WO2018004341A1

WO2018004341A1 - Cross-linked enzyme aggregate comprising magnetizable particles

Info

Publication number: WO2018004341A1
Application number: PCT/NL2017/050431
Authority: WO
Inventors: Sander VAN PELT; Michiel Hubertus Arnold Janssen; Jo-Anne Michelle RASMUSSEN; Menno Jort Sorgedrager; Roger Arthur Sheldon; Pieter KONING
Original assignee: Clea Technologies B.V.
Priority date: 2016-06-29
Filing date: 2017-06-29
Publication date: 2018-01-04
Also published as: NL2017064B1

Abstract

The invention relates to a CLEA comprising magnetizable particles, characterized in that the magnetizable particles comprise particles of a zerovalent metal selected from the group of iron, nickel, cobalt and any mixture thereof, which particles are fully covered with a coating. The CLEAs of the invention were applied in catalytic processes, wherein it was demonstrated that the CLEAs can be recycled without any substantial loss of activity and with a virtually complete catalyst recovery. In addition, no leaching of metal and enzyme from the CLEAs was observed during their application.

Description

Cross-linked enzyme aggregate comprising magnetizable particles

The invention relates to a cross-linked enzyme aggregate (CLEA), to a method for making such CLEA and to a process that makes use of the CLEA.

In enzymatic processes that are performed at an industrial scale, it is important that the enzyme has a long-term operational stability and that recovery and re-use of the enzyme are efficient and convenient. These requirements can be met by using enzymes in an immobilized form. Efficient recovery of an immobilized enzyme from a reaction medium can be achieved by means of standard separation techniques, such as filtration, centrifugation or settling.

There are processes, however, wherein other solids are present in the reaction medium besides the immobilized enzyme. In such cases, separation only by filtration, centrifugation or settling is not sufficient, since this does not separate the immobilized enzyme from the other solid(s). It is known that this problem can be overcome by providing the immobilized enzyme with magnetizable components. In this way, the separation from other solid components in the mixture is possible by making use of a magnet.

Generally, methods of enzyme immobilisation can be divided into three categories (R.A. Sheldon, S. van Pelt, Chem. Soc. Rev., 2013, 42, 6223 - 6235); 1 ) binding the enzyme to a support (carrier); 2) entrapment of the enzyme (encapsulation); and 3) cross-linking of enzyme aggregates (CLEAs) or enzyme crystals (CLECs). The advantage of the third method - as compared to the first and the second method - is that it produces carrier-free macroparticles wherein non-catalytic material can be virtually absent. Therefore, CLEAs have a particularly high activity per weight unit as compared to carrier-based and encapsulated enzymes. When a CLEA is to be provided with a magnetizable component, however, the absence of any carrier or other non-catalytic auxiliary material makes it more difficult to introduce and immobilize magnetizable components in the immobilized enzyme matrix. A known method to prepare magnetizable CLEAs is to incorporate (nano)particles of magnetite in the CLEA. The magnetite is usually

functionalized {e.g. with (3-aminopropyl)triethoxysilane or

(3-aminopropyl)trimethoxysilane) to be able to co-crosslink the magnetic particles with the enzyme aggregates, so that the magnetite particles are fixated in the final CLEA. Although reports on the catalytic action of CLEAs thus obtained do not seem to mention any undesired leaching of iron, caused by the dissolution of magnetite, it has been found by the present inventors that iron leaching indeed takes place under certain conditions, such as a lower pH of reaction (pH < 7) and can be accelerated in presence of free carboxylic acids. The dissolution of magnetite has undesirable effects. Such effects concern the presence of iron ions in the reaction medium with undesirable properties, a decrease in the catalytic activity of the CLEA and an incomplete recovery of the CLEA using magnetic recovery techniques. It appears that the undesirable effects of iron leaching become apparent only after long residence times of the CLEA in the reaction medium and after a plurality of CLEA recovery cycles. This is indeed different from the conditions wherein the magnetizable CLEAs known in the art are applied (i.e. neutral or basic pH, organic solvent, or short residence times and only a few or no recovery cycles). Since industrial processes often require a long catalyst lifetime and comprise multiple catalyst recoveries, the use of conventional magnetizable CLEAs poses serious problems for their application on an industrial scale.

Another clear disadvantage of the conventional magnetizable CLEA is the relatively low saturation magnetizability of magnetite, which makes it challenging to separate the conventional magnetizable CLEA from large scale applications with high flow rates and/or challenging media using cheap permanent magnet recovery set-ups.

It is therefore an object of the present invention to provide improved magnetizable CLEAs, in particular with respect to their tendency to leach iron (or any other metal present in the magnetizable particles), their

magnetizability, the decay of their activity in catalytic processes, and/or their recovery in catalytic processes. Therefore, the invention relates to a CLEA comprising magnetizable particles, characterized in that the magnetizable particles comprise particles of a zerovalent metal selected from the group of iron, nickel, cobalt and any mixture thereof.

A CLEA of the invention is in principle present as a collection of

CLEA particles. The dimensions of CLEA particles of the invention are usually within the range of those of CLEA particles described in the art. Typically, the average diameter of the CLEA particles in a CLEA of the invention is 0.5 μιη or more. Preferably, it is 1 μιη or more. It may also be 2 μιη or more, 3 μιη or more, 5 μιη or more, 7 μιη or more, 10 μιη or more, 15 μιη or more, 20 μιη or more, 30 μιη or more, 40 μιη or more or 50 μιη or more. The average diameter of the CLEA particles is usually 100 μιη or less, 80 μιη or less, 60 μιη or less, 40 μιη or less, 30 μιη or less, 20 μιη or less, 15 μιη or less or 10 μιη or less. By "average diameter" is meant that the diameter is the average of the diameters of all CLEA-particles in a particular CLEA. Usually, the average diameter of the CLEA-particles is in the range of 1 -100 μιη, preferably it is in the range of 1 -50 μιη. The diameter of the CLEA-particles are measured with laser diffraction, using e.g. a Malvern MasterSizer 3000 Hydro MV.

The size distribution of CLEA particles of the invention usually has a

D10 value in the range of 1-9 μιη, a D50 value in the range of 10-30 μιη, and a D90 value in the range of 35-75 μιη (measured with laser diffraction using a Malvern MasterSizer 3000 Hydro MV). The values of D10, D50 and D90 indicate that 10 percent of the particle population has a particle diameter below the D10 value; half of the population has a particle diameter below the D50 value; and 90 percent of the population has a particle diameter below the D90 value.

In some cases, CLEA particles of the invention tend to form clusters, likely due to protein-protein interactions. Such clusters may have a size of up to 500 μιη. The breaking up of such clusters in smaller clusters or in separate CLEA particles is possible, yet sometimes difficult. Such clustering appears to have no negative influence on the catalytic properties of the CLEAs of the invention, such as their leaching of metal ions, their leaching of magnetizable particles, their catalytic activity or their recovery efficiency in catalytic processes.

CLEA particles of the invention may in principle have any shape. Usually, they are compact, i.e. not of a flat or elongated form. They may for example be essentially spherical. The largest dimension (largest diameter) of the CLEA particles may also be up to 1 .1 times, up to 1 .3 times, up to 1 .5 times, up to 2 times or up to 3 times larger than its smallest dimension. A CLEA particle of the invention may also deviate more from a compact shape in that its largest dimension is up to 5, up to 7, or up to 10 times larger than its smallest dimension.

Figure 1 displays a Scanning Electron Microscope (SEM) picture of CIP-silica particles measured at 5 kV and 15000 x magnification on a

TableTop SEM Hitachi TM3030Plus by Sysmex.

Figure 2 displays a Microscope image of a CLEA of the invention, comprising particles of silica-coated CIP and cross-linked glucose amylase. The cross-linked protein content of the CIP-silica mCLEA is 34 wt%.

Figure 3 displays a Microscope image of a CLEA of the invention, comprising particles of silica-coated CIP and cross-linked glutaminase. The cross-linked protein content of the CIP-silica mCLEA is 12 wt%.

The entities that are visible on the photographs of Figures 2 and 3 are the CLEAs. The black areas of the CLEAs are the magnetizable particles or clusters thereof, while the lighter areas of the CLEAs between the magnetizable particles are the cross-linked proteins that hold together the CIP particles in a CLEA of the invention. The CLEA particles in Figure 3 have a lower cross-linked protein content than those in Figure 2.

The cross-linked protein content of a CLEA of the invention is defined as the weight percentage of cross-linked protein in the CLEA. The lower loading in Figure 3 is confirmed by the presence of less cross-linked protein {i.e. light areas) between the magnetizable particles, than in Figure 2.

Some magnetizable CLEA particles only comprise a thin layer of cross-linked protein. Figures 2 and 3 demonstrate that magnetizable particles are indeed present in a CLEA of the invention and that such CLEAs can have various shapes. Further evidence for the incorporation of magnetizable particles was obtained by staining the protein of the CLEA with Coomassie Blue. In the recorded photographs (not shown), the black magnetizable particles were embedded in a blue environment of stained protein.

The enzyme that is immobilized in a CLEA of the invention is usually selected from the group of hydrolases, such as esterases, proteases, amidases, cellulases, nitrilases, xylanases and glycosylases; lyases, such as hydroxynitrile lyases and aldolases; oxidoreductases, such as alcohol oxidases, peroxidases, ketoreductases and imine reductases; and

transferases, such as transaminases.

A CLEA of the invention may also contain two or more different enzymes in the aggregate {i.e. in the same CLEA particle). Such a CLEA is also known as a "combi-CLEA". Using combi-CLEAs allows multiple process steps in a single operation (a one-pot process), which minimizes solvent usage, waste generation and energy.

A CLEA of the invention comprises magnetizable particles of a zerovalent metal selected from the group of iron, nickel, cobalt and mixtures thereof. Preferably, the amount of the zerovalent metal(s) in the magnetizable particles is as high as possible, for example 90 wt% or more, 95 wt% or more, 96 wt% or more, 97 wt% or more, 98 wt% or more, 98.5 wt% or more, 99 wt% or more, 99.5 wt% or more or 99.8 wt% or more. More preferably, the magnetizable particles essentially consist of the zerovalent metal or a mixture of the zerovalent metals. In the event that the magnetizable particles comprise other components, these components are typically non-zerovalent metals, in particular selected from the group of iron, nickel and cobalt; and more in particular oxides and/or mixed oxides thereof. The magnetizable particles may also contain one or more of the elements nitrogen, carbon and oxygen.

In a CLEA of the invention, the magnetizable particles in particular comprise zerovalent iron. More in particular, the particles comprise so-called "carbonyl iron particles" (CIP) or are derived thereof. CIP is an iron powder that is prepared by thermal decomposition of iron pentacarbonyl. This well- known method produces spherical iron particles of high purity, typically with particle sizes of a few micrometers.

Alternatively, when the magnetizable particles are made of nickel or cobalt, these metals may also be used in a form analogous to that of CIP, i.e. nickel particles prepared from "carbonyl nickel" {e.g. nickel tetracarbonyl) and cobalt particles prepared from "carbonyl cobalt" {e.g. dicobalt octacarbonyl).

In a CLEA of the invention, it is preferred that the particles of the zerovalent metal are covered with a coating. More preferably, the metal particles are completely covered by such a coating. The main function of such a coating is to maintain the physical properties of the particles and to protect the particles against corrosion by the aqueous and/or oxidative environments wherein a CLEA of the invention usually resides during its synthesis and/or during the catalytic processes wherein the CLEA is used.

Thus, the coating is preferably of a compound that is capable of protecting the zerovalent metal from degradation by water, oxygen and/or acid. The coating in particular comprises silica {i.e. SiO₂), more in particular non-mesoporous silica, or a material comprising a silicate anion, such as orthosilicate {i.e. SiO₄ ^4"). The coating may also comprise a compound selected from the group of carbon, metal oxides such as AI₂O₃, and polymers such as polystyrene. Preferably, the coating consists of any of the materials mentioned above.

The coating usually has a thickness of at least 1 nm. It may in particular be 5 nm or more, 10 nm or more, 20 nm or more, 30 nm or more, 40 nm or more, 50 nm or more, 60 nm or more, 70 nm or more, 80 nm or more, 90 nm or more, 100 nm or more, 125 nm or more, 150 nm or more, 200 nm or more or 300 nm or more.

The coating usually has a thickness of 400 nm or less. It may in particular be 300 nm or less, 200 nm or less, 150 nm or less, 100 nm or less, 75 nm or less or 50 nm or less.

The thickness may be in the range of 10-500 nm, in the range of 25-

150 nm, in the range of 50-100 nm, or in the range of 100-150 nm.

Optionally, the coating is functionalized. By such functionalization of the coating is meant that particular functional goups are provided {e.g. grafted) at the surface of the magnetizable particles via a spacer of a particular length, which can realize an interaction (in particular a covalent interaction) between the magnetizable particles and the enzyme.

A covalent interaction can be realized by allowing a functional group to participate in the cross-linking reaction of the CLEA synthesis. For example, the surface may be functionalized with alkoxysilanes having a pendant amine functionality that can co-cross-link with the enzyme

aggregates. The surface is for example functionalized with

(3-aminopropyl)triethoxysilane or (3-aminopropyl)ethyldiethoxysilane. In this way, the binding strength of the magnetizable particles with the cross-linked protein can be increased, which diminishes the release {i.e. leaching) of the magnetizable particles from the CLEA according to the invention.

A covalent interaction between the magnetizable particles and the cross-linked protein can also be realized by epoxide or carboxylic acid groups.

Other means of functionalization result in a hydrophobic interaction between the cross-linked protein and the magnetizable particles (by using e.g. hydrophic groups such as phenyl, octyl or octadecyl groups) or in an ionic interaction between the cross-linked protein and the magnetizable particles (by using e.g. charged groups such as tertiary and quaternary amines).

In the invention, however, functionalities on the coating are in principle not necessary. It appeared for example that smooth, non-porous, silica-coated magnetizable particles can effectively be included in a CLEA to yield a CLEA of the invention and that leaching of the magnetizable particles from a CLEA of the invention in essence does not occur during its synthesis and neither when the CLEA of the invention is applied in a catalytic process. Thus, in a CLEA of the invention, a functionalization of the magnetizable particles with bifunctional molecular entities that provide a covalent attachment of the magnetizable particles to the cross-linked protein, in particular those entities having an alkyl chain, such as

aminoalkylalkoxysilanes, may be absent. This absence does however not exclude any other covalent interaction between the cross-linked protein and the particles, since it may indeed be possible that the surface of the magnetizable particles as such is capable of forming a covalent interaction with a particular group of the cross-linked protein.

As described above, the magnetizable particles in a CLEA of the invention comprise a zerovalent metal particle that is preferably surrounded by a coating. In case of the coating, the magnetizable particles can be regarded as particles comprising a core of the zerovalent metal and a surface layer of another material.

For the purpose of the present invention, the description of the size of the magnetizable particles is based on the outer dimensions of the

magnetizable particles. When the magnetizable particles comprise a coating, their outer dimensions are defined by the size of the metal particle and the thickness of the coating. When the magnetizable particles do not comprise a coating, their outer dimensions are defined by the size of only the metal particle.

Typically, the average diameter of the magnetizable particles in a

CLEA of the invention is in the range of 10 nm - 20 μιη, in particular it is in the range of 1 -15 μιη. For example, it is 20 nm or more, 30 nm or more, 40 nm or more, 50 nm or more, 75 nm or more, 100 nm or more, 200 nm or more, 300 nm or more, 400 nm or more, 500 nm or more, 750 nm or more, 1 μιη or more, 2 μιη or more, 3 μιη or more, 4 μιη or more, 5 μιη or more, 6 μιη or more, 7 μιη or more, 8 μιη or more, 9 μιη or more, 10 μιη or more, 12 μιη or more, 14 μιη or more, or 20 μιη or more. The average diameter of the magnetizable particles is 20 μιη or less, 16 μιη or less, 13 μιη or less, 10 μιη or less, 8 μιη or less, 7 μιη or less, 6 μιη or less, 5 μιη or less, 4 μιη or less, 3 μιη or less, 2 μιη or less, 1 μιη or less, 800 nm or less, 600 nm or less, 400 nm or less, 200 nm or less, 100 nm or less, 80 nm or less, 60 nm or less, 40 nm or less, 30 nm or less, or 20 nm or less. By "average diameter" is meant that the diameter is the average of the diameters of all magnetizable particles in a particular CLEA of the invention.

The magnetizable particles may in principle have any shape. They may have a flat shape {e.g. flake-like), an elongated shape {e.g. rod-like) or a more compact shape {e.g. spherical, ovoid or cubic-like). When the magnetizable particles have a compact shape, the largest dimension (largest diameter) of the particles is typically up to 2 times, up to 3 times, or up to 5 times larger than their smallest dimension. When a magnetizable particle is nearly spherical, the largest dimension (largest diameter) of the magnetizable particle is typically up to 1 .05 times, 1 .1 times, 1 .2 times, 1 .3 times or 1 .5 times larger than its smallest dimension.

In an embodiment, the magnetizable particles have an essentially spherical shape. In such case, the ranges of the average diameter as provided hereinabove are the ranges of the actual diameter of the sphere. When the magnetizable particles are spherical or nearly spherical, they preferably comprise "carbonyl iron particles" (CIP), as described hereinabove.

The magnetizable particles, and in particular magnetizable particles that have a spherical shape, may be present in a particle size distribution having a median value D50 in the range of 25 nm - 15 μιη, the D50 value being the particle size that splits the distribution with half above and half below this size. The relative span of the distribution of magnetizable particles is usually in the range of 1 -10, wherein the relative span is defined by dividing the absolute span D10-D90 by the D50 value, D10 and D90 being the particle size that splits the distribution with 10% of the particles below the size of D10 and 90% of the particles below the size of D90, respectively {i.e. the relative span is (D90-D10)/D50). The relative span may also be 10 or less, 7 or less, 5 or less, 3 or less, 2 or less, 1 or less, 0.7 or less, 0.5 or less, 0.3 or less, 0.2 or less, or 0.1 or less.

In a particular embodiment, the particle size distribution has a D10 value in the range of 1 -3 μιη, in particular 1 .7-2.8 μιη, a D50 value in the range of 3.5-6 μιη, in particular in the range of 3.9-5.3 μιη and a D90 value in the range of 6.5-10 μιη, in particular 7.2-9.3 μιη. The thickness of the coating of such magnetizable particles is usually in the range of 10-500 nm, in the range of 25-150 nm, in the range of 50-100 nm, or in the range of 100-150 nm.

The magnetizable particles may also be present in a CLEA of the invention in an aggregated form. For example, when compact particles, in particular particles comprising CIP, stick together, they may be incorporated in the CLEA of the invention as a small cluster of particles, e.g. of 2 or more, 5 or more, 10 or more, 20 or more, 50 or more or 100 or more particles. It is also possible that during a coating process of the magnetizable particles, the metal particles are not completely separated, so that the resulting coating embraces a small cluster wherein a plurality of metal particles are present, e.g. 2 or more, 5 or more, 10 or more, 20 or more, 50 or more or 100 or more. An aggregate of magnetizable particles as described above {i.e. wherein each magnetizable particle is separately coated or wherein an aggregate of metal particles is coated as a whole) may as such be incorporated into a CLEA of the invention. The skilled person is capable of finding the appropriate conditions to reach any of these conditions wherein the magnetizable particles reside {e.g. as completely separated particles or as one of the aggregates as described hereinabove), without undue burden and without exerting an inventive effort.

It is also possible that the magnetizable particles are obtained by the sintering of smaller metal particles, followed by an eventual coating the sintered product under appropriate conditions. In this way, particles with an irregular shape can be incorporated into a CLEA of the invention. For example, the magnetizable particles in a CLEA of the invention may comprise sintered particles that comprise CIP. The number of sintered particles may be 2 or more, 5 or more, 10 or more, 20 or more, 50 or more or 100 or more.

It was surprisingly found that the zerovalent magnetizable particles could effectively and permanently be incorporated into a CLEA of the invention by cross-linking a precipitated enzyme in the presence of the magnetizable particles, even when the particles were not functionalized.

Accordingly, the invention further relates to a process for preparing a CLEA, comprising

providing a solution of an enzyme; then

precipitating the enzyme by forming an enzyme aggregate; then - cross-linking the enzyme aggregate with a cross-linking agent in the presence of magnetizable particles comprising particles of a zerovalent metal selected from the group of iron, nickel, cobalt and any mixture thereof. The magnetizable particles are present during the cross-linking step, and thus are added to the enzyme before this step is carried out. The magnetizable particles may be added before, during or after the step of precipitating the enzyme.

In a method of the invention, the magnetizable particles are preferably completely covered by a coating so as to ensure that the

zerovalent metal in the particles does not degrade under the conditions that are present during the CLEA synthesis or during a catalytic process wherein the CLEA is used. Usually, such conditions are aqueous and/or aerobic conditions, under which the zerovalent metal in the particles is prone to oxidation followed by dissolution.

In a method of the invention, it is preferred to use CIP-based magnetizable particles, i.e. iron particles derived from CIP. It is also preferred that these particles comprise a coating comprising silica {i.e.

non-(meso)porous silica). The coating in particular consists of silica {i.e.

non-(meso)porous silica). It appears not necessary to use a functionalization on the coating, thus the coating may be unfunctionalized.

The CIP-based magnetizable particles in particular have a particle size distribution wherein D10 is in the range of 1 -3 μιη, in particular 1 .7-2.8 μιη, D50 is in the range of 3.5-6 μιη, in particular in the range of 3.9-5.3 μιη and D90 is in the range of 6.5-10 μιη, in particular 7.2-9.3 μιη. The thickness of the coating on the CIP is usually in the range of 10-500 nm, in the range of 25-150 nm, in the range of 50-100 nm, or in the range of 100-150 nm.

It is surprising that a CLEA of the invention comprising

non-functionalized magnetizable particles, did not measurably leach magnetizable particles into the medium of this CLEA, while this CLEA contained a high loading of protein on the magnetizable particles. This is surprising because the silica coating on the particles has a low surface area as compared to mesoporous silica materials that are generally used in the art as a support for catalyst immobilization, so the area for interaction of the cross-linked protein with the magnetizable particles is more limited than in the art. In addition, no functionalization was present, which - if present - would compensate for that. Further, given that the size of the magnetizable particles is in general above 1 μιη, the surface area per gram of particle material is even more remote from that of commonly used silica supports. For these reasons, it was not expected that the use of silica-coated CIP would yield a CLEA with magnetizable particles that has improved properties, in particular a virtually absent leaching of CIP in combination with a high protein loading. A further advantage is that functionalization of the magnetizable particles is not necessary in a CLEA of the invention. In contrast, the magnetite used in the art often requires functionalization for a successful and long-lasting

incorporation into a CLEA.

Iron leakage experiments were carried out on CLEAs of the invention, by incubating them in an acid environment and performing an iron detection assay on the incubation supernatant. Although the conditions chosen for incubation were extreme with respect to common enzyme applications, iron leakage from the CIP-silica mCLEA was found to be negligible.

It is further surprising that CLEAs of the invention were not substantially affected by a magnetic field used for their recovery. In particular, no leaching of the magnetizable particles could be detected. Due to the relatively low surface area of the magnetizable particles and the lack of functionalization thereon, it was considered a likely possibility that the magnetizable particles would be pulled out of a CLEA of the invention by an external magnetic field, leaving the CLEA of the invention in a state in which it cannot be magnetized anymore. This not at all being the case, it can be concluded that the magnetizable particles in a CLEA of the invention do not only withstand the forces resulting from vibration and other thermally induced movements, but that they also withstand the magnetic force of an external magnetic field.

It further appears that the CIP-based magnetizable particles are not only prevented from leaching into the medium, but they also maintain a fixed position within a CLEA of the invention. By a fixed position is meant that the particles do not undergo substantial translational movement within a CLEA of the invention. Further investigations revealed that there is essentially no interaction between the free dissolved enzyme and the CIP-based particles. In view of these findings, it is even more surprising that the particles are so well incorporated in a CLEA of the invention.

The leakage of enzyme from CLEAs of the invention into the reaction medium was also investigated. Surprisingly, no leaching of enzyme could be demonstrated, even after stirring the CLEA for proloned times, such as 10 days, in an aqueous medium at 32 °C.

Recycling experiments were also carried out with CLEAs of the invention, to mimic their application on an industrial scale. It was found that at least 10 cycles could be made without any substantial loss of activity and with a virtually complete catalyst recovery.

The measured saturation magnetizations of CLEAs of the invention were surprisingly high. Values of up to 200 emu-g^"1 were found, while known CLEAs containing APTES functionalized magnetite generally have a saturation magnetization in the range of 10^10 emu-g^"1.

Generally, the saturation magnetization of a CLEA of the invention is lower when it contains more protein, since protein is a non-magnetizable material. Thus, a large amount of protein would result in a poorer separability and recovery of the CLEA in a catalytic process. On the other hand, it is an advantage that the amount of protein in a CLEA of the invention is high, because this increases the catalytic activity per gram of CLEA. It has been found that the use of the zerovalent metal {i.e. iron, nickel, cobalt) in the magnetizable particles provides a good balance between the saturation magnetization and the catalytic activity per gram of CLEA.

The number of magnetizable particles that is incorporated in a CLEA-particle of the invention is 1 or more. The number of magnetizable particles that is on average incorporated in a CLEA-particle of the invention (the "average number") may be in the range of 1-10,000, in particular it is in the range of 1 -100. The average number may be 2 or more, 3 or more, 5 or more, 7 or more, 10 or more, 15 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more or 80 or more. The content of magnetizable particles in a CLEA of the invention may also be defined as a weight percentage of the entire CLEA. The content is usually in the range of 1 -99 wt%. It may also be in the range of 25-90 wt%, in the range of 50-95 wt% or in the range of 70-90 wt%. It may be 3 wt% or more, 5 wt% or more, 10 wt% or more, 15 wt% or more, 20 wt% or more, 30 wt% or more, 40 wt% or more, 50 wt% or more, 60 wt% or more, 65 wt% or more, 70 wt% or more, 75 wt% or more, 80 wt% or more, or 85 wt% or more.

The cross-linking agent may be selected from the group of formaldehyde, glyoxal, glutaraldehyde and aldehyde-comprising cross-linking agents derived from polysaccharides.

The invention further relates to a CLEA obtainable by the preparation method described hereinabove.

The invention further relates to a process comprising the use of a CLEA as described hereinabove for a catalytic conversion, wherein the CLEA is separated from the reaction medium by collecting the CLEA with the aid of a magnetic field, in particular with the aid of a permanent magnet.

EXAMPLES

General

Magnetic strength

Magnetic strength measurements were carried out on a Model 2000 Alternating Gradient Magnetometer of the company Princeton Measurement Cooperation. The reported values are an average of three measurements.

SEM

Scanning Electron Microscope (SEM) measurements were

performed on a TableTop SEM Hitachi TM3030Plus by Sysmex.

Iron detection assay

The organic ligand 2,2'-bipyridine forms stable, strongly colored complexes with Fe(ll). Very small quantities of iron in solution can be detected by using the intense light absorption properties of these complexes. Hydroxylammonium chloride was used as a reducing agent to convert Fe(lll) into Fe(ll) as Fe(lll) and 2,2'-bipyridine form a complex that absorbs less strongly and at a different wavelength. Using this method, Fe(ll)

concentrations of 1 to 5 ppm (1 - 5 mg- L^"1) could easily be detected (Vogel, A Textbook of Quantitative Inorganic Analysis, 3rd Ed., p. 294, 310 and 787).

Example 1. Synthesis of silica coated CIP (CIP-silica)

Preparation of CIP

CIP particles were prepared according to known methods of iron carbonyl decomposition, such as described in e.g. GB684054A. Besides, the CIP particles were also directly obtained from BASF.

Providing the CIP with a silica coating

The silica coating layer is introduced in order to maintain the physical and physicochemical properties of the CIP, such as to protect the CIP against oxidation and subsequent dissolution, or as to serve as a matrix that allows for chemical modification for the introduction of functional groups. The majority of these CIP-silica particles have been produced relying on a method known as the Stober method, which was originally reported by W. Stober, A. Fink, and E. Bonn, Journal of Colloid Interface Science, 1968, 26, 62-68, or a sol-gel method.

The reaction was performed under a nitrogen atmosphere. A 500 imL reaction vessel was charged with 30 g of CIP (OM grade, BASF), 210 imL of ethanol, 2.9 g of TEOS (tetraethyl orthosilicate), 33 imL of water and 6 imL of 25% ammonia. The mixture was vigorously stirred with an overhead stirrer for 30 minutes after which it was heated to 60 °C. Next, 15 g of TEOS was added over a period of 2 hours after which stirring was continued for another 4 hours. The mixture was then cooled to room temperature and the particles were repeatedly collected with a handheld magnet and resuspended in ethanol for washing. Finally, the CIP-silica particles were dried under reduced pressure at 80 °C for 3 days. Figure 1 shows a Scanning Electron

Microscope (SEM) picture of the CIP-silica particles measured at 5 kV and 15000 x magnification on a TableTop SEM Hitachi TM3030Plus by Sysmex. Iron leakage

Iron leakage from CIP-silica particles was determined by using the iron detection assay as described above.

The absorption of the supernatant in the iron detection assay at 100 wt% iron leakage was determined by completely dissolving 6 mg of CIP-silica particles in concentrated HCI and subsequent dilution.

The iron leakage from CIP-silica particles was then studied by contacting 6 mg of the CIP-silica particles with 18 ml of a 1 wt% aqueous lactic acid solution of pH 3. The iron leakage (wt%) after incubation was determined by directly comparing the absorption of the incubation

supernatant of the CIP-silica particles to that of the absorption after total dissolution (100 wt% iron leakage). After gently shaking at 32 °C for 72 h, leakage of Fe(ll,lll) was <1 wt%. Commercial CIP without any protective coating, on the other hand, appeared to dissolve completely in these incubation mixtures within 24 hours {i.e. 100 wt% leakage).

Magnetic strength

The saturation magnetization of dry CIP-silica particles was measured at 235 emu-g^"1 (Am²/kg).

Example 2. CIP-silica magnetizable CLEA (mCLEA) of glucose amylase - high protein loading

Synthesis

In a 2 L plastic beaker, a mixture of 0.727 L of saturated ammonium sulfate and 0.225 kg of CIP-silica particles (prepared according to Example 1 ) was stirred with an overhead stirrer (Velp Scientifica ES overhead stirrer) at room temperature for 1 h, to form a suspension. Thereafter, 0.3 L of glucose amylase (Zibo Guoao, Shandong, China) was added slowly to the

suspension, followed by stirring the suspension at room temperature for 1 hour. After the addition of 0.153 L of 25 wt% glutaraldehyde, the reaction mixture was stirred at room temperature for 18 h. The resulting CIP-silica mCLEA was removed with a handheld magnet (ERIEZ Mega Rare Earth Tube Magnet 150 mm, 10700 Gauss) and washed five times with 4.5 L of water. The final CIP-silica mCLEA was suspended in 1 L of water. The total dry weight of the CIP-silica mCLEA was 341 grams. Figure 2 shows a

Microscope image of the CIP-silica mCLEA of glucose amylase, obtained with a Bresser microscope using a 40x magnification lens and Mikro CamLab software (Version 6.1 .4.0).

Activity assay

In a 100 imL conical flask were added 30 imL of maltodextrin

(33% w/v in water) and 0.2 imL of glucose amylase CIP-silica mCLEA suspension. The suspension was shaken at 150 rpm at 32 °C for 3 hours in a Stuart Orbital Incubator SI500. After 1 , 2 and 3 h, a 0.1 mL sample was taken from the suspension and added to 0.9 mL of water. The mixture was then centrifuged, filtered (0.2 μιη Phenex™ syringe filter), transferred to an HPLC vial and subjected to HPLC analysis under the following conditions to determine the glucose content:

HPLC: Shimadzu LC-20AT Prominence Liquid

Chromatograph using a SIL-20AC Prominence auto sampler

Column: Bio-Rad Aminex^® HPX-87H 300x7.8 mm

Mobile phase: 5 mM H₂SO₄

Flow rate: 0.6 imL-min^"1

Temperature: 60 °C

Detector: Shimadzu RID 10 A

Sample volume: 10 μΐ

Runtime: 1 1 min.

Retention time: 9.2 min. (glucose)

The HPLC analysis demonstrated 18% activity recovery versus the free enzyme in this assay. Cross-linked protein content of the CIP-silica mCLEA was 34 wt%. Glucose amylase activity was neither detected in the supernatant of the CLEA preparation nor in the washing water.

Enzyme leakage study

In a 30 mL glass bottle were added 20 mL of maltodextrin (33% w/v in water) and 2 mL of CIP-silica mCLEA. The resulting suspension was shaken at 150 rpm at 32 °C for 10 days in a Stuart Orbital Incubator SI500. After this time the CIP-silica mCLEA was removed with a magnet (40 x 20 x 10 mm, neodymium magnet) and the supernatant was kept. The supernatant (0.2 mL) was added to 30 mL fresh maltodextrin (33% w/v in water) and shaken at 150 rpm at 32 °C for 3 hours in a Stuart Orbital Incubator SI500. After 1 , 2 and 3 h, a 0.1 mL sample was added to 0.9 mL water, centrifuged, filtered (0.2 μιη Phenex™ syringe filter) and transferred to an HPLC vial and subjected to HPLC analysis for glucose determination. No additional glucose formation could be detected and therefore it was concluded that no enzyme was leaking from the CIP-silica mCLEA. Recycles

In a 100 imL conical flask were added 30 imL of maltodextrin (33% w/v in water) and 0.2 imL of CIP-silica mCLEA. The resulting suspension was shaken at 150 rpm at 32 °C for 48 hours in a Stuart Orbital Incubator SI500. Samples of 0.1 imL were taken after 24 and 48 h. The 0.1 imL sample was added to 0.9 imL water, centrifuged, filtered (0.2 μιη Phenex™ syringe filter) and transferred to an HPLC vial and subjected to HPLC analysis. After 48 h, the CIP-silica mCLEA was removed with a magnet (40 x 20 x 10 mm, neodymium magnet), washed five times with 10 imL water and was then reused by adding another 30 imL maltodextrin mixture. This reuse constitutes the first recycle. In total, ten of such recycles were performed without loss of glucose amylase activity of the CIP-silica mCLEA, indicating excellent immobilised enzyme stability and highly effective magnetic separation.

Iron Leakage

Iron leakage experiments were carried out with an amount of

CIP-silica mCLEA that contains 6 mg of CIP-silica particles. This amount was added to 18 mL of incubation mixture (1 wt% lactic acid of pH 3). Incubation of the CIP-silica mCLEA was performed by shaking the mixture for 72 hours at 32 °C at 150 rpm in a Stuart Orbital Incubator S1500. Absorption in the iron detection assay after dissolution of the total amount of CIP-silica particles was determined by dissolving the sample in concentrated HCI and subsequent dilution. The iron leakage (%) after incubation was determined by directly comparing the absorption of the incubation supernatant of the CIP-silica mCLEAto that of the absorption after total dissolution. Leakage of Fe(ll,lll) was <1 %. Although the conditions chosen for incubation were extreme with respect to common enzyme applications, iron leakage from the CIP-silica mCLEA was negligible. Magnetic strength

The saturation magnetization of the dry glucose amylase CIP-silica mCLEA was measured at 161 emu*g^"1 (Am²/kg).

Example 3. CIP-silica magnetizable CLEA (mCLEA) of

glutaminase - low protein loading

Synthesis

In a 500 mL beaker were added 100 mL of a glutaminase enzyme solution, which was prepared by dissolving 10 grams of glutaminase (Amano glutaminase SD-C100S) in 50 imM potassium phosphate buffer of pH 6 to a total volume of 100 mL. An amount of 20.05 g of CIP-silica particles (prepared according to Example 1 ) was added to the enzyme solution. The resulting suspension was stirred for 15 minutes with an overhead stirrer (Velp

Scientifica ES overhead stirrer). Subsequently 400 mL of saturated

ammonium sulfate solution was added to the suspension to precipitate the enzyme in the presence of the CIP-silica particles. To allow full precipitation of the enzyme, the suspension was stirred for 1 hour at room temperature. After precipitation, 20 mL of a 25 wt% glutaraldehyde solution were added and the mixture was cross-linked overnight. The CIP-silica mCLEA was removed with a hand held magnet (ERIEZ Mega Rare Earth Tube Magnet 150 mm, 10700 Gauss) and was washed 4 times with H₂O (500 mL end volume). Each wash was stirred for 15 minutes with the overhead stirrer. After this, the CIP-silica mCLEA was stored in 100 mL 50 imM potassium

phosphate buffer pH 6. Total dry weight of the CIP mCLEA was 22.7 grams. Figure 3 shows a Microscope image of the CIP-silica mCLEA of glutaminase obtained from a Bresser microscope using a 40x magnification lens and Mikro CamLab software (Version 6.1 .4.0).

Activity assay

An L-glutamine substrate solution of 250 imM was prepared by dissolving 3.65 g of L-glutamine in 100 mL of 50 imM potassium phosphate buffer of pH 6. In a glass reaction vial, 10 mL of substrate solution were added. The vial was placed in a Stuart Orbital Incubator SI500 at a

temperature of 50 °C. The vial was left to shake for 5 minutes to allow reaching the desired temperature. Hereafter, 10 μΙ_ of the enzyme solution or CIP-silica mCLEA suspension were added to start the catalytic conversion. After 5, 10 and 20 minutes, a sample of 500 μΙ_ was taken and placed in an Eppendorf tube. The sample was heated in a water bath at 95 °C for

10 minutes. Glutaminase-catalysed L-glutamic acid release from L-glutamine was determined using an L-glutamic acid assay kit (K-GLUT) from Megazyme International, Ireland.

The glutaminase CIP-silica mCLEA showed 81 % activity recovery versus the free enzyme in this assay. Cross-linked protein content of the CIP-silica mCLEA was 12 wt%. Glutaminase activity was neither detected in the supernatant of the CLEA preparation nor in the washing water.

Enzyme leakage study and CIP-silica mCLEA stability

A part of the CIP-silica mCLEA suspension prepared as described above was incubated for 7 days at 50 °C, while shaking in a Stuart Orbital Incubator SI500. After incubation the supernatant was separated from the CIP mCLEA and the CIP mCLEA was washed three times with 50 imM potassium phosphate buffer of pH 6. Activity assays where then carried out on the supernatant and the CIP-silica mCLEA. No activity was detected in the supernatant and the glutaminase CIP-silica mCLEA had a residual activity of 98%, indicating that the CIP-silica mCLEA is stable at 50 °C and that no or negligible leakage takes place during 1 week of incubation at 50 °C.

Recycles

A 250 imM glutamine solution was prepared by dissolving 3.5 g of glutamine in 100 mL of 250 imM potassium phosphate buffer of pH 6. In a 250 mL round bottomed flask, 100 mL of the substrate solution were added. The solution was heated to 50 °C in an oil bath, while stirring with an overhead stirrer (Velp Scientifica ES overhead stirrer). Once the desired temperature of 50 °C was reached, 2.26 mL of glutaminase CIP-silica mCLEA suspension (513 mg of CIP-silica mCLEA) as prepared above was added to the reaction. After stirring for 15 hours a sample of 1 imL was taken from the reaction mixture. Glutaminase-catalysed L-glutamicacid release from

L-glutamine was determined using an L-glutamic acid assay kit (K-GLUT) from Megazyme International Ireland. The glutaminase CIP-silica mCLEA was removed from the reaction mixture using a magnet (40 x 20 x 10 mm, neodymium magnet) and washed once with 50 imM potassium phosphate buffer pH 6, after which it was used in the next cycle. Five consecutive cycles were performed under the same conditions, all leading to a conversion of 68 ±2%, indicating a stable immobilized enzyme and an effective magnetic separation.

Magnetic strength

The saturation magnetization of the dry glutaminase CIP-silica mCLEA was measured at 208 emu-g^"1 (Am²/kg).

Example 4. APTES functionalized magnetite magnetic CLEA preparation glucose amylase

APTES functionalized magnetite was produced as described in WO2012023847A2. The magnetic particles were washed with (NH₄)₂SO₄ before use. In a 2 L plastic beaker, a mixture of 0.7 L of saturated ammonium sulphate and 0.124 kg of APTES functionalized magnetite was stirred with an overhead stirrer (Velp Scientifica ES overhead stirrer) at room temperature for 1 h. Thereafter, 0.3 L of glucose amylase (Zibo Guoao, Shandong, China) was added slowly and the mixture stirred at room temperature for 1 hour. After the addition of 0.153 L of 25 wt% glutaraldehyde, the reaction mixture was stirred at room temperature for 18 h. The resulting mCLEA was removed with a hand held magnet (ERIEZ Mega Rare Earth Tube Magnet 150 mm, 10700 Gauss) and washed five times with 4.5 L of water. The final mCLEA was suspended in 1 L of water. The total dry weight of the APTES

functionalized magnetite mCLEA was 245 grams. Activity assay

Following the activity assay and HPLC analysis as described in Example 2, it was found that the glucose amylase APTES functionalized magnetite mCLEA showed 20% activity recovery versus the free enzyme in this assay. Cross-linked protein content of the APTES functionalized magnetite mCLEA was approximately 50 wt%. Glucose amylase activity was neither detected in the supernatant of the CLEA preparation nor in the washing water.

Iron Leakage

Iron leakage experiments were carried out with an amount of APTES functionalized magnetite mCLEA that contains 6 mg of APTES functionalized magnetite in 18 mL of incubation mixture (1 wt% lactic acid of pH 3).

Incubation of the CIP-silica mCLEA was performed by shaking for 72 hours at 32 °C and 150 rpm in a Stuart Orbital Incubator S1500. Absorption in the iron detection assay after dissolution of the total amount of APTES coated magnetite was determined by dissolving the sample in concentrated HCI and subsequent dilution. The iron leakage (wt%) after incubation was determined by directly comparing the absorption of the incubation supernatant of the APTES functionalized magnetite mCLEA to that of the absorption after total dissolution. After 72 h of incubation in 1 % lactic acid of pH 3, 35 wt% of the magnetite in the CLEA had dissolved.

Magnetic strength

The saturation magnetization of the APTES functionalized magnetite was measured at 37.3 emu*g^"1 (Am²/kg) and dry glucose amylase APTES functionalized magnetite mCLEA was measured at 16.1 emu*g^"1 (Am²/kg).

Claims

1 . A cross-linked enzyme aggregate (CLEA) comprising magnetizable particles, characterized in that the magnetizable particles comprise particles of a zerovalent metal selected from the group of iron, nickel, cobalt and alloys thereof, which particles are fully covered with a coating.

2. A CLEA according to claim 1 , comprising a plurality of CLEA particles having an average diameter of 1 μιη or larger, in particular in the range of 2-100 μιη.

3. A CLEA according to claim 1 or 2, wherein the average diameter of the

magnetizable particles is in the range of 0.010-20 μιη, in particular in the range of 1 -15 μιη.

4. A CLEA according to any of claims 1 -3, wherein the magnetizable particles have a spherical form, or deviate from a spherical form in that their largest dimension is up to two times larger than their smallest dimension.

5. A CLEA according to any of claims 1 -4, wherein the magnetizable particles are derived from carbonyl iron powder (CIP).

6. A CLEA according to any of claims 1 -5, wherein the average number of

magnetizable particles in a CLEA particle is in the range of 1 -100 and/or wherein the content of magnetizable particles is in the range of 50-95 wt%.

7. A CLEA according to any of claims 1 -6, wherein the coating is functionalized.

8. A CLEA according to any of claims 1 -7, wherein the coating comprises silica.

9. A CLEA according to any of claims 1 -8, wherein the coating has a thickness in the range of 10 - 500 nm.

10. A CLEA according to any of claims 1 -9, wherein the saturization

magnetization of the CLEA is in the range of 80-200 emu-g^"1.

1 1 . A CLEA according to any of claims 1 -10, wherein the enzyme is selected from the group of hydrolases, such as esterases, proteases, amidases, cellulases, nitrilases, xylanases and glycosylases; lyases, such as hydroxynitrile lyases and aldolases; oxidoreductases such as alcohol oxidases, peroxidases, ketoreductases and imine reductases; and transferases, such as

transaminases.

12. Method for preparing a CLEA according to any of claims 1 -1 1 , comprising

- providing a solution of an enzyme; then

- precipitating the enzyme to form an enzyme aggregate; then

- cross-linking the enzyme aggregate with a cross-linking agent in the

presence of magnetizable particles comprising a particle of a zerovalent metal selected from the group of iron, nickel, cobalt and alloys thereof, which particle is fully covered with a coating.

13. Method according to claim 12, wherein the cross-linking agent is selected from the group of glutaraldehyde and aldehyde cross-linkers derived from polysaccharides.

14. CLEA obtainable by the method of claim 12 or 13.

15. Process comprising the use of a CLEA of claims 1-1 1 and 14 for a catalytic conversion, wherein the CLEA is separated from the reaction medium by collecting the CLEA with the aid of an external magnetic field, in particular with the aid of a permanent magnet.