-xylose using readily available enzymes that allows kineticevaluation of hemicellulolytic enzymes using natural xylooligosaccharidesubstrates. Hydrogen peroxide is generated as an intermediaryanalyte, which allows flexibility in the choice of the chromophoreor fluorophore used as the final reporter. Thus, we presentD-xylose quantitation results for solution-phase assays performedwith both the fluorescent reporter resorufin, generated fromN-acetyl-3,7-dihydroxyphenoxazine (Amplex Red), and 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonate)(ABTS), whose corresponding radical cation has an absorbancemaximum at $~$ 400 nm. We also describe a useful solid-phase variationof the assay performed with the peroxidase substrate 3,3'-diaminobenzidinetetrahydrochloride, which produces an insoluble brown precipitate.In addition, kinetic parameters for hydrolysis of the naturalsubstrates xylobiose and xylotriose were obtained using thisassay for a glycosyl hydrolase family 39 ß-xylosidasefrom Thermoanaerobacterium sp. strain JW/SL YS485 (Swiss-Protaccession no. O30360). At higher xylobiose substrate concentrationsthe enzyme showed an increase in the rate indicative of transglycosylation,while for xylotriose marked substrate inhibition was observed.At lower xylobiose concentrations k_cat was 2.7 ± 0.4s^–1, K_m was 3.3 ± 0.7 mM, and k_cat/K_m was 0.82± 0.21 mM^–1 · s^–1. Nonlinear curvefitting to a substrate inhibition model showed that for xylotrioseK_i was 1.7 ± 0.1 mM, k_cat was 2.0 ± 0.1 s^–1,K_m was 0.144 ± 0.011 mM, and k_cat/K_m was 14 ±1.3 mM^–1 · s^–1.

INTRODUCTION

Top

Introduction

There is much current interest in enzymes that degrade starch,cellulose, and hemicellulose due to their increasing use inbiobleaching and for the conversion of biomass to chemical feedstocksand fuels (3, 12, 13). Our group has been pursuing gene discoveryand improvement of enzymes that degrade biomass through theuse of directed molecular evolution (7, 17-20), in which a libraryof mutated enzymes is screened to select for clones with improvedperformance under the screening assay conditions employed (10,14). One enzyme family of interest is the ß-xylosidases(EC 3.2.1.37). These enzymes are exo-type glycosidases thatremove xylose monomers from the nonreducing end of xylooligosaccharidesand serve as one component of a multienzyme milieu that biodegradeshemicellulose (13). Thus far, ß-xylosidase activityassays have relied mainly on the use of substrate analogs consistingof xylose O- linked to a chromophore or fluorophore, such asnitrophenyl or ß-umbelliferyl. While the substrateanalogs mentioned above are extremely useful for gene discoveryand protein structure-function studies (2, 9, 16), use of thesecompounds in directed evolution efforts could be misleading.Thus, one could select from an enzyme library variants withimproved activity for substrate analog hydrolysis that for xylooligosaccharideglycosidic bond hydrolysis are either inactive or not improved.The converse is also possible, so that enzymes in the librarywith high activity for hydrolysis of the native glycosidic bondare either inactive or not improved for hydrolysis of the substrateanalogs. Thus, for improved selection an in vitro screeningprocedure that emulates the anticipated final commercial-scalebioreactor process conditions is desired, and it requires aselection assay that allows monitoring of depolymerization ofnative xylan or xylooligosaccharide substrates to xylose monosaccharides.Various chemical methods to quantitate monosaccharide formationbased on the reducing capacity of the sugars are not usefulsince the natural substrates themselves contain a reducing groupand are usually present in large excess. We describe here anew enzyme-coupled xylose assay and an accompanying method tocharacterize a Thermoanaerobacterium sp. strain JW/SL YS485glycoside hydrolase family 39 ß-xylosidase (XylB;Swiss-Prot accession no. O30360) (9). In addition, this assayappears to be amenable to conversion to both solution and solid-phasehigh-throughput formats for use in gene discovery and directedevolution studies, and it is potentially useful as a reporteractivity assay for characterization of other hemicellulose-degradingenzymes, such as xylanases.

MATERIALS AND METHODS

Top

Materials and Methods

Plasmid DNA of p-Xylo 1.1 was isolated using a Qiaprep spinminiprep kit (QIAGEN). The plasmid DNA was digested using EcoRI(New England Biolabs [NEB], Beverly, MA) and electrophoresedusing a 1.5% low-melting-point agarose gel (Cambrex Biosciences,NJ), and the desired band containing xylB was excised and purifiedusing Zymoclean gel DNA recovery (Zymo Research, Orange, CA).PCR primers were designed to subclone the gene into a pET-22b(+)expression vector with a C-terminal His₆ tag (Novagen, San Diego,CA). The sequence of the forward primer was 5'-ATG ATT AAA GTGATA GTG CCA GAT TTT TCC-3', and the sequence of the reverseprimer was 5' GCG CTC GAG ATA TCC GTT TAT CTT GCT ATC ATC AAGTCC T-3'.

The reverse primer generated a XhoI site that was used for directionalcloning into the vector. The cloning strategy consisted of a5' blunt-end fragment and a 3' XhoI cohesive end. PCR amplificationwas performed by traditional methods using Pfu Ultra HotstartDNA polymerase (Stratagene, La Jolla, CA) and 0.5 µM ofeach primer. PCR generated a 1.5-kb fragment with a 5' bluntend and a 3' XhoI restriction site. The reaction product waspurified using DNA clean and concentrator-25 (Zymo Research,Orange, CA). A XhoI site was generated by restriction enzymedigestion, and the fragment was cleaned up using a Min Elutegel extraction kit (QIAGEN). The vector was prepared for cloningby restriction digestion with NdeI (NEB, Beverly, MA). The desiredband was purified using a Min Elute gel extraction kit (QIAGEN),and the 5' overhang was filled using T4 DNA polymerase (Promega,Madison, WI). After this, the pET-22b(+) vector was purifiedand cut with XhoI. The cleaned vector was dephosphorylated usingcalf intestine phosphatase (NEB, Beverly, MA). Ligation wasperformed using T4 ligase (NEB, Beverly, Madison) and a 5:1insert/vector ratio. Chemical transformation was done usingJM109 (Promega, Madison, Wisconsin) at 16°C overnight. Positiveclones were confirmed by colony PCR of plasmid preps (QIAGEN)using the primers mentioned above.

The expression vector was used to transform chemically competentEscherichia coli BL21(DE3) cells, and transformants were selectedby overnight growth at 37°C on Luria-Bertani agar platesamended with 75 µg/ml carbenicillin (LB_carb). Overexpressionwas achieved by incubating a single colony overnight at 37°Cin 10 ml LB_carb with 0.5% glucose, from which 8 ml was usedto inoculate 250 ml LB_carb. The culture was allowed to growto an optical density of 2 to 3 at 30°C, and then isopropyl-ß-D-thiogalactopyranoside(IPTG) (1 mM) was added and growth was allowed to occur foran additional 16 h at 30°C. Then 50-ml aliquots were pelleted,and each pellet was lysed at room temperature for 20 min byadding 3 ml of a Bug-Buster solution (Novagen 70584) containing1 µl/ml Benzonase (Novagen 712053), 1 µl/ml r-Lysozyme(Novagen 71114), 1 µl/ml protease inhibitor cocktail setIII (Calbiochem 539134), and 5 mM ß-mercaptoethanol.The supernatant solution was adjusted to contain 300 mM NaCl,10 mM imidazole, and 50 mM phosphate buffer (pH 8.0) beforebinding to Ni-nitrilotriacetic acid resin (QIAGEN 1018142) at4°C. The resin was transferred to a small column (Bio-Rad7311550) and washed with 50 mM phosphate buffer (pH 8.0) containing1 mM ß-mercaptoethanol, 1 µl/ml protease inhibitorcocktail, 300 mM NaCl, and 10 mM imidazole. The protein waseluted using the same buffer except that the imidazole concentrationwas increased to 250 mM. Fractions containing the enzyme werebuffer exchanged using NAP-5 desalting columns (Amersham 17085301)into 50 mM phosphate (pH 6.0), 10% glycerol, and 200 µMphenylmethylsulfonyl fluoride and stored at –80°C.Protein concentrations were obtained using Coomassie Plus reagent(Pierce 1856210) by following the manufacturer's protocol. Enzymefractions were analyzed using polyacrylamide gel electrophoresisby following the manufacturer's protocol (Invitrogen, Carlsbad,CA), and the final protein purity was estimated to be 73% fromgel densitometry using an AlphaImager imaging system (AlphaInnotech Corp., San Leandro, CA).

Enzymatic assay.
The solution-phase assay for ß-xylosidase naturalsubstrate hydrolysis activity is performed in two steps. Thefirst step is an endpoint assay of the ß-xylosidaseenzyme being evaluated, in which the glycosidic bonds of a solublexylooligosaccharide substrate (xylobiose or xylotriose in thisstudy; Wako Chemicals USA, Inc., Richmond, VA) are hydrolyzedby the enzyme, and D-xylose is generated. The ß-xylosidaseenzyme evaluated here is a family 39 retaining glycoside hydrolasecloned with a C-terminal His₆ tag from a Thermoanaerobacteriumsp. genomic library (Swiss-Prot accession no. O30360) (9). Inthe second step, the amount of xylose liberated is quantifiedusing an enzyme-coupled kinetic assay in which pyranose oxidase(pyranose:oxygen 2-oxidoreductase) converts the D-xylose hydrolysisproduct to 2-keto-D-xylose with concomitant formation of H₂O₂(equation 1). In a subsequent (more rapid) reaction catalyzedby horseradish peroxidase, the hydrogen peroxide generated oxidizesthe chromogen 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonate)(ABTS) (Sigma A-9941) to form the colored ABTS radical cation(equation 2), and the rate of formation is monitored at 420nm.

(1)

(2)
For liquid-phase assays we also employedN-acetyl-3,7-dihydroxyphenoxazine (Amplex Red; Molecular Probes,Eugene, OR) in place of ABTS, which reacts with a 1:1 stoichiometrywith H₂O₂ in a reaction catalyzed by peroxidase to produce highlyfluorescent resorufin, which was detected by excitation at 550nm and emission at 590 nm (22).

Time course studies were performed initially to determine thelinear range of the assay. It was established that the rateof substrate hydrolysis using either xylobiose or xylotrioseat a concentration of 4 mM was in each case linear (R² >0.99) for at least 90 min at 45°C. Generally, 16 differentsubstrate concentrations were used to assess the kinetic parameters,and the amount of enzyme was chosen so that the proportion ofsubstrate hydrolyzed ranged from 1 to 3% when xylobiose wasused at various concentrations as the substrate and from <1%to 7% when xylotriose was used at various concentrations asthe substrate. The judicious use of enzyme concentration tolimit substrate conversion, coupled with the previously establishedlinearity of enzyme activity with respect to time at the assaytemperature, served to clearly established that there was linearityof the hydrolysis rate for the substrate concentrations andreaction times used to assess the kinetic parameters. In a typicalsolution-phase assay 105 µl of 50 mM phosphate (pH 6.0)containing 0.1% bovine serum albumin (BSA) and the ß-xylosidaseenzyme (2.7 nM for the xylotriose assays and 18 nM for the xylobioseassays) was placed into PCR strip tubes and chilled to 0 to4°C. The reaction was initiated by adding 20 µl ofan appropriately diluted xylobiose or xylotriose solution, gentlemixing, and incubation at 45°C for 45 min using a PCR machine(MJ Research, Watertown, MA) preequilibrated at 45°C. Atthe end of the incubation period, the temperature was programmedto immediately ramp up to 100°C for 7 min to quench thereaction, followed by cooling to 0°C. This completed theß-xylosidase endpoint assay, and the samples werethen stored in a freezer or immediately analyzed using the xyloseenzyme-coupled quantitation assay. When the reaction being studiedwas catalyzed by a His-tagged enzyme and the enzyme was notreadily heat inactivated, it was also possible to quench thereaction by freezing on dry ice and to prevent further substratehydrolysis by adsorption of the enzyme at 0 to 4°C ontoNi-nitrilotriacetic acid resin (data not shown). No-incubationcontrols were included, in which the enzyme was heat inactivatedin the presence of the various substrate concentrations, followedby incubation in parallel with the assays. This compensatedfor small D-xylose impurities in the xylobiose and xylotriosesubstrates, as well as for any substrate hydrolysis that mayhave occurred either before the 45°C incubation temperaturewas reached or during the ramp-up of temperature involved inthe heat inactivation quenching step. After brief centrifugation,the contents of the PCR strip tubes were transferred to MilliporeMulti-Screen 96-well filtration plates (Millipore MAHVN4510)and centrifuged to obtain a filtrate free of precipitated protein,which would have interfered with the subsequent spectrophotometricassay. Aliquots (100 µl) of the filtrate were transferredto standard flat-bottom 96-well plates (Greiner Bio-One 655101)along with 100 µl of a reaction buffer consisting of 125mM phosphate (pH 7.4), 400 µM ABTS or 400 µM AmplexRed, 1.7 U/ml peroxidase (Sigma P6782), and 1 U/ml pyranoseoxidase (Sigma P4234). The plates were gently mixed for 30 s,and the rates were obtained at 25°C using a Spectramax M2plate reader (Molecular Devices Corporation, Sunnyvale, CA)from the initial slopes (200 to 800 s; R² > 0.99) of thechange in absorbance at 420 nm for ABTS or the change in fluorescence(excitation at 550 nm, emission at 590 nm) for Amplex Red. TheD-xylose concentration in the assays was calculated using aD-xylose standard curve generated on the same microtiter plate.

To compare the kinetic parameters obtained with the native substrateswith those obtained using chromophore-tagged xylose substrateanalogs, kinetic spectrophotometric assays were also performedusing the substrates o- and p-nitrophenyl-ß-D-xylopyranosideat 35°C in 50 mM phosphate (pH 6.0) containing 0.1% BSAand measurement of the change in absorbance at 400 nm. Ratesof hydrolysis were obtained from the initial slopes measuredfrom 60 to 400 s (R² > 0.99). The rate of nitrophenyl-ß-D-xylopyranosidehydrolysis was quantified using standard curves generated witheither the ortho- or para-nitrophenyl analog.

A solid-phase variation of the assay was also developed, inwhich the chromogen 3,3'-diaminobenzidine tetrahydrochloride(DAB) (Aldrich 261890), which in the presence of H₂O₂ in a peroxidase-catalyzedreaction produces an insoluble brown precipitate, is used. Totest the solid-phase assay, E. coli BL21(DE3)(pLysS) (Novagen,Madison, WI) was transformed either with empty pET-22b(+) vectoras a negative control (vector only, Fig. 1) or with the pET-22b(+)vector containing the ß-xylosidase gene insert asa positive control (ß-xyl, Fig. 1) and grown overnightin LB_carb. The cultures were then aliquoted into 96-well microtiterplates and robotically gridded (Q-Bot; Genetix USA Inc., Boston,MA) onto a Nytran SPC membrane (Schleicher & Schuell Bioscience,Keene, NH). The membrane was then placed onto LB_carb growthand induction agar containing 1 mM IPTG and incubated overnightat 30°C. The membrane was then carefully lifted from thegrowth and induction agar and placed face down onto a combinedlysis and indicating agar containing 50 mM phosphate, pH 7,1.5% (wt/vol) low-melting-point agar, 0.86 mg/ml mixed xylooligosaccharides(approximately 1 mM xylobiose and 1 mM xylotriose; Wako ChemicalsUSA, Inc., Richmond, VA), 1 U/ml peroxidase, 0.5 U/ml pyranoseoxidase, 1 mM DAB, 1 mg/ml hen egg white lysozyme (Sigma L7651),1 mg/ml polymyxin B sulfate (Sigma P1004), and 1 µl/mlprotease inhibitor cocktail III (Calbiochem 539134). The lysisand indicating agar was prepared by dissolving the agar in thephosphate buffer and allowing it to cool to less than 40°Cbefore the remainder of the components were added. Lysis andhydrolysis of the mixed xylooligosaccharides were allowed tooccur overnight at room temperature, and the membrane was analyzedusing an AlphaImager imaging system (Alpha Innotech Corp., SanLeandro, CA).

View larger version (22K):
[in this window]
[in a new window]
FIG. 1. Solid-phase assay of E. coli BL21(DE3) transformed with vector only or with the ß-xylosidase gene (ß-xyl). Bacterial colonies were grown and expressed in the presence of 1 mM IPTG at 30°C overnight on a Nytran SPC membrane placed on LB_carb. Cell lysis and detection were performed as described in the text by incubating the membrane overnight at room temperature using DAB as the peroxidase substrate. The image was obtained using an AlphaImager imaging system (Alpha Innotech Corp., San Leandro, CA).

RESULTS

Top

Results

The standard curve generated using the enzyme-coupled xyloseassay and ABTS as the chromophore was linear up to the highestconcentration of xylose tested (320 mM), and the limit of detectionwas in the low µM range. The assay worked equally wellusing Amplex Red, as shown in Fig. 2 for concentrations up to30 µM for Amplex Red and 60 µM for ABTS. The limitof detection of the Amplex Red fluorescence-based assay wason the order of 1 µM or less, and we were able to generatea linear standard curve from 0.1 µM to 1.0 µM xylosewith a correlation coefficient (R²) of >0.99 (data not shown).However, for use in the ß-xylosidase assay (see below)the sensitivity achieved using ABTS was more than sufficientsince the concentration of xylose released in the assays wasin the range from 3 to 330 µM. In order to determine thesuitability of the assay for detecting xylose in the presenceof a large xylooligosaccharide substrate concentration, thebackground signal was measured and found to be 2.5% and 1.5%for xylobiose and xylotriose, respectively. The observed backgroundmay have been due to the presence of trace levels of xylosein the xylooligosaccharide substrates, as revealed by thin-layerchromatography (data not shown). Also, it was found that thexylose standard curve generated in the presence of 12 mM xylobiosewas identical to that generated in the absence of xylobiose,while the xylose standard curve generated in the presence of12 mM xylotriose showed at most a 2% reduction in sensitivity.

View larger version (21K):
[in this window]
[in a new window]
FIG. 2. D-Xylose standard curves generated using ABTS as the chromophore or Amplex Red as the fluorophore. The error bars are generally about the size of the symbols. mOD/min 420 nm, 10^–3 unit of optical density at 420 nm/minute; rfu/sec Ex/Em 550/590 nm, relative fluorescence units/second at an excitation wavelength of 550 nm and an emission wavelength of 590 nm.

The kinetic data obtained for the nitrophenyl-ß-D-xylopyranosidesand xylobiose at high substrate concentrations all deviatedfrom Michaelis-Menten kinetics, presumably due to transglycosylation.Thus, increases in both k_cat and K_m values were observed withincreasing substrate concentration, and K_m and k_cat values weretherefore obtained using both low- and high-substrate-concentrationranges (Table 1). The low-substrate-concentration ranges usedto calculate the kinetic parameters were 10 µM to 250µM (0.11 to 2.8x K_m) for p-nitrophenyl-ß-D-xylopyranoside,6 µM to 375 µM (0.08 to 4.7x K_m) for o-nitrophenyl-ß-D-xylopyranoside,and 104 µM to 2.08 mM (0.03 to 0.63x K_m) for xylobiose.The high-substrate-concentration ranges used were 150 µMto 2,000 µM (0.19 to 2.5x K_m) for p-nitrophenyl-ß-D-xylopyranoside,150 µM to 2,000 µM (0.42 to 5.6x K_m) for o-nitrophenyl-ß-D-xylopyranoside,and 2.08 mM to 16 mM (0.20 to 1.6x K_m) for xylobiose. Kineticconstants were calculated by nonlinear regression fitting tothe Michaelis-Menten equation using the program GraFit 5 (6),as shown in Fig. 3A for xylobiose at low substrate concentrations.The nonlinear fit values obtained using either the lower- orhigher-substrate-concentration data for the xylopyranosidesand xylobiose were similar to those obtained using Lineweaver-Burkplots, as shown for the lower-xylobiose-concentration rangeused (Fig. 3A, inset). At low substrate concentrations it wasfound that the K_m values for the ß-D-xylopyranosideswere both around 0.1 mM, while the K_m for the natural substratexylobiose was appreciably higher, about 3.3 mM. In contrast,the k_cat values obtained under these conditions were similarfor the ß-D-xylopyranosides and xylobiose, rangingfrom 2.7 s^–1 to 4.2 s^–1 (Table 1). When the kineticparameters obtained at low substrate concentrations were comparedto those measured at high substrate concentrations (Table 1),it was found that the K_m values for o-nitrophenyl-ß-D-xylopyranosideand xylobiose showed three- to fivefold increases, while thek_cat roughly doubled. For p-nitrophenyl-ß-D-xylopyranoside,the K_m increased about 10-fold and the k_cat increased 3-fold,with the result that at high substrate concentrations the k_cat/K_mvalues were the same within the experimental error for botho- and p-nitrophenyl-ß-D-xylopyranosides.

View this table:
[in this window]
[in a new window]
TABLE 1. Michaelis-Menten parameters for hydrolysis of nitrophenyl-ß-D-xylopyranoside and xylooligosaccharide substrates by Thermoanaerobacterium saccharolyticum ß-xylosidase

View larger version (17K):
[in this window]
[in a new window]
FIG. 3. Reaction conditions were 45°C, 50 mM phosphate (pH 6.0), and 0.1% BSA. (A) Nonlinear regression fitting to the Michaelis-Menten equation of hydrolysis rate versus xylobiose concentration (only 0.104 mM to 2.08 mM shown) using 0.018 µM enzyme. (Inset) Lineweaver-Burk reciprocal plot of the data. (B) Nonlinear regression fitting to the Michaelis-Menten equation with substrate inhibition of hydrolysis rate versus xylotriose concentration using 0.0027 µM enzyme. R² = 0.993.

The substrate concentration range used for evaluating xylotrioseas a substrate was 52 µM to12 mM (0.36 to 84x K_m). Inthe course of xylotriose hydrolysis xylobiose is released, whichcan then be a substrate for hydrolysis. However, the K_m forxylobiose is 3.3 mM, and for the assay using xylotriose theextent of substrate conversion was such that the concentrationof the xylobiose released was in the range from 4 to 68 µM.Thus, the effect of subsequent xylobiose hydrolysis on the ratemeasured using xylotriose was minimal. Surprisingly, markedsubstrate inhibition was observed with xylotriose, necessitatingfitting of the data to a substrate uncompetitive inhibitionmodel in which an inactive ES₂ ternary complex is formed (equation3).

(3)
Nonlinear curve fitting of thexylotriose hydrolysis data to equation 3 using GraphPad Prism4 (GraphPad Software, Inc. San Diego, CA) provided estimatesfor the kinetic parameters, with an R² of 0.993 (Fig. 3B). Thus,the K_i for xylotriose was 1.7 ± 0.1 mM, the K_m was 0.143± 0.011 mM, the k_cat was 2.0 ± 0.1 s^–1,and k_cat/K_m was 14 ± 1.2 mM^–1 · s^–1(Table 1). It was found that the K_m for xylotriose was about23-fold lower than that for xylobiose, while the k_cat was similar,with the result that the k_cat/K_m for xylotriose was 17-foldhigher than that for xylobiose measured at lower substrate concentrations.

In the solid-phase variation of the assay utilizing DAB as theperoxidase substrate, an appreciable D-xylose content in themixed xylooligosaccharide substrate employed resulted in a higherbackground signal in the agar compared to the results when xylotriosewas used as the substrate. However, this did not appear to affectthe ability of the assay to detect positive colonies, and themixed xylooligosaccharide substrate was substantially cheaperthan the more chemically pure xylotriose substrate, making itclearly preferable for use in high-throughput screening applications.While clearly visible after 24 h (Fig. 1), the brown precipitatesignal was increased after 48 h. Also, development of the signalmay have been accelerated by incubating the plates at an elevatedtemperature. For the imaging step, we found that it was usefulto use a flat-bed scanner to image the bottom of the plate withappropriate adjustment of the contrast to minimize any backgroundinterference. The image was then flipped about the horizontalaxis using standard image-editing software to restore the orientationso that it matched that of the original gridding, and then itwas printed and finally scanned using an AlphaImager imagingsystem (Alpha Innotech Corp., San Leandro, CA).

DISCUSSION

Top

Discussion

An enzymatic coupled reaction that quantitatively measures D-xylosehas previously been reported for monitoring the consumptionof NADH using D-xylose isomerase and D-glucitol dehydrogenase(4). In the xylose quantitation method presented here a coupledenzyme strategy is also used, and this method has the advantagethat the enzymes are readily available and hydrogen peroxideis generated as an intermediary analyte, thereby allowing considerableflexibility in the choice of the chromophore or fluorophoreused as the final reporter. We found that the xylose assay presentedhere can be used to obtain useful relative kinetic parametersfor a series of natural substrates when it is used to measurethe ß-xylosidase catalytic activity as described above.Alternatively, the assay can be used to rank a series of ß-xylosidaseseither from naturally occurring enzyme sources or generatedin vitro.

Retaining xylosidases have been shown to follow a double-displacementmechanism (5) involving the formation of an enzyme-bound covalentintermediate (xylosylation; k₂), followed by base-catalyzedhydrolysis to form a ß-sugar hemiacetal product (dexylosylation;k₃) (equation 4) (2, 16, 21). Alternatively, another substratemolecule or other nucleophile can react with the covalent intermediateE-X in a transglycosylation (4)reaction (k₄). {zam0090558090004} By monitoring the hydrolysis of xylobioseby thin-layer chromatography, it has previously been shown thata similar ß-xylosidase from a Thermoanaerobacteriumsp. (94% amino acid identity; Swiss-Prot accession no. P36906)had xylobiose transglycosylation activity (8). Furthermore,detailed kinetic studies by Vocadlo et al. showed that for substrateanalogs with pK_as less than 9.0, dexylosylation (k₃) was therate-limiting step, and deviation from Michaelis-Menten kineticswas observed at higher substrate concentrations due to transglycosylation(16). For the similar ß-xylosidase in this study (Swiss-Protaccession no. O30360) we also observed deviation from Michaeliankinetics at higher o- and p-nitrophenyl-xylopyranoside substrateconcentrations (leaving group pK_as, 7.22 and 7.18, respectively),presumably due to transglycosylation. Moreover, transglycosylationwas also manifested in the kinetic data for xylobiose hydrolysis,consistent with the conclusion that dexylosylation (k₃) is therate-limiting step. The rate-limiting nature of the dexylosylationstep (k₃) in the xylobiose hydrolysis mechanism is corroboratedby the observation (Table 1) that the k_cat values are similarfor the nitrophenyl-ß-D-xylopyranoside substratesand the natural substrates, in spite of the large differencein the pK_as of the leaving groups (pK_a of xylose, $~$ 12.2). Ithas been pointed out that natural substrates can have k_cat valuesas good as artificial substrates with better leaving groups,owing to (evolutionary) optimization of the stability of thetransition state associated with their hydrolysis, as discussedbriefly in the review of Zechel and Withers (21).

Interestingly, significant substrate inhibition was observedusing the natural substrate xylotriose. Substrate inhibitionhas been observed previously for a Bacillus licheniformis 1,3-1,4-ß-glucanase(11) and a Streptomyces sp. strain QM-B814 ß-glucosidase(Bgl3) (15). At low substrate concentrations the ß-xylosidaseenzyme appears to be a more specific catalyst for hydrolysisof xylotriose than for hydrolysis of xylobiose, with a 17-fold-higherk_cat/K_m (Table 1), due mainly to a much lower K_m.

The solid-phase assay described above has the potential to effectivelyscreen genomic and in vitro mutagenized enzyme libraries, whichshould have applications in gene discovery and directed-evolutionstudies. This assay is particularly useful since previouslydescribed solid-phase ß-xylosidase assays involvedetection of the release of a chromophore or fluorophore moietyfrom the corresponding xylopyranoside substrate. The fluorophoresand chromophores employed tend to diffuse rapidly in the agar,making detection of bacterial colonies carrying the gene ofinterest problematic. Also, the use here of natural xylooligosaccharidesubstrates rather than artificial substrates results in increasedselectivity for detection of bona fide hemicellulase activity.Additionally, we demonstrated (data not shown) that when theexpression host is gridded in either a 96- or 384-well format,the membrane can be imaged and the pixel density can be readilyquantitated using an AlphaImager imaging system (Alpha InnotechCorp., San Leandro, CA). This is especially useful in screeninglibraries of enzyme variants since a list of candidates canbe generated for automated picking in a subsequent rearrayingstep. Moreover, it may be possible to add further screeningcriteria to the solid-phase assay to allow simultaneous screeningand selection for desired properties, such as protein expression,substrate specificity, thermal stability, and/or pH stability.

Recently, the regiospecificity of deacetylation of a seriesof monoacetylated 4-nitrophenyl-ß-D-xylopyranosidesubstrate analogs by a deacetylase activity has been evaluatedby using ß-xylosidase enzyme activity as a reporter(1). The ß-xylosidase assay described here in whichnatural substrates are used should extend this use of ß-xylosidaseas an auxiliary reporting enzyme to quantitation of the xylosegenerated from natural hemicellulose substrates by various mixturesof wild-type or mutant enzymes or, alternatively, the xyloseincorporation rate in the case of biosynthetic pathways. Thus,we anticipate that the extent of breakdown of xylan by a commercialxylanase preparation can be measured with the assay describedhere, using the resulting soluble xylooligosaccharide mixtureas the substrate source and monitoring the xylose monosaccharidereleased. This could be useful as a process development strategythat directs the in vitro evolution of multienzyme modules ofbiodegradation or biosynthetic pathways by measuring the concentrationof only the final or initial metabolite whose rate of formationor incorporation is to be optimized. Also, the hydrolyzabilityof various xylooligosaccharide substrate mixtures can be assessed,which is important in estimating the ability of a given ß-xylosidaseto efficiently hydrolyze potential xylooligosaccharide mixturesgenerated by upstream hemicellulolytic steps.

In conclusion, here we describe a sensitive enzyme-coupled assayfor xylose in which readily available enzymes that can be coupledto a large selection of chromophores or fluorophores are used.We have also demonstrated the utility of this assay in ß-xylosidasesolution and solid-phase assays using natural xylooligosaccharidesubstrates. Work is in progress to employ these assays in ahigh-throughput format to screen ß-xylosidase enzymelibraries for mutants with improved stability and kinetic propertiesfor use in the enzymatic breakdown of hemicellulose.

ACKNOWLEDGMENTS

We gratefully acknowledge the generous gift of the Thermoanaerobacteriumsp. strain JW/SL YS485 ß-xylosidase gene (NationalCenter for Biotechnology Information accession no. AAB68820)from Juergen Wiegel, Department of Microbiology, Universityof Georgia, Athens, and W. Walter Lorenz, Warnell School ofForest Resources, University of Georgia, Athens, and the manyhelpful corrections and suggestions of the reviewers.

Reference to a company and/or products is for purposes of informationand does not imply approval or recommendation of the productto the exclusion of others which may also be suitable. All programsand services of the U.S. Department of Agriculture are offeredon a nondiscriminatory basis without regard to race, color,national origin, religion, sex, age, marital status, or handicap.

FOOTNOTES

* Corresponding author. Mailing address: USDA Agricultural Research Service, Western Regional Research Center, 800 Buchanan Street, Albany, CA 94710. Phone: (510) 559-6453. Fax: (510) 559-5940. E-mail: kwagschal{at}pw.usda.gov.

REFERENCES

Top