Biochemical Characterization of an L-Xylulose Reductase from Neurospora crassa

An L-xylulose reductase identified from the genome sequenceof the filamentous fungus Neurospora crassa was heterologouslyexpressed in Escherichia coli as a His₆ tag fusion protein,purified, and characterized. The enzyme may be used in the productionof xylitol from the major pentose components of hemicellulosicwaste, D-xylose and L-arabinose.

INTRODUCTION

Top

Introduction

L-Xylulose reductase (LXR; EC 1.1.1.10) is the third of fiveenzymes, in fungi, involved in the assimilation of L-arabinoseinto the pentose phosphate pathway (1). As a member of the short-chaindehydrogenase/reductase (SDR) superfamily, the catalytic actionof LXR for the reduction of L-xylulose to xylitol requires aconcomitant oxidization of a nicotinamide cofactor. With theexception of that from the yeast Ambrosiozyma monospora, allknown LXRs are known to be strictly NADPH dependent (16). AlthoughLXRs are also found in higher eukaryotes such as humans, guineapigs, pigeons, and rats and catalyze the same reaction, theirmetabolic role is fundamentally different, realized in the glucuronicacid/uronate cycle of glucose metabolism (5, 15), and also thecatalytic oxidation of ${alpha}$ -dicarbonyl compounds (6, 12), ratherthan in L-arabinose metabolism.

Very few LXRs have been isolated and characterized to date;in fact, the identification of an LXR encoding gene in fungiwas discovered only recently in Trichoderma reesei (Hypocreajecorina) (13). In addition, LXRs have not found much commercialuse, and their natural substrate, L-xylulose, is relativelyexpensive. We realize that the potential of LXR could be actualizedin utilization of the pentose fraction hemicellulosic waste,which is primarily D-xylose and L-arabinose, for the productionof xylitol, a pentitol sugar alternative with several favorableproperties (7-11). Such a process would require concurrent utilizationof three enzymes: a xylose reductase, an L-arabinitol-4-dehydrogenase,and an L-xylulose reductase. To the best of our knowledge, afungal pathway for the co-utilization of D-xylose and L-arabinoseto produce xylitol has never been attempted, either in vitroor in vivo. To implement such a strategy, a highly active andstable LXR from a fungal source needs to be identified, whichhas not yet been done. Here we report the identification, heterologousexpression, purification, and characterization of a highly activeand stable LXR from Neurospora crassa.

Gene identification.

Top

Gene identification.

We postulated with the discovery of the highly catalyticallyefficient xylose reductase from Neurospora crassa (18) thatthe other enzymes involved the same L-arabinose utilizationpathway may also be extremely active. Using the protein sequenceof T. reesei LXR (GenBank accession no. AF375616.1) as a queryfor a BLASTP search (www.ncbi.nlm.nih.gov), the 271-amino-acidhypothetical protein NCU09041.1 in the N. crassa sequenced genome(3) was found as a putative LXR. A sequence alignment revealeda $~$ 79% identity at the amino acid level with the search query.In addition, the protein has the conserved SDR catalytic triadof serine, tyrosine, and lysine (14). A BLAST search for homologywith D-arabinitol dehydrogenases, enzymes with similar reactionsubstrate profiles, revealed $≤$ 38% identity. These findings suggestedthat this hypothetical protein is most likely an LXR.

Cloning, heterologous expression, and purification.

Top

Cloning, heterologous...

For experimental details, please see the supplemental material.PCR-amplified reverse-transcribed cDNA from L-arabinose-inducedN. crassa 10333 showed a specific high-yield product of theexpected gene size for lxr. This product was cloned into pET-28a(+)plasmid vector using NdeI and SacI restriction sites, whichincluded an N-terminal His₆ tag via a thrombin-recognized linkersequence. The circularized plasmid was electroporated into E.coli BL21(DE3) and selected on kanamycin-supplemented LB plates.

After confirming soluble, active enzyme expression by sodiumdodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)analysis using cell lysates of isopropyl-ß-D-thiogalactopyranoside(IPTG)-induced cultures, as well as by activity assay, the His₆-taggedLXR was purified by using single-step gravity immobilized metalaffinity column chromatography. The protein concentration wasdetermined by using the extinction coefficient as calculatedby San Diego Supercomputer Center Biology Workbench (http://workbench.sdsc.edu)as 32,670 M^–1 cm^–1 at 280 nm. The enzyme puritywas confirmed by SDS-PAGE analysis and stained with Coomassiebrilliant blue. The final yield of tagged LXR was 18 mg ( $~$ 60mg/liter of culture) of >95% pure LXR with a molecular massof $~$ 30 kDa, which corresponds well to the calculated combinedmass of $~$ 31 kDa for the His₆-LXR (Fig. 1). Although the yieldcould be further improved by optimizing fermentation techniques,the acquired quantity and purity is sufficient for characterizationpurposes. The His₆ tag was cleaved by incubating the taggedLXR with thrombin (Fig. 1), and it was found that cleaving thetag restored ca. 16% of the activity (data not shown). Therefore,the untagged LXR was used for all further analysis.

View larger version (38K):
[in this window]
[in a new window]

FIG. 1. SDS-PAGE analysis of purified, heterologously expressed N. crassa LXR. Lanes: 1 and 8, MW standard; 2, purified LXR after His₆ tag cleavage; 3, purified His₆-LXR fusion; 4 and 5, soluble and insoluble fractions, respectively, of cell lysate expressing His₆-LXR; 6 and 7, soluble and insoluble fractions, respectively, of cell lysate harboring empty plasmid (6). MW, molecular weight.

Steady-state kinetics.

Top

Steady-state kinetics.

Reaction kinetics were studied as for other NADP(H)-consumingoxidoreductases (18) by observing the initial change in absorbanceat 340 nm using a UV-Vis spectrophotometer (as described inthe supplemental material). All kinetic parameters for bothforward and reverse reactions were determined at 25°C in100 mM morpholinepropanesulfonic acid (MOPS; pH 7.0) and aresummarized in Table 1. Purified N. crassa LXR had a strong preferencefor NADPH, displaying activity below the level of detectionin the presence of NADH (see the description of high-performanceliquid chromatography [HPLC] analysis for NADH acceptance below).The enzyme does not catalyze the reduction of L-arabinose, D-xylose,D-glucose, and D-galactose or the oxidation of L-arabinitolat detectable levels, which is consistent with other characterizedLXRs (12, 13). Although the kinetic parameters for L-xyluloseand D-xylulose could not be determined because the reactionvelocity did not plateau in the range of concentrations tested,limited by cost of the expensive substrate, the K_m values seemto be unusually high. However, the enzyme displays the highestspecific activity toward L-xylulose compared to any other substratetested, including D-xylulose or D-ribulose (Table 1). LXRs havebeen known to have lower K_m values for D-ribulose than for L-xylulose(16). These data confirm that it is neither an D-arabinitoldehydrogenase nor an D-xylulose reductase but rather an LXR.The K_m values for all three pentulose substrates are at least1 to 2 orders of magnitude higher than for any other characterizedLXR, whereas the V_max is comparable to that of the yeast A.monospora but significantly higher than that of the fungus T.reesei (Table 2).

View this table:
[in this window]
[in a new window]

TABLE 1. Steady-state kinetics for N. crassa LXR^a

View this table:
[in this window]
[in a new window]

TABLE 2. Comparing properties of LXRs from fungi and yeast

HPLC analysis for NADH acceptance.

Top

Hplc analysis for nadh...

The acceptance of NADH as cofactor by N. crassa LXR was examinedby HPLC. The separation of NAD⁺ and NADH was carried out asdescribed elsewhere (19). We set up 100-µl reaction mixturesconsisting of 1 mM NADH and 10 mM D-ribulose in 100 mM MOPS(pH 7.0) and, after the addition of approximately 64.5 µgof enzyme, the reaction was allowed to proceed for 1 h at 25°C.Relative to the control reactions, this reaction yielded a significantlylarger peak, with a retention time corresponding to authenticNAD⁺ (see the supplemental material). This indicates that N.crassa LXR does accept NADH, although with a low average specificactivity of 0.083 µmol/min/mg compared to 7.8 µmol/min/mgfor NADPH as calculated for the same reactant concentrations.

Temperature dependence.

Top

Temperature dependence.

The optimal temperature for catalysis was determined by assayingLXR activities at temperatures ranging from 15 to 60°C byusing a recirculating water bath connected to the UV-Vis spectrophotometerwith a jacketed cuvette holder. The data show the optimum reactiontemperature to be approximately 37°C (Fig. 2A). At highertemperatures the enzyme rapidly inactivates, whereas at lowertemperatures the rate decreases with temperature in accordancewith Arrhenius' equation. The energy of activation for LXR wasdetermined to be 19.8 ± 1.1 kJ/mol by fitting the datafrom 15 to 30°C to the Arrhenius equation. N. crassa LXRwas found to be naturally stable, since it retained >85%activity after incubation at 25°C for over 6 h and >75%activity at 37°C after the same period of incubation (datanot shown). Thermal inactivation was studied by incubating LXRin a heating block with a heated lid at 45°C in 100 mM MOPS(pH 7.0). Samples were removed at various times to detect residualactivity at 25°C. The percent residual activity was plottedas a function of incubation time (Fig. 2B) and followed a first-orderexponential decay with a half-life of about 13 min.

View larger version (15K):
[in this window]
[in a new window]

FIG. 2. (A) The temperature optimum for the reductase activity of N. crassa LXR was determined to be $~$ 37°C when assayed between temperatures from 15 to 60°C. (B) Thermal inactivation of N. crassa LXR at 45°C. The heat inactivation followed first-order kinetics with a half-life of 12.9 min. (C) The pH optimum for the reductase activity ( ${blacksquare}$ ) of N. crassa LXR was at 7.0 and at 9.0 for oxidase activity ( ${triangleup}$ ). The enzyme shows a broad range for activity with >50% reductase activity between pH 4.5 and 8.5.

pH dependence.

Top

pH dependence.

The effect of pH on LXR's activity was studied in constant-ionic-strengthbuffers to prevent the introduction of artifacts due to differingbuffer components. For reduction, 2 mM D-ribulose and >200µM NADPH in either acetate-2-(N-morpholino)ethanesulfonicacid (MES)-Tris buffer (pH 4.5 to 8.0) or MES-3-(cyclohexylamino)ethanesulfonicacid (CHES)-Tris buffer (pH 7.0 to 9.5) were used. Similarly,for oxidation 1,000 mM xylitol and 2 mM NADP in either acetate-MES-Trisbuffer (pH 6.5 to 8.5) or Tris-CHES-3-(cyclohexylamino)-1-propanesulfonicacid (CAPS) buffer (pH 8.0 to 11.0) were used. The optimal pHwas determined to be 7.0 for reduction and 9.0 for oxidation(Fig. 2C). The enzyme demonstrated a broad range of activitywith >50% reductase activity between pH 4.5 and 8.5.

Determination of quaternary structure.

Top

Determination of quaternary...

To determine the quaternary structure, size-exclusion HPLC ona Bio-Sil SEC-250 column (300 by 7.8 mm) was performed by usingan Agilent 1100 series HPLC system with 100 mM Na₂HPO₄, 150mM NaCl, and 10 mM NaN₃ (pH 6.8) as the mobile phase. A Bio-Radstandard was used to standardize the column's retention timewith respect to molecular mass (see the supplemental material).The mass determined from the retention time was 62.2 kDa, closeto a 58.2-kDa estimated mass for an LXR homodimer with a GlyHisSerlinker remnant from the His₆ tag at the N terminus. Multimericstates are not uncommon, with several mammalian LXRs, includinghuman, being tetrameric (2).

Homology modeling.

Top

Homology modeling.

Using the coordinates (Protein Data Bank [PDB]; www.wwpdb.org)for the human LXR (PDB accession codes 1PR9 and 1WNT) (2) andthe NADP-dependent mannitol 2-dehydrogenase from Agaricus bisporus(PDB accession code 1H5Q) (4), a homology model was createdwith the Molecular Operating Environment (MOE; Chemical ComputingGroup). Templates were chosen on the basis that the human LXRis the only LXR with a solved crystal structure, and mannitol2-dehydrogenases are known to share high catalytic and sequencesimilarity with LXRs (13). The model was verified for consistencywith known protein folds and allowed ${phi}$ and ${psi}$ angles (see the supplementalmaterial). Compared to the human LXR, the N. crassa LXR is longerby 14 amino acids in the N terminus. As can be seen from Fig.3, the N. crassa LXR N terminus is modeled as an ${alpha}$ -helix thatis not present in the human LXR. The long distance of the N-terminalhelix from the catalytic site explains why the addition of alinker and a His₆ tag does not significantly affect the catalyticefficiency of the protein. In addition, the N. crassa LXR is27 amino acids longer, which would account for the longer loopregions. The catalytic triad Ser136, Tyr149, and Lys153 of humanLXR have similar locations and orientations in the N. crassaLXR. The differing orientation of NADP in the active site isprobably due to the steric effects between the bulky Tyr57 sidechain and the adenine group of NADP in N. crassa LXR. In thehuman LXR, Ser37 is not likely to cause similar steric effects.The low levels of activity toward NADH can possibly be explainedby the presence of hydrophobic Ala58 and Gly214 in the placeof Arg39 and Ser185 in human LXR that stabilizes the 2'-phosphategroup of NADPH. Although residues in the substrate binding site,other than the catalytic triad (Ser, Tyr, and Lys), are notstrongly conserved between human and N. crassa LXR, they remainin general moderately to highly hydrophobic, with the exceptionof Asp220 in N. crassa LXR in place of Tyr191 (6), which wouldmake the pocket more hydrophilic. However, the role of thisresidue seems unclear since it is conserved between the fungalLXRs, N. crassa and T. reesei, but not with the LXR from theyeast A. monospora, having a strongly hydrophobic Ile217 inthe corresponding position.

View larger version (39K):
[in this window]
[in a new window]

FIG. 3. Comparison of the crystal structure of human LXR (A) and homology model for N. crassa LXR with NADP (B). The catalytic triads (Ser, Tyr, and Lys) are shown and are approximately similar in position and orientation. Position of NADP differs in the two possibly because of Tyr57 in N. crassa LXR, offering steric hindrance to the adenine group of the cofactor. The longer N terminal is distant from the catalytic site, possibly accounting for minimal effect of the His₆ tag.

In conclusion, we have identified and characterized a highlyactive LXR from N. crassa that can be heterologously expressedand purified with high yield in E. coli. The enzyme was characterizedas stable and active over a wide range of temperatures and pH.Although it shows unusually high K_m values toward pentulosesubstrates, it displays one of the highest turnover numbers(k_cat) among characterized LXRs and may prove useful in theco-utilization of D-xylose and L-arabinose for the productionof xylitol or ethanol.

ACKNOWLEDGMENTS

Support for this research was provided by the BiotechnologyResearch and Development Consortium (project 2-4-121) and theDepartment of Chemical and Biomolecular Engineering at the Universityof Illinois at Urbana-Champaign.

Access to the Insight II and MOE programs was provided by TheUniversity of Illinois School of Chemical Sciences' ComputerApplication and Network Services. We thank Ryan Sullivan andRyan Woodyer for help with the HPLC analyses and Hua Zhao forhelp in cloning and characterizing the L-xylulose reductase.

FOOTNOTES

* Corresponding author. Mailing address: Department of Chemical and Biomolecular Engineering, University of Illinois at Urbana-Champaign, 215 RAL, Box C3, 600 S. Mathews Ave., Urbana, IL 61801. Phone: (217) 333-2631. Fax: (217) 333-5052. E-mail: zhao5{at}uiuc.edu.

Published ahead of print on 19 January 2007.

Supplemental material for this article may be found at http://aem.asm.org/.

REFERENCES

Top