Gr/S25203/01

Final Scientific Report for grant ref GR/S25203/01- Tailored ionic liquids as solvents for biocatalysts.

Background/Context
Enzymes offer a number of major advantages over standard chemical catalysts, in terms of their stereoselectivity, efficiency and ability to perform difficult reactions. However, their use in industry has been impeded by their requirement for water as a solvent. Many of the key precursors and products in pharmaceuticals and fine chemicals processes are very poorly soluble in water and in certain cases may be hydrolysed by it. This solubility problem by extension leads to poor overall yields and very slow reaction rates which often renders enzyme catalysis impractical as an industrial tool. Work by various groups over the last 25 years has conclusively shown that enzymes can retain both catalytic activity in some organic solvents, thereby enabling the generally much higher solubility and stability of organic substrates and products in these media to be exploited. In general, however, enzyme activity in dry molecular organic solvents is fundamentally hampered by the essential insolubility of proteins in non-denaturing solvents (forcing them to act as heterogeneous rather than homogeneous catalysts) and the drastically reduced structural flexibility permitted by organic liquids as opposed to water. The net effect of these factors generally results in a loss of enzyme activity of several orders of magnitude compared to the native environment. Furthermore, the use of more polar, hydrogen-bonding solvents (such as DMSO) to physically dissolve enzymes generally results in denaturation, by virtue of stronger solvent-protein interactions disrupting the secondary structure. The addition of low levels of water to solvents such as acetonitrile usually elicits similar results – the conformational flexibility engendered by the water acting as a “molecular lubricant” facilitating entropic rearrangement. For these reasons, most practical studies using enzymes in conjunction with organic solvents have resulted in the adoption of “compromise” solutions, such as immobilized or encapsulated proteins or the use of biphasic solvent/water mixtures, in which the activity occurs either in the aqueous phase or interfacially, with the organic solvent functioning merely as a substrate/product reservoir.

Substantial interest nevertheless remains in the potential for performing enzyme catalyses and other biochemical reactions in single-phase, non-aqueous environments. From an academic standpoint, such media would represent an invaluable investigative tool for probing the effects of water displacement or replacement with other species, whilst commercially they could potentially facilitate major improvements in both the performance and practicality of biocatalytic reactors. To this end, various classes of “neoteric” solvents have been evaluated as alternative biochemical reaction media, including supercritical fluids (SCFs), hydrofluorocarbons (HFCs) and room temperature ionic liquids (RTILs). These latter materials offer a number of advantages over both classical organic solvents and other neoterics, which have facilitated their extensive study for a wealth of widely differing applications.

RTILs are salts, composed entirely of ions, with melting points at or below room temperature. Whereas most ionic materials exhibit high melting points, resulting from the high lattice energies arising from strong electrostatic attractive forces between cations and anions, the combination of bulky, charge-delocalized, desymmetrized ions with minimized hydrogen bonding and Van der Waals interactions can raise the entropy of such materials to such an extent as to lower their melting points, in several cases to well below 0°C. For example, the well-characterized heterocyclic RTIL 1-butyl-3-methylimidazolium hexafluorophosphate (BMIm PF6) remains in the liquid state at temperatures as low as –75°C. By virtue of their ionic character, these materials possess a number of unique and useful properties rendering them desirable for solvent applications. Unlike molecular solvents, the enthalpy of vaporization is so low that RTILs possess effectively zero vapour pressure under normal conditions. They are thus non-flammable and non-volatile. In addition, the uniquely ordered liquid structure arising from coulombic ion-ion interactions endows most RTILs with phenomenally high solvation capacities for a wide range of organic and inorganic materials, including natural and artificial polymers, biological macromolecules and even minerals. This environment can also serve to stabilize otherwise short-lived reaction intermediates, thereby facilitating novel chemistries. Importantly, RTILs possess far greater diversity and versatility than do other media – the number of ammonium- and phosphonium-based ionic liquids which can exist, based upon molecular dynamics simulations, has been predicted to be of the order of 1018 individual materials, each possessing its own unique combination of physicochemical properties. The range of functionalities present within such a vast array of materials offers the potential to “tailor” solvents for specific applications; hence ionic liquids are currently the focus of considerable research interest in the field of “medium engineering” – designing the solvent to fit the process. Of the numerous areas in which this capability is being exploited, the emergent sphere of biological and biochemical applications for RTILs is of particular interest, due to the remarkable levels of biomolecular compatibility which have been observed with a number of these solvents.

Prior to the start of this project activity in RTILs had been shown with simple enzyme systems such as hydrolases. Cofactor-dependent oxidoreductases are a more complicated system inasmuch as that the critical redox processes which occur between enzyme and cofactor require a polar, protic environment (normally provided by water).

The objectives of the project were to:
• Deliver tailored ionic liquids as solvents for enzyme catalysed redox reactions, in particular those catalysed by NAD(P)-dependent enzymes
• Develop functionalized ionic liquids to optimise biocompatibility and enhance enzyme and cofactor solubility, stability and activity
• Investigate methods of recycling of nicotinamide cofactors in ionic liquids
• Transfer the technology into industry via the Pro-Bio Partnership.

The project was selected for funding by the EPSRC as a Pro-Bio Flagship Project. The Pro-Bio Faraday Partnership was a consortium of Universities and companies with common interests covering the whole supply chain for biocatalytic chemical manufacturing in the UK. Funding for the Partnership was provided by EPSRC and also DTI. A stated aim of the Partnership was to accelerate the implementation of new biocatalytic manufacturing processes in industry. EPSRC funding (£1.0M) was approved in April 2001 to fund research projects within Pro-Bio and this project was peer reviewed by Pro-Bio and selected as a Flagship Project for support by EPSRC. The Flagship Project system was designed to support projects that can deliver benefits to the whole industrial community aiming to use biocatalysts, through development of novel technology that can be implemented quickly by industry.

Key advances and Supporting Methodology

Tailored ionic liquids as solvents for enzyme catalysed redox reactions, in particular those catalyzed by NAD(P)-dependent enzymes

Tailored dialkylimidazolium ionic liquids
We have demonstrated that the introduction of labile protic groups into an anhydrous ionic liquid facilitates cofactor-dependent enzyme activity. Prior to this discovery, enzyme activity in ionic liquids had primarily been shown with robust lyases and lipases, which have also shown activity in conventional molecular organic solvents. Cofactor-dependent enzymes (such as NADPH-dependent morphine dehydrogenase) are more complicated systems in that the critical redox processes which occur between enzyme and cofactor require a polar, protic environment (normally provided by water). Anhydrous ionic liquids have never previously been shown to permit cofactor-dependent enzyme activity. We demonstrated that ionic liquids can be tailored for redox-dependent biotransformations by functionalizing the cation, the substitutents on the cation heteroatoms and the anion. In the first instance, this was demonstrated by introducing a single hydroxyl group onto the carbon chain, e.g 1-(3-hydroxy-n-propyl)-3-methylimidazolium hexafluorophosphate (3-HOPMIm PF6 ).

Major differences in enzyme activity were apparent between individual ionic liquids. For examples with morphine dehydrogenase activity buffer exceeded that in other solvents; however, product yield was severely depleted through hydrolysis. Negligible conversion occurred in the molecular organic solvents studied; by contrast, the ionic media generally permitted the retention of a limited level of enzyme activity, even at very low levels of water. Whilst the catalytic rates in these media did not approach those observed in water, product solvolysis was effectively suppressed. In certain cases, this permitted higher isolable yield of codeinone than in water. Activity in the hydrophobic ionic liquids correlated closely with water content. The barely significant residual activity observed in BMIm PF6 at < 100 ppm H2O was not replicated in 1-butyl-2,3-dimethylimidazolium PF6, which differed only in the absence of the acidic (and hydrogen bonding) proton at the 2-position. Profound differences were observed between these solvents and the more strongly hydrogen-bonding, hydrophilic RTILs BMIm glycolate and HOPMIm PF6 and HOPMIm glycolate.

The hydroxylated cation in HOPMIm PF6 altered the water dependence of the catalytic activity, permitting its retention at far lower water levels than those possible in the hydrophobic RTILs or in molecular organic solvents. By contrast, the use of hydrogen-bonding anions was significantly less successful, the BMIm glycolate salt being inferior to the HOPMIm PF6 salt. When the hydroxylated cation was employed, replacement of PF6 with a more hydrophilic anion significantly impaired performance; the HOPMIm chloride eliciting near-total abolition of activity.

Walker, A. and Bruce, N. C., Cofactor dependent enzyme catalysis in functionalized ionic solvents. Chem. Commun., 2004, 2570-2571.

We went on to demonstrate that that the opioid analgesic oxycodone could be produced from codeine in a ‘one pot’ reacton in a RTIL, using a combination of chemical and biological catalysis in HOPMIm glycolate. Codeine was converted by MDH to codeinone which exists in dynamic equilibrium with its β,γ-unsaturated isomer neopinone. Direct Markovnikov addition of water across the double bond of this isomer, in contrast to that of codeinone, leads directly to oxycodone. In the absence of any known hydratase capable of performing the hydration of neopinone biocatalytically, a series of chemical reagents known to elicit Makovnikov hydration of substituted alkenes were applied. The most successful of these was bis(acetylacetonato)cobalt (II) with molecular oxygen and phenylsilane.

Walker, A. and Bruce, N. C., Combined biological and chemical catalysis in the preparation of oxycodone. Tetrahedron, 60, 2004, 561-568.

Novel RTILs based on functionalized ammonium nuclei
Dialkylimidazolium RTILs suffer from a number of disadvantages, including issues of cost, viscosity, toxicity, hydrolytic stability (particularly with perfluorinated anions), poor biodegradability and environmental impact and inherent problems in their preparation and purification.

We overcame these shortcomings by designing and developing a range of novel RTILs based upon functionalized ammonium nuclei. These materials address many of the above disadvantages, being orders of magnitude cheaper and easier to synthesize and purify than imidazolium salts (the alkylammonium cations can be easily combined with many anions typically by direct neutralization); they are also effectively transparent from approximately 250-800nm, non-toxic and low environmental impact (we have shown that diluted solutions of the cations can be completely degraded to ammonium, CO2 and water by ordinary soil microorganisms). The cations may be functionalized with a practically limitless range of chemical moieties and combined with any anion by direct neutralization or metathesis, thus permitting the ready preparation of a solvent exhibiting almost any desired combination of physical and chemical properties. For example, viscosity, electrical conductivity, specific heat capacity, proticity and many other properties may be varied across a very wide range by correct selection of the cation/anion pair. As part of this project a library of RTILs was synthesized that included those that were uv/visible-transparent, low-viscosity functionalized alkylammonium-based RTILs, encompassing both N-protic and tetraalkyl-substituted cations.

The chosen cation/anion combinations covered a wide range of degrees of relevant properties, including dissociative proticity, “polarity”, coordinating capability, specific heat capacity, protonic and electronic conductance, hydrogen bond donor/acceptor capability and partition coefficient. Exposed and hindered functional groups were incorporated, in order to compare and contrast bulk and localized effects. Studied functionalities included alkyl, amino, hydroxyl, alkoxyl, oxoalkyl and alkenyl groups.

Ionic liquids based upon hydroxyalkylammonium and aminoalkylammonium cations. UK Priority Patent Filing (0407908.3), priority date of 07/04/04 , converted to PCT Application on 07/04/05. Also filed as separate UK application. Published 12/10/05 as GB2412912.

Our investigations showed that dehydrogenases including Thermoanaerobium brockii alcohol dehydrogenase, yeast alcohol dehydrogenase, morphine dehydrogenase and glucose dehydrogenase exhibit activity levels in these media which compare favourably to those seen in 2. Overall it is clear that the dehydrogenases function well in a range of alkanolammonium RTILs. Levels of water in all of these ionic liquids was low, typically below 1%, but it was not always possible to reduce water content below 100 ppm as this class of ionic liquids is hydrophilic and hygroscopic they tend to absorb moisture from the air. As these alkylammonium ionic liquids are water miscible, we investigated the effect of water content on enzyme catalysis. In RTILs with less nucleophilic anions, such as glycolate, or methansulfonate increasing the ionic liquid concentration up to 40 % results in a loss of activity, as the enzyme is present in an increasingly strong salt solution. But then, as the ionic liquid concentration is increased further, the activity of the enzyme increases again, as the solvent takes on the properties of an ionic liquid instead of a salt solution. To investigate these observations further samples of the enzyme in different concentrations of RTIL were analyzed by circular dichroism. Interestingly, in 10 % RTIL, the CD signal from the secondary structure of the protein is similar to the native one, when the activity levels are close to those seen in buffer. When more water is present, essentially creating a strong salt solution, the CD signal is significantly reduced and the enzyme activity is near zero. Then as the concentration of ionic liquid begins to exceed that of water, and reasonable activity is observed, the CD shows that the protein has refolded to an active confirmation.

We also investigated the activity of proteases such as subtilisin and chymotrypsin in a range of alkyllammonium ionic liquids. We found that proteases are generally inactive in the RTILs tested, which came as a surprise since these enzyme are generally very robust and are known to be active in organic solvents. CD analysis revealed that the proteases were unfolded in the RTILs. This work has led to further funding from the BBSRC to characterize the stability, structure and conformation and catalytic activity of a range of enzymes in functionalized ionic liquids. As part of this new project we are also investigating the “solvation” environment (i.e. protein/solvent interactions) of enzymes in RTILs and examining the effect of water content upon the above factors.

Develop functionalized ionic liquids to optimize biocompatibility and enhance enzyme and cofactor solubility, stability and activity
A long term study of stability of yeast ADH and MDH was performed in several ionic liquids and buffer. Activity of the enzymes stored in buffer drops significantly over the first few days as expected, whilst the activity of the enzymes stored in three different alkylammonium ionic liquids has remained constant over several months, even though they have absorbed up to 2 % water during this time. An equivalent study was set up for NAD(P). The amount of functional cofactor was assessed at different time points using an enzyme assay. Again it is clear that under ambient conditions the cofactor is more stable in certain of the ionic liquids than it is in buffer.

Investigate methods of recycling of nicotinamide cofactors in ionic liquids
Many of the NAD(P)-dependent enzymes catalyze reactions of significant commercial interest, since the cofactors need to be added in stoichiometric quantities cofactor regeneration is, therefore, an important consideration when processes involving NAD(P)-dependent oxidoreductases are to be applied in a commercial setting. In cell-free systems cofactor must be supplied, albeit at lower-than-stoichiometric concentrations (catalytic amounts), when cofactor generation is achieved. Alcohol dehydrogenases and glucose dehydrogenase were routinely used to recycle NAD(P) in homogenous biocatalytic systems in RTIL. Under cofactor-recycling conditions, where an NADPH-dependent alcohol dehydrogenase in an RTIL was used, the MDH reaction was shown to be significantly enhanced.

Other Applications for ionic liquids with alkylammonium cations
In addition to their use as solvents for non-aqueous homogeneous enzyme catalysis the novel solvents developed during this project have also found utility as solvents for biological assays (including immunoassays), affinity chromatography and natural product extraction. Three patents have been filed covering these applications.

Biological assays in ionic liquids. UK Priority Patent Filing (0408853.0), priority date of 21/04/04, converted to PCT Application on 20/04/05.

Affinity chromatography in ionic liquids. UK Priority Patent Filing (0408854.8), priority date of 21/04/04, converted to PCT Application on 21/04/05.

Extraction of alkaloids using ionic liquids. UK Priority Patent Filing (0606147.7), priority date of 28/03/06. Filed by Bioniqs Ltd.

Research Impact and Benefits to Society
The synthesis of a new class of environmentally benign ionic liquids offers the possibility of designing solvents for a wide range of application. To date these solvents have found potential use for biocatalysis, affinity assays, natural product extraction, cleaning in place and polymer dissolution. Academic interest in this project has resulted in collaborations with Prof Peter Halling (University of Strathclyde), Dr Seishi Shimizu (University of York), Prof James Clark (University of York), Prof David Goodall (University of York), Dr Victor Chechik (University of York), Dr Jen Potts (University of York), Dr Chris Anderson (Massey) and Prof Roy Daniel (Waikato). The work has also resulted in invitations to speak at national and international conferences including the 2003 ProBio Annual Conference (Edinburgh), 2004 ProBio Annual Conference (Warwick), Biocatalysis in Non-Convential Media (Manchester), 7th International Conference on Protein Stabilisation (Exeter) and Biotrans 2007 (Oviedo).

The work has resulted in two publications with another two papers planned for submission this summer. Three patent applications have been filed by the University of York and a spin out company Bioniqs Ltd has been formed to commercialize the Technology. Bioniqs was formed in November 2004 and was created through the vehicle Amaethon, a joint venture between the University of York and CNAP, which was capitalized by IP2IPO (now IPGroup).

Further Research or Dissemination Activities
The research has led to additional research funding from the BBSRC (BBD522989/1 - Protein interactions in ionic media) and through the BBSRC/MoD Joint grant Scheme (BB/E000576/1 - Advanced optical waveguide biosensors for the detection of illicit drugs and explosives). The work also formed a significant component of a successful FP 6 Marie Curie EST Programme ‘CHEMCELL’.