DOI: 10.1002/cbic.200700765 Evolution in Reverse: Engineering a d-Xylose-Specific Xylose Reductase Nikhil U. Nair and Huimin Zhao* [a] Xylitol (1) is a pentitol and is used not only as a sweetener but also as a platform chemical for the production of industrially important chemicals. [1] As a sweetener, it has been shown to possess several favorable properties in comparison to other sugar substitutes, such as anticariogenicity, [2] good gastrointes- tinal tolerance, low caloric content, and minimal insulin de- pendence for metabolism. As an alternative to direct chemical reduction with gaseous hydrogen over Raney nickel catalyst, safer and environmentally-friendly biosynthetic routes of pro- ducing xylitol by fermentation and enzymatic reduction of d- xylose (2) into xylitol by using xylose reductases (XR) have also been studied extensively (Scheme 1). [3,4] However, the unspecif- ic nature of chemical reduction has not been addressed by the use of XRs. These enzymes have evolved to act as promiscuous aldose reductases and can reduce a number of pentoses and hexoses efficiently, of which l-arabinose (3) is of particular ACHTUNGTRENNUNGimportance. l-Arabinose, which occurs in abundance with its epimer d-xylose in plant hemicellulose, is difficult to remove, [5] and if left unpurified, will be reduced to l-arabinitol (4)—an unwanted byproduct. To the best of our knowledge, no one has attempted to implement methods to alleviate this issue, which remains one of the primary obstacles in the economical production of xylitol. We propose a “kinetic resolution” of the two reacting epimers by selective reduction of xylose from a mixture of sugars. Here, we present an engineered XR with partially reversed promiscuity, which results in increased prefer- ence of the enzyme for d-xylose over l-arabinose. Very few ex- amples exist in the literature of enzymes engineered to have narrowed substrate acceptance, and none for highly promiscu- ous sugar-utilizing enzymes. XRs from the yeasts Pichia stipitis and Candida tenuis are most popular for xylitol production; however, these enzymes have higher catalytic efficiencies toward l-arabinose than d- xylose (Table 1). [6,7] Consequently, we decided that the recently isolated fungal XR from Neurospora crassa (NcXR) was a better choice for engineering due to its innate 2.4-fold preference for d-xylose, high activity, and high expression level in E. coli. [8] Semirational-design approaches, targeted site-saturation muta- genesis (TSSM), and combinatorial active-site saturation testing (CASTing) have been successfully applied to shift the substrate specificity of the human estrogen receptor a LBD, [9] and to alter enantioselectivity or substrate scope of lipases, respec- tively. [10,11] Thus, we sought to use a similar method to engi- neer a d-xylose-specific XR. In addition, random mutagenesis by error-prone polymerase chain reaction (epPCR) was used to recognize possible contributions by distant residues through allosteric interactions. For TSSM, first-shell residues within interacting distance (ex- tended from a standard 4–5 to 8  for NcXR with a wide, sol- vent-accessible active site) were identified by docking and energy minimizing d-xylose and l-arabinose into a homology model. [8] Of the thirteen residues identified, two catalytic resi- dues (Y49 and K78) were not mutated. [8] D48, F112, and N307 were noted to be particularly important due to their proximity to C4 of the two sugars 2 and 3. Mutant S (F112S) was identi- fied after screening as one with maximum increase in substrate specificity (Table 2). A second round of TSSM and screening on the remaining ten residues in the mutant S background did not, however, result in the identification of any improved mu- tants (see the Experimental Section for details of screening). Mutagenesis by epPCR with this template followed by selec- Scheme 1. Reduction of d-xylose (2) to xylitol (1), and l-arabinose (3) to l- arabinitol (4) by xylose reductase (XR); the epimeric carbons are indicated with arrows. Table 1. Selectivities of two popular yeast XRs and that from N. crassa. Organism Selectivity [a] Neurospora crassa [8] 2.4 Pichia stipitis [7] 0.625 Candida tenuis [6] 0.5 [a] Selectivity = (k cat /K M ) xylose /ACHTUNGTRENNUNG(k cat /K M ) arabinose . [a] N. U. Nair, Dr. H. Zhao Department of Chemical and Biomolecular Engineering University of Illinois at Urbana–Champaign 600 S. Mathews Avenue, Urbana, IL 61801 (USA) Fax: (+ 1)217-333-5052 E-mail: zhao5@uiuc.edu Supporting information for this article is available on the WWW under http://www.chembiochem.org or from the author. ChemBioChem 2008, 9, 1213 – 1215 # 2008 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim 1213