Pyranose dehydrogenase for biofuel cells

Enzymatic biological fuel cells (biofuel cells) are electrochemical devices that transform chemical to electrical energy via enzymatic catalysis. Microscale enzymatic biofuel cells are a promising future power source for implantable medical devices such as the cardiac pacemaker or prosthetic applications (artificial hearing or vision), as well as for power generation from ambient fuels for small electronic devices. In enzymatic biofuel cells, suitable redox enzymes are connected to electrodes by redox active substances (mediators) such as osmium redox polymers. Glucose oxidase is most commonly used as a model enzyme. There are, however, a number of alternative carbohydrate oxidoreductases that offer distinct advantages over glucose oxidase. One of them is pyranose dehydrogenase (PDH), a flavin-dependent oxidoreductase from litter-decomposing Agaricaceae. PDH is unable to utilize oxygen as electron acceptor and is capable of oxidizing a broad variety of monosaccharides, oligosaccharides and glycosides. Additionally, PDH oxidizes glucose and other sugars at more than one C-atom, increasing the attainable electron output per molecule of substrate. The suitability of PDH for biofuel cell applications has already been demonstrated, resulting in a biofuel cell with increased coulombic efficiency. We propose to engineer and improve certain properties of PDH for use in biofuel cells employing random strategies as well as semi-rational protein engineering. A mutant library will be constructed by a combination of error-prone PCR and gene shuffling (recombination), and screened in microtiterplate-format using a standard assay with suitable modifications for improved catalytic activity, improved stability of the biocatalyst, improved activity at lower temperatures and improved activity with primary oxidation products of PDH and other enzymes such as gluconic acid, 2-keto-glucose and 2-keto-gluconic acid. Selected variants with improved properties will be characterized with respect to several properties, especially kinetic constants for various electron acceptors such as cycloruthenate or complexed Os3+. We will also employ semi-rational protein engineering in order to improve certain properties of PDH. As soon as a three-dimensional structure of PDH is available, amino acid residues at the active site of PDH will be selected and subject to saturation mutagenesis and screening of the resulting mutant libraries for improvement of the above mentioned properties. Additionally we will target amino acid residues that display a high flexibility upon thermal motion by saturation mutagenesis, and screen for increased thermostability as a result of amino acids conferring increased rigidity. Variants of interest will also be characterized electrochemically using redox polymers with different redox potentials, and prototype biofuel cells consisting of a PDH-based anode and a suitable cathode will be constructed and their performance tested.

The project resulted in the construction of several variants of pyranose dehydrogenase with changed glycosylation patterns and superior properties on electrodes of the anodic side of bio-fuelcells or biosensors, due to improved electron transfer to the electrode in the absence of certain glycosylation chains. We exchanged the four asparagines that are naturally glycosylated against all other possible amino acids, and determined which variants retain their catalytic activity. This was first done with single sites, and led to the observation that two sites (N75 and N252) did not alter the biochemical properties of PDH when exchanged against glutamin or glycin, which cannot be glycosylated. Another site shows a minor effect on stability (N175), and the last one cannot be exchanged at all, presumably because the glycosyl chain attached to this asparagine (N319) serves a non-dispensable function during production of the protein. One of the three exchangable sites, N252, located more distantly to the active site, is responsible for the significant over-glycosylation that is observed in strains of Pichia pastoris such as X-33. The other two, N75 and N175, are situated in vicinity to the site where the carbohydrate substrate accesses the active site, and where electrons are transferred to the electrode, either directly or by a mediating substance. Abolishing glycosylation on either of these sites leads to higher current densities to the electrode. Abolishing both has a significantly higher effect and produces current densities that are almost tenfold compared to the wild-type enzyme, due to the reduced distance of the active site to the electrode surface. Additionally, this modification allows the electrons to pass to the electrode surface directly, without the use of a mediating substance that "shuttles" the electrons to the electrode. This Direct Electron Transfer has not been observed before for simple flavoproteins. We also constructed a variant lacking N75, N175 and N252, which did not show further improvements, but was less stable than the others. We investigated the production of these enzyme variants, and found that the best version, lacking both glycosylation chains near the active site (N75 and N175), can be produced well in recombinant Pichia pastoris, but at reduced volumetric yields. Use of an alternative strain, that adds modified glycosylation chains to the respective asparagines, did not bring the expected results, as the modification did not occur homogenously, and the resulting enzymes did not, unlike the variants lacking glycosylation, perform better on electrodes than the wild- type. Further efforts to modify the enzyme so that it can be produced in E. coli (lacking glycosylation entirely) did not meet with success. Work done in this project has led to substantial interest in PDH and related enzymes for bio- fuelcell or biosensor applications, and further developments are in discussion, also with potential industrial partners.