Gels have a wide variety of applications, for instance in the biomedical field as coatings, in contact lenses and in drug delivery systems but also in the non medical field such as coatings and cosmetic applications.

A hydrogel has been defined (Peppas et al., Hydrogels in Medicine and Pharmacy, Vol. I fundamentals, Chapter I, pages 1-25, (1986) CRC Press Inc. Boca Raton, USA) as a water swollen network (cross-linked structure) of hydrophilic homo-polymers or copolymers with a three dimensional appearance of which the cross-links can be formed by covalent and/or ionic bonds, and/or by van der Waals and/or hydrogen bonds.

Analogously a gel can be defined as such a network of homo-polymers or copolymers which is swollen by means of a solvent or a mixture of solvents.

US 4,587,127 describes the fermentative production of liquid seasoning. According to this disclosure, lactobacilli and yeasts may be immobilized on any conventional carrier material, an example being an alginate gel, entrapment of cells in alginate being a widely used mild immobilization procedure. However, alginate gel is not suitable because it is sensitive to abrasion and chemically unstable. The inventors theorize that alginate gel weakens under the influence of salt (NaCl) due to the replacement in the gel network of calcium ions with sodium ions. Chemically cross-linked gels, for example a polyethylene-oxide gel, have been used instead.

The viability of the cells after immobilization is however very poor with these gels.

To solve the problem associated with (physical) weakness of the immobilization-or carrier material such as alginate gels, one can choose a different material, such as cross-linked polymer materials, obtained by cross-linking of (pre-) polymers.

Conventionally, for such cross-linked polymer carrier material a pre-polymer is first made (or directly obtained from supplier) by polymerisation of suitable monomers, which polymerisation can be started by e. g. using an initiator. Thereafter cross-linking (to provide sufficient rigidity of the carrier and to allow swelling) is effected by adding a cross-linking agent after the pre-polymers are formed. In order to ensure cells of the micro-organisms are distributed on/in the carrier, cross- linking is effected in the presence of the micro- organisms which are to be immobilized.

Although usually at least a small percentage (e. g. 3-.) will survive the standard polymerization procedure, which can then left to grow and propagate first to higher numbers before the material is used in fermentation, this is unattractive. The formed gel will need to be held under growing conditions for the micro- organisms to grow in the carrier material to a desirable number in order to get economical conversion rates, before actual fermentative production can start.

Hence, there is a need for immobilized micro- organisms, which micro-organisms are immobilized on or in a carrier material that can resist medium to high shear, in particular under conditions of high salt (e. g. 5k or higher, preferably at least 6t, more preferably at least 8% wt.). Preferably, the material should be such shear resistant in the presence of salt that it can be used in stirred tank reactors. It is also desired that the micro- organisms can be immobilized onto or into the carrier material in such a way that a substantial number of cells of the micro-organisms survive the immobilization

technique. Preferably, at least 10% of cells should survive, more preferably at least 20%.

An object of the present invention is to provide polymer immobilized living cells and applications thereof in gels, which combat the above referred to problem.

According to a first aspect the present invention provides a process for immobilizing living cells in a polymer according to claim 1-17.

The inventors have shown that polymer immobilized living cells according to the present invention exhibit high survival rates and accordingly good viability.

According to a second aspect of the present invention there is provided a polymer having an immobilized living cell therein.

According to a third aspect the present invention provides a gel comprising a solvent swollen network of cross-linked polymer (s), with one or more living cells immobilized therein.

For the present invention the following definitions are used: At least one of the (pre-) polymers utilised preferably has at least one multi-functional site.

A functional group is a chemical entity which is capable to directly or indirectly (via a reagent) participate in a reaction or interaction, for instance cross-linking.

A multi-functional site is a sequence of more than one functional group.

A multi-functional polymer is a polymer comprising one or more multi-functional sites and/or more than one functional group.

An end-multi-functional polymer is a polymer (including star-like polymers) having at least at one end of the polymer chain a multi-functional site. In one embodiment the multi-functional sites are at both or at all ends of the polymer.

A core-multi-functional polymer is a polymer having as a core a multi-functional site and grafted/bound thereto at least one polymer chain resulting in more than one polymer end-group. Within one polymer more than one core may be present.

A pendant-multi-functional polymer is a polymer having along the polymer chain as a side group more than one functional group.

A homo-multi-functional polymer is a polymer with only one type of functional group whereas a hetero- multi-functional polymer is a polymer with more than one type of functional group.

By utilising polymers with multi-functional sites, as described above, gels with regular network structures and/or a better incorporation of the polymers in the final network are provided.

Advantages of the use of polymers with multi- functional units over the use of polymers with mono- functional units for the preparation of polymeric networks for immobilizing living cells are: i) enhanced reactivity; the statistical chance of coupling with multi-functional units is increased.

This results in shorter reaction and gelation times and a large degree of freedom in reaction conditions and in a better incorporation of gel components in the network; and ii) enhanced stability of the resulting network preventing or retarding the release of network components. Possibly this arises from multi-point attachment per functional site.

These two advantages results in a better control over the gel properties (chemical, physical and mechanical) enabling to control important gel-parameters like the cross-link density, permeability, swelling and mechanical strength.

Additionally, with the multi-functional sites an additional phase is incorporated in the gel that may retain functionality. This can be used for the

immobilization (chemically or physically) of additional components that need to be released (drugs) or taken up (toxic substances) or to influence the swelling of the gel. This phase can also be functional in the adhesion/immobilization of the gel to a substrate.

Further, an enlarged flexibility in network compositions can be covered. The use of end-multi- functional polymers offers the possibility to obtain mechanically stable networks with very high swelling characteristics.

In the case the functional sites are positioned at the ends of a polymer chain this type of polymer has an absence of cross-link sites between the chain ends, providing control over the degree and positioning of cross-links within the gel network.

By choosing a specific polymer chain length between the multi-functional sites, the swelling characteristics and the distance between two cross-links of the desired gel can be controlled.

The cross-linker may be an organic molecule, multivalent ions and an energy source, for instance light, ultraviolet or gamma-radiation. For the cross- linking of aqueous systems containing both carboxylic acid and amine functional groups a preferred cross-linker for the preparation of hydrogels is N-ethyl-N'- (3- dimethylaminopropyl)-carbodiimide.

The chemical nature of the chains and the chemical nature of the multi-functional groups/coupling groups can be different, for example these being non- ionic, anionic or cationic.

By pre-selecting the chain lengths and type of solvent, and the chemical nature of the chains and multi- functional site (s), gels can be produced with a combination of desired characteristics.

The multi-functional site is preferably selected from the group oligomers or polymers based on: acrylates, acrylic acid, allylamines, amino-acids, sugars/saccharides, dendrimers, ethylene imine,

vinylalcohol, styrene, azido-compounds, etc., and derivatives thereof like for instance cellulose, dextran, chitosan, chitine, etc.

In the case of end-multi-functional polymers that substantially lack functional groups between the chain ends, and wherein the chemical nature of the multi- functional groups is substantially the same (homo-end- multi-functional polymers), the polymer is preferably selected from the group consisting of: poly (ethylene imine)-poly (ethylene oxide)-poly (ethylene- imine) triblock copolymer, poly (acrylic acid)-poly (ethylene oxide)-poly (acrylic acid) triblock copolymer, poly (ethylene imine)-poly (ethylene oxide-co-propylene oxide)-poly (ethylene imine) triblock copolymer, poly (acrylic acid)-poly (ethylene oxide-co-propylene oxide)-poly (acrylic acid) triblock copolymer, poly (ethylene imine)-poly (ethylene oxide)-poly (propylene oxide)-poly (ethylene oxide)-poly (ethylene imine) pentablock copolymer, poly (ethylene imine)-poly (propylene oxide)-poly (ethylene oxide)-poly (propylene oxide)-poly (ethylene imine) pentablock copolymer, poly (acrylic acid)-poly (ethylene oxide)-poly (propylene oxide)-poly (ethylene oxide)-poly (acrylic acid) pentablock copolymer, poly (acrylic acid)-poly (propylene oxide)-poly (ethylene oxide)-poly (propylene oxide)-poly (acrylic acid) pentablock copolymer.

In the case of core-multi-functional polymers, the polymer is preferably selected from the group consisting of: poly (ethylene oxide) n-poly (ethylene imine), poly (ethylene oxide) n-poly (acrylic acid), poly (ethylene oxide)-poly (propylene oxide) copolymern- poly (ethylene imine), poly (ethylene oxide)-poly (propylene oxide) copolymern- poly (acrylic acid).

In the case of pendant multi-functional polymers, the polymer is preferably selected from the group consisting of pendant multi-functional homo-and copolymers: carboxymethyl cellulose, carboxymethyl dextran, carboxymethyl amylose, chitine, chitosan, chondriotine sulphate, dendrimers, dextran sulphate, heparan sulphate, heparin, poly (acrylic acid), poly (methacrylic acid), poly (amino acids) including copolymers, poly (ethylene imine), poly (hydroxyethyl methacrylate-methacrylic acid), polysaccharides, proteins.

Examples of anionic functional groups include COO-, SO3-, S04-, PO42-, etc., examples of cationic functional groups include N+RR'R", N+R2R', N+R3, N+R2H, N+RR'H, N+H3, etc. and examples of non-ionic functional groups include aldehyde, amino, carboxy, epoxy, halogen, hydroxyl, thiol, etc.

It should be noted that certain functional groups have been accorded the non-ionic and ionic status, for instance the carboxylic acid group and amino group.

The reason for this is that these groups are used for covalent binding in some applications and in other applications for ionogenic interaction. In the first instance these groups are considered to be non-ionic and in the second instance ionic.

Further aspects of the invention include applications of the gels and uses thereof. For example as an anti-thrombogenic lubricious coating of a medical article or system.

The so-prepared immobilized microorganisms may be used in a further aspect of the present invention, which relates to a process for cultivating microorganisms and/or fermentation by microorganisms for example.

In the process according to the invention, it appears that a considerable higher percentage of micro- organisms survive the immobilization procedure for immobilization on a cross-linked polymer carrier then

using a conventional technique. As such cross-linked polymer carrier material is sufficiently rigid, it can be applied in reactors in which medium to high shear exists.

Without wishing to be bound by theory, it is believed that apparently the pre-polymers that are contacted with the cross-linking agent, incubated, and are still reactive (to form cross-links) via the second biopolymer), are less toxic than the cross-linking agent to the cells.

The process according to the invention can be used for immobilizing many types of microorganisms, especially the living cells as defined in claim 20.

As the process according to the invention can suitably be used for preparing flavouring components, it is preferred to choose those microorganisms that produce flavouring components, and in particular seasoning and savoury flavouring components.

Examples of such flavouring components are 4- ethylguaiacol, guaiacol, sotolon, furanone, furaneol, homofuraneol, norfuraneol, hydroxyfuraneol, methoxyphenol.

Also, the process according to the invention may be carried out as (part of) a process for preparing soy sauce (or shoyu), and more in particular in the moromi or fermentation phase of preparing the soy sauce.

As the microorganisms are now immobilized, the process may be carried out conveniently in a continuous way. Although such continuous processing may be carried out in any conventional equipment for continuously fermenting, it is in particular preferred to apply such in reactor equipment in which medium to high shear regimes exist. Examples of such equipment are stirred tank reactors.

Background of the invention The soy-sauce yeasts, Zygosaccharomyces rouxii and Candida versatilis, are important flavour producers in soy-sauce processes. In these processes, Z. rouxii

produces ethanol and other flavour components like 4- hydroxy-2 (or 5)-ethyl-5 (or 2)-methyl-3 (2H)-furanone (Nunomura et al., 1976), while C. versatilis produces phenolic compounds, like 4-ethylguaiacol (Roling, 1995).

These phenolic compounds give the characteristic flavour to soy-sauce.

In order to shorten the process time, to increase the production efficiency and to make continuous operation of soy-sauce processes easier, much attention has been paid in the last decade to the application of immobilized soy-sauce yeasts (Osaki et al., 1985; Hamada et al., 1989; Hamada et al., 1990a; Hamada et al., 1990b; Horitsu et al., 1990; Horitsu et al., 1991; Iwasaki et al., 1991; Hamada et al., 1992; Motai et al., 1993). It is shown that the immobilization of soy-sauce yeasts considerably decreased the total time for the soy-sauce process. In most of these investigations, alginate gel was used as immobilization material because immobilization in this gel is a mild and convenient method that can be scaled up.

However, as stated above, alginate gel has the disadvantage of mechanical weakness (Horitsu et al., 1990; Muscat et al., 1996) and furthermore, if is chemically unstable towards high salt concentrations (Martinsen et al., 1989; Horitsu et al., 1990). The latter is also expected for alginate gel beads in soy sauce medium in which the salt content is high (Horitsu et al., 1990). Because of these disadvantages, a continuous long-term process with soy-sauce yeasts immobilized in alginate gel is not feasible (Horitsu et al., 1990). For this reason, a ceramic carrier to be used as support material was developed instead.

Leenen et al. (1996) made a polyethylene-oxide gel by adding a cross-linker to a mixture of a prepolymer solution and a cell suspension. This polyethylene-oxide- based gel has better characteristics than alginate gel.

The gel does not dissolve in the presence of salt and is insensitive to abrasion; therefore, the durability of

this polyethylene-oxide gel is expected to be high (Leenen et al., 1996). However, this immobilization process was toxic which resulted in low survival of the cells. Leenen et al. (1996) estimated that 0.5% of their Nitrosomonas europaea cells survived the immobilization process. Similarly low survival percentages were measured by Tanaka et al. (1996) for nitrifying sludge immobilized in a polyethylene-glycol gel. In a preliminary experiment, the inventors did not even observe any survival of the soy-sauce yeast Z. rouxii at all, after immobilizing in the polyethylene-oxide gel according to the procedure as described by Leenen et al. (1996).

The inventors have directed their efforts to improving this immobilization procedure for Z. rouxii cells. The effect of the concentration of the prepolymer solution and the cross-linker on the activity of Z. rouxii was determined. The effect of the contact time with the cross-linker on the activity of Z. rouxii was also determined. Subsequently, a new, mild immobilization procedure for the soy-sauce yeasts in polyethylene-oxide gel was developed which resulted in high survival percentages of Z. rouxii and C. versatilis.

This new gel was investigated further in rheological studies. The formation of the polyethylene- oxide gel network was followed in time by measuring the storage modulus. This modulus is a measure for the elastic energy stored in the material (Whorlow, 1992); it will increase during gelation and level off when the gelation is completed (Verheul et al, 1998). Furthermore, the sensitivity of the polyethylene-oxide and alginate gel to abrasion was studied with oscillation experiments as was done before by Martins dos Santos et al. (1997).

Oscillation experiments show the fatigue of gel materials which is likely to be related to the abrasion sensitivity of these materials in bioreactors (Martins dos Santos et al., 1997). In addition, the salt effect on the abrasion sensitivity for both gels was determined. Finally, the

absence of fracture during a large-deformation experiment showed the flexibility of the new polyethylene-oxide gel.

Experimental The invention will now be described with reference to the following experimental results and figures which show: Figure 1 Effect of the incubation time in soy- sauce medium on the mean force at fracture of alginate gel beads. The soy-sauce medium contained 13% (w/v) NaCl.

To 500 ml of this medium, 100 ml gel beads were added.

Error bars show the 95k confidence interval of the measurements.

Figure 2 Effect of the contact time with cross- linker (g) and the cross-linker yeast ratio (t) on the survival of Z. rouxii. The applied conditions were respectively a cross-linker yeast ratio of 0.4 g. g-1 and a contact time of 50 min.

Figure 3 The formation of the polyethylene- oxide gel network, as shown by the storage modulus, in time. The timer was started after all the gel components were thoroughly mixed.

Figure 4 The effect of support material and the presence of salt on the evolution of the resistance to compression. Legend to the figure: gel resistance (N. m-2) as a function of number of compressions. Closed symbols for alginate gel (circles for PBS and squares for PBS + 12.5 w/v% NaCl), open symbols for PEI-PEO-PEI/CMC gel (circles for PBS and squares for PBS + 12.5 w/v% NaCl).

Figure 5 The effect of deformation on the storage modulus of polyethylene-oxide gel.

The immobilization method according to the present invention (and alginate gel as reference) was investigated on sensitivity to abrasion with oscillation experiments as was done before Martins dos Santos et al.

(Martins dos Santos, V. A. P., Leenen, E. J. T. M., Rippoll, M. M., Van der Sluis, C., Van Vliet, T., Tramper, J. and Wijffels, R. H. Relevance of rheological properties of gel

beads for their mechanical stability in bioreactors.

Biotechnol. Bioeng. 1997,56,517-529). Oscillation experiments show the fatigue of gel materials which is likely to be related to the abrasion sensitivity of these materials in bioreactors (Martin dos Santos et al., 1997). In addition, the salt effect on the abrasion sensitivity for both gels was determined as well.

Finally, the absence of fracture during a large deformation experiment showed the flexibility of the new polyethylene oxide gel.

Materials and methods Yeast cultures Zygosaccharomyces rouxii CBS 4021 and Candida versatilis CBS 4019 were cultivated in 300 ml Erlenmeyer flasks, containing 100 ml medium, on a rotary shaker (Gallenkamp, Orbital Incubator) at 28°C and 200 rpm.

Media The medium used for Z. rouxii had the following composition per liter of demineralized water: 40 g glucose. laq (Merck), 5 g pepton (Sigma) and 5 g yeast extract (Oxoid). The components were separately autoclaved for 20 min at 120°C. For C. versatilis, a defined medium with the following composition per liter of demineralized water was used: 22 g glucose. laq (Merck), 21 g citric acid. laq and 6.7 g Bacto Yeast Nitrogen Base without amino acids (Difco). The pH of this medium was adjusted to 5 using NaOH and the medium was filter sterilized (Nalgene, 0.2 Um filters).

Preparation of yeast suspension The yeast cells were concentrated by centrifugation for 15 minutes at 9,500 g and 4°C (Beckman J2-MC centrifuge), when they were in the early stationary growth phase. After centrifugation, the concentrated cell

suspension was washed with sterilized PBS-buffer (pH 7.4) of the following composition per liter of demineralized water: 8.2 g NaCl (Merck), 1.9 g Na2HPO4.2aq (Merck) and 0.3 g NaH2PO4.2aq (Merck).

Hereafter, the yeast suspension was centrifuged again. The pellet obtained was used for the determination of the toxicity of the prepolymer solution and cross- linker or for immobilization, after appropriate dilution in PBS-buffer.

Example 1 Determination of the toxicity of the prepolymer solution and cross-linker The toxicity of the prepolymer solution, containing carboxy-methyl-cellulose and polyethylene- oxide, and cross-linker (carbodiimide) was determined by separately incubating them for 1 hour with a Z. rouxii suspension and hereafter, estimating the survival percentage with the respiration activity assay. The toxicity of the cross-linker on Z. rouxii was determined further as a function of the contact time with the yeast and the cross-linker yeast ratio. For this, both respiration activity assay and colony count on agar plates were used.

Immobilization procedure Example 2 N-hydroxysuccinimide (NHS, 20 mg) was added to 18 ml of a carboxymethyl cellulose (CMC) solution in phosphate buffered saline (PBS, 5 mg CMC/ml). After complete dissolution of the NHS, N-ethyl-N'- (3- dimethylaminopropyl)-carbodiimide (EDC, 70 mg) was added and the solution was mixed for 5 minutes to allow activation of carboxylic acid groups.

In the meantime 2 ml of a solution of poly (ethylene imine)-poly (ethylene oxide)-poly (ethylene imine) triblock copolymer with a molecular weight of 21.200 D (molecular

weight of poly (ethylene imine) blocks 600 D) in PBS (100 mg triblock copolymer/ml PBS) was mixed with 1.72 ml of a yeast suspension and the resulting suspension was subsequently added to the activated CMC solution while mixing. Directly after this addition the complete yeast suspension was poured into a stirred, baffled reaction vessel which contained 200 ml hexadecane. After gelation, the spheres containing the immobilized yeast were isolated by sieving and rinsed with a solution of glucose in PBS (10 gram/liter).

Example 3 In the process according to the present invention, carbodiimide (cross-linker) was added to only a carboxy- methyl-cellulose solution and not to a mixture of prepolymer solution, containing both carboxy-methyl- cellulose and polyethylene-oxide, and yeast suspension as previously had been done. For this, 0.125 ml cross-linker solution (0.1 mg/pl) and 4.5 ml carboxy-methyl-cellulose solution were mixed and allowed to react for 5 minutes.

Then a mixture of 0.4 ml yeast suspension and 0.475 ml polyethylene-oxide solution was added and thoroughly mixed. After 1 hour of gelation, the cylindrical gel was washed with PBS-buffer in order to remove possible excess of cross-linker.

Alginate gel was made by mixing 5 ml of 6% (w/v) alginate solution (Protanal LF 10/60, high guluronic acid content, Pronova Biopolymers, Norway) with 1 ml yeast suspension. Hereafter, 4 ml of a 12.5 (w/v) CaCl2 solution (Merck) was added to make a cylindrical gel.

To determine the survival percentage with the respiration-activity assay after immobilization, both polyethylene-oxide and alginate gel were cut into small pieces. For the rheological tests, the gels were made without yeast.

Alginate gel beads with a diameter of about 1 mm were made by a dropwise extrusion of a 3% (w/v)

alginate solution through a hollow needle using air pressure. The droplets were collected in a stirred 5% (w/v) CaCl2 solution and left for at least 24 hours for hardening. The resulting gel beads were used to evaluate their chemically stability towards a soy-sauce medium.

For this, 100 ml gel beads were added to 500 ml soy-sauce medium, containing 13% (w/v) NaCl. The medium with beads was stored at room temperature for three days and during that time, the force to fracture these beads was measured.

Example 4 Respiration activity assay The respiration activity assay was used in order to determine the survival percentage after incubating the cells with prepolymer solution and cross- linker or after immobilizing them. For this, the oxygen consumption rate of the incubated or immobilized cells and that of untreated cells was measured. From this, survival percentages were calculated. The oxygen consumption rate was measured at 30°C in a Biological Oxygen Monitor (BOM, Yellow Springs Instruments, Ohio, USA). To a 24 cm3 vessel, containing PBS-buffer, a known amount of cells (free or immobilized) was added and aerated for 10 minutes. After aeration, the vessel was sealed with an oxygen electrode (model 5331, Yellow Springs Instruments, Ohio, U. S. A.) and the decrease in oxygen concentration was recorded as a function of time.

The initial oxygen consumption rate was used to calculate the respiration activity.

Example 5 Colony count on agar plates The colony count on agar plates was used to determine the survival percentage after incubating the cells with cross-linker. For this, the colony count from a yeast suspension incubated with cross-linker was divided by that from an untreated yeast suspension. For

this, an appropriate dilution of the yeast suspension in sterilized PBS-buffer was made and added to agar plates.

After the yeast was grown on the plates for two weeks at 25°C, the number of colonies was counted. The agar plates were made by adding 4.0 g glucose. laq (Merck), 1.0 g Bacto Yeast Nitrogen Base without amino acids (Difco), 20 g agar (Technical agar (no. 3), Oxoid) to 1 1 demineralized water. Before this the agar was autoclaved for 20 minutes at 120°C and the glucose and Bacto Yeast Nitrogen Base were filter sterilized (Nalgene, 0.2 Um filters).

Example 6 Rheoloqical tests The force to fracture beads was measured in order to evaluate the chemical stability of alginate gel beads towards a soy-sauce medium. For this, at different points of time during the incubation, a sample was taken and 12 gel beads from this sample were separately compressed until fracturing. This compression was done at room temperature with a tension-compression device (Overload Dynamics Table model S100) fitted with a 50 N load cell. On the fixed bar of this device, a single gel bead was placed on a wet filter paper in order to prevent dehydration. The gel bead was compressed using the moving bar at a fixed compression speed of 60 mm/min. The force needed for fracturing the bead was recorded.

The storage modulus during formation of the polyethylene gel network was measured with a CVO Rheometer System (Bohlin Instruments). This system was used with a concentric cylinder measuring geometry which consists of a rotating inner cylinder located in a fixed outer cylinder. In the annular gap between the two cylinders, the polyethylene-oxide gel was made according to the procedure described above. During gel formation, a sinusoidal oscillation was applied to the inner cylinder.

The frequency of the oscillation applied was 0.1 Hz.

Through the oscillation of the driving system, the inner

cylinder will also oscillate sinusoidal but at a smaller amplitude and with a phase difference, if the gel has a linear visco-elastic behaviour. Linear visco-elastic behaviour means that the storage modulus is independent of the magnitude of deformation and deformation rate applied (Whorlow, 1992). This will be the case when the amplitude of the oscillation is small enough (Roefs et al., 1990; Whorlow, 1992). In our case, the amplitude was kept low enough to ensure linear behaviour. From the amplitude and phase difference, some rheological properties like the storage modulus can be calculated (Roefs et al., 1990; Whorlow, 1992). The storage modulus was measured at 20°C and during the measurement the gelling solution was covered with paraffin oil in order to prevent drying out.

The CVO Rheometer System was also used for a large-deformation experiment, after the gelation was complete. During this experiment the storage modulus was measured as a function of the relative shear deformation in order to determine the maximum relative deformation at fracture.

Oscillation compression tests, as described by Martins dos Santos et al. (1997), were applied on gel cylinders in order to evaluate the fatigue of the immobilization materials. For this, polyethylene-oxide and alginate gel samples were made according to the procedures as described above. After gelation the samples were cut into gel cylinders of 20 by 20 mm. At least two cylinders of each sample were tested. The oscillation compression tests were done at room temperature with the same tension-compression device as used to measure the force to fracture gel beads. For the compression tests this device was fitted with a 2000 N load cell. The gel cylinders, which were placed in a beaker containing PBS- buffer in order to avoid dehydration, were 6 mm compressed with the moving bar at a compression speed of 50 mm/min. The applied relative deformation was small (0.3'-.) in order to stay within the linear region for both

gels. After compression the moving bar returned to its original position. This oscillation was repeated 1000 times for one gel cylinder. During these oscillations the resistance to compression in time was determined.

Furthermore, the salt effect on the fatigue of both polyethylene-oxide and alginate gel was determined by incubating the gel cylinders for at least 24 hours in a modified PBS-buffer containing 12.5% (w/v) NaCl (Merck) and doing the compression tests with the cylinders placed in the same modified buffer.

Results and discussion: Chemical stability of alginate gel beads towards soy- sauce medium Horitsu et al. (1990) expected that the high salt content of the soy-sauce medium will adversely affect the chemical stability of alginate gel beads.

Therefore, we determined the chemical stability of alginate gel beads towards soy-sauce medium by following the force to fracture the alginate beads during incubation in this medium. If the stability of the alginate gel is adversely affected by the salt, the gel will start to dissolve and the force to fracture will decrease because of a lowered gel concentration (Martinsen et al., 1989). For this experiment alginate with a high guluronic acid content was used because this kind of alginate gel has the highest tolerance to salt (Martinsen et al., 1989; Smidsrd and Skjak-Braek, 1990).

In Figure 1 the effect of the incubation time in soy-sauce medium on the mean force at fracture of alginate gel beads is shown. It appears from this figure that, after about 24 hours of incubation, this mean force was reduced by almost 50% and became more or less constant. This reduction in force was attended by some white colouring in the medium. These observations show that alginate gel was partly dissolved in the soy-sauce medium. Because the soy-sauce medium we used had a lower

salt content (13%) than the conventional one (16-18%, Osaki et al, 1985; Rowing et al., 1995), it is likely that alginate gel will dissolve even more in the latter.

From this, we concluded that alginate is chemically unstable towards soy-sauce medium.

Toxicitv of prepolymer solution and cross-linker In a preliminary experiment, we did not observe any survival of the soy-sauce yeast Z. rouxii after immobilizing in the chemically cross-linked polyethylene- oxide gel that we used before (Leenen et al., 1996). For this reason, we studied the toxicity of the cross-linking reaction of this gel. The polyethylene-oxide gel was made by mixing a carbodiimide (cross-linker) with a mixture of a prepolymer solution, containing carboxy-methyl- cellulose and polyethylene-oxide, and a cell suspension (see materials and methods section). We separately determined the effect of the prepolymer solution and cross-linker on the activity of Z. rouxii. This was done by incubating Z. rouxii cells with the prepolymer solution or cross-linker for 1 hour. This incubation time was chosen because the contact time of the yeast with prepolymer solution or cross-linker during the immobilization process needs to be 1 hour. After incubating the survival was determined using respiration activity assays. From this (data not shown), it appeared that the prepolymer solution was not toxic at all for Z. rouxii while the cross-linker was very toxic. This toxic effect of the cross-linker was not found by Leenen et al.

(1996) for Nitrosomonas europaea cells.

The toxicity of the cross-linker was studied further by determining the effect of the contact time with cross-linker and cross-linker yeast ratio on the activity of Z. rouxii. For this, both the respiration- activity assay and the colony count on agar plates were used. The contact time and cross-linker yeast ratio applied had the same order of magnitude as during the preliminary experiment (respectively 1 hour and 2 g. g.-').

The effect of the contact time and cross-linker yeast ratio on the survival of Z. rouxii, as determined with the respiration-activity assay, can be seen in Figure 2. Figure 2 shows that the survival of Z. rouxii cells strongly decreased with an increase in contact time and cross-linker yeast ratio. Similar results were obtained with the colony count on agar plates (data not shown). It can be seen in Figure 2 that, after about 1 hour, less than 20% of the Z. rouxii cells did survive at a cross-linker yeast ratio of 0.4 g. g-1. Figure 2 also shows that less than 300 of the Z. rouxii cells was still alive after incubating for 50 minutes at a cross-linker yeast ratio of 2 g. g-'. Hence, an even lower survival percentage can be expected when Z. rouxii cells are incubated for about 1 hour at a cross-linker yeast ratio of 2.0 g. g-1 like in the preliminary immobilization experiment. For this reason, the severe toxic effect of the cross-linker seems to be a good explanation for the absence of survival of Z. rouxii cells after immobilization in the polyethylene-oxide gel.

In order to reduce the severe toxic effect of the cross-linker during the immobilization process, it was tried to avoid the direct contact between the cross- linker and yeast. For this, the immobilization process was started without yeast by using the cross-linker to activate carboxy-methyl-cellulose. After this, the activated carboxy-methyl-cellulose was added to a mixture of polyethylene-oxide and yeast suspension, as described in the materials and methods section.

Example 7 Survival of soy-sauce yeasts in polyethylene-oxide gel according to the present invention The survival of the soy-sauce yeasts Z. rouxii and C. versatilis after immobilization in the polyethylene-oxide gel was determined with the respiration-activity assay in order to examine the toxicity of the new immobilization process. The same was

done for alginate. In Table 1 the survival percentages of the yeasts in both gels can be found. It can be seen in this table that the survival of both yeasts in the new polyethylene-oxide gel was excellent (more than 80°S).

This clearly demonstrates that the new immobilization process was much milder than the original process. It can also be seen in Table 1 that there was not a big difference between the survival percentages in both gels.

This shows that conditions for immobilizing the soy-sauce yeasts in the new polyethylene-oxide gel are comparable to the condition in alginate gel which are very mild (Smidsrd and Skjak-Braek, 1990; Leenen et al. 1996).

Table 1 Survival percentages of the soy-sauce yeasts Z. rouxii and C. versatilis* after immobilization in the new polyethylene-oxide and alginate gel. new alginate polyethylene-gel oxide gel Z. rouxii 86 74 C. versatilis 90 84 90 96 Example 8: Survival of yeast To study the effect of the cross-linker on the yeast, the survival of Z. rouxii (suspended in PBS) during incubation with cross-linker (EDC, 0.4 gram per gram yeast) was determined. The survival of the micro- organisms was determined by a respiration activity assay using a Biological Oxygen Monitor (BOM, Yellow Springs Instruments, Ohio, U. S. A.). The results showed that within 25 minutes the survival had fallen below 50k

whereas at 60 minutes the survival was less than 25%.

After 2 hours of incubation, survival was less than 10%.

When different EDC to yeast ratio's were applied at a fixed incubation time of 50 minutes, it was shown that less than 30% of cells survived at an EDC to yeast ratio of 2 gram per gram.

To study the effect of immobilization on yeast survival, Z. rouxii and C. versatilis were both immobilized in the PEI-PEO-PEI/CMC gel system as well as in alginate and the survival of both immobilisation procedures was compared.

For yeast immobilisation according to the PEI- PEO-PEI/CMC gel system, the polymer solutions as mentioned in the above example (see Yeast immobilization during microsphere preparation) were prepared and 0,400 ml of the yeast suspension was added to 0.475 ml of the PEI-PEO-PEI solution. Subsequently, 0.125 ml of an EDC solution (100 mg/ml) was added to 4,5 ml of the CMC solution in a petri dish and after thorough mixing this solution was allowed to react for 5 minutes. Then the yeast suspension was added to the CMC solution in the petri dish and again after thorough mixing the suspension was allowed to gelate. After 1 hour gelation, the gel was rinsed with PBS and the gel was processed into small particles using a scalpel knife.

For yeast immobilization in alginate, 1 ml of a yeast suspension in PBS was added to 5 gram of a 6 w/v% alginate (Protanal LF 10/60, high guluronic acid content, Provova Biopolymers, Norway) solution in demineralized water. After mixing, 4 ml of a 12,5 w/v% CaCl2 solution was added and after mixing the solution was allowed to gelate for 1 hour. Then the gel was processed into small particles using a scalpel knife.

Again the survival of the micro-organisms was determined by a respiration activity assay using a Biological Oxygen Monitor (BOM, Yellow Springs Instruments, Ohio, U. S. A.). The results of these experiments showed that the survival of the micro-

organisms was higher than 70% for all micro- organism/immobilisation system combinations and was more or less the same (differences less than 15%) for both immobilisation systems.

Rheolocrical tests with the gel according to the present invention Mechanical properties of the new polyethylene- oxide gel was investigated by using rheological tests.

The gelation was followed and the sensitivity to abrasion and the maximum relative deformation at fracture were determined.

Example 9 Gelation The gelation of polyethylene-oxide was followed in time by measuring the storage modulus, which is a measure for the elastic energy stored in the material (Whorlow, 1992). In increase in the storage modulus shows the formation of the gel network while the storage modulus becomes constant when the gelation is completed (Verheul et al., 1998). In Figure 3 the gelation of polyethylene-oxide is shown. In this figure it can be seen from the increase in storage modulus that the formation of the polyethylene-oxide gel network started about 20 minutes after all the gel components were mixed.

It can also be seen in this figure that the gelation is largely completed within about 4 hours. At that time, the storage modulus became more or less constant. During immobilization experiments we already started to wash the gel with PBS-buffer after 1 hour of gelation (see materials and methods section). This was done to remove a possible excess of the toxic cross-linker and appeared to have no effect on the formation of the gel network.

Example 10 Gel stability For the evaluation of the fatigue of the immobilization gel, oscillation compression tests (Martins dos Santos V. A. P., Leenen E. J. T. M., Rippoll M. M., Van der Sluis C., Van Vliet T., Tramper J., Wijffels R. H.,'Relevance of rheological properties of gel beads for their mechanical stability in bioreactors', Biotechnol. Bioeng. 56,517-529,1997) were applied on PEI-PEO-PEI/CMC and alginate gel cylinders (2 cm diameter, 2 cm height) of the same composition as mentioned in the example above (survival of yeast). To include the effect of NaCl on the fatigue properties of the gels, oscillation compression tests were also performed with cylinders that had been incubated in PBS that was enriched with 12.5 w/v% NaCl for at least 24 hours.

The oscillation experiments were performed at room temperature and to avoid dehydration, the cylinders were placed in PBS or PBS that was enriched with 12.5 w/vo NaCl. A compression of 6 mm was applied at a compression speed of 50 mm/min after which the moving bar returned to its original position. During a total of 1000 oscillations, the resistance of the gel compression was determined.

The results of the experiments are shown in figure 4 which clearly shows that the PEI-PEO-PEI/CMC gel is not affected by the number of compressions whereas the properties of the alginate gel decrease with increasing number of compressions. In case of the alginate gel the experiments even had to be stopped before a total number of 1000 compressions was reached due to excessive deformation of this gel. In the presence of high NaCl concentrations (PBS enriched with 12.5 w/v% NaCl) this was already the case after 30 compressions whereas at low salt concentration (PBS) 350 compressions could be performed.

Example 11 Abrasion sensitivitv The sensitivity to abrasion of the polyethylene-oxide gel was determined by oscillation experiments as described by Martins dos Santos et al.

(1997). They found that the sensitivity to abrasion of gel materials in bioreactors is likely to be related to the fatigue of the gel materials. The fatigue of gel materials can be determined with oscillation compression experiments. In these experiments, the gel material is exposed to repetitive compression and the change in resistance to compression is measured. If the resistance changes only slightly, the gel material will be less liable to fatigue and for that reason, less sensitive to abrasion than when the resistance alters considerably.

The oscillation compression experiments were done with both the new polyethylene-oxide and alginate gel. These experiments were also done with both gels in the presence of salt in order to predict their abrasion sensitivity in soy-sauce-like processes. For alginate again a gel with a high guluronic acid content was used, because this kind of alginate gel has not only the highest tolerance to salt, but also the highest mechanical stability (Smidsrd and Skjak-Brsek, 1990).

In Figure 4, the representative results of the oscillation experiments with the polyethylene-oxide and alginate gels can be found. This figure shows that the resistance to compression of the alginate gels at the start of the experiment was higher than that of the polyethylene-oxide gels, which reflects the difference in rheological behaviour of the two gels. Alginate gels are more viscous than polyethylene-oxide gels (Leenen et al., 1996). For this reason, alginate gels resist more against a relatively small compression, like we applied during these experiments, than polyethylene-oxide gels. However, a higher resistance does not give any information on the mechanical stability of the gel material to abrasion because the magnitude of the resistance measured in these

experiments is much higher than the maximum shear stress gel materials may encounter in bioreactors (Martins dos Santos et al., 1997). For the information about abrasion sensitivity, the change in resistance during the oscillation experiment is important.

Figure 4 shows that there was hardly any change in the resistance to compression of the new polyethylene- oxide gels, like the polyethylene-oxide gels made by Leenen et al. (1996). There was also no effect of salt on the changes in resistance. Therefore, the polyethylene- oxide gel will not be sensitive to abrasion in bioreactors, even in the presence of salt.

On the other hand, Figure 4 shows that the resistance to compression of the alginate gels clearly decreased with the number of oscillations, like found by Vogelsang et al. (submitted). The oscillations for the alginate gels had even to be stopped before the end of the experiment (1000 oscillations) was reached because the gels were too much deformed, which confirmed that fatigue. In the presence of salt, this already happened after 30 oscillations while without salt around 350 oscillations could be done. These results confirm that alginate gel is sensitive to abrasion in bioreactors (Muscat et al, 1996), especially in the presence of salt.

In addition, Figure 4 shows that salt also caused a lower resistance to compression of the alginate gel confirming the lowered strength of the gel at high salt concentration, like in the soy-sauce medium.

Our results of the oscillation compression experiment were in agreement with the conclusion of Leenen et al. (1996) that'elastic'gels like polyethylene-oxide will accommodate stresses better and for that reason, be relatively insensitive to abrasion than more'viscous'gels like alginate and carrageenan.

For this reason, the new polyethylene-oxide gel will be more suitable than alginate for use in a continuous process for a long period of time, especially in the presence of high salt concentrations. In our laboratory,

the new polyethylene-oxide gel has already been successfully applied for the continuous processes with immobilized soy-sauce yeasts in the presence of high salt concentrations.

Example 12 Maximum relative deformation at fracture Finally, a large-deformation experiment was done with the new polyethylene-oxide gel in order to determine the maximum relative deformation at fracture.

This was done by measuring the storage modulus as a function of the relative shear deformation. The storage modulus, which is a measure of the elastic energy stored in the material (Whorlow, 1992), will increase during a large deformation until fracture occurs. After that, the storage modulus will decrease.

In Figure 5 the effect of deformation on the storage modulus of polyethylene-oxide gel can be seen.

This figure shows that the storage modulus kept increasing until a large relative deformation of 1.0.

This means that fracture did not occur and that the new polyethylene-oxide gel, like rubber and gelatine but unlike alginate gel, is very flexible.

Example 13 Yeast immobilization during microsphere preparation For the immobilisation of the yeast in a microsphere system the following polymer solutions were prepared: Carboxymethyl cellulose (CMC) solution: low viscosity carboxymethyl cellulose sodium salt (Fluka, Buchs Switzerland) was dissolved in phosphate buffered saline (PBS, NPBI, Emmercompascuum, Netherlands) to a concentration of 5 mg/ml. After complete dissolving of the CMC, N-hydroxy succinimide was added to this solution to a concentration of 1 mg/ml.

Polyethylene imine-polyethylene oxide-polyethylene imine triblock copolymer (PEI-PEO-PEI) solution: PEI-PEO-PEO

with respective block lengths of 600,20. 000 and 600 was dissolved in PBS to a final concentration of 100 mg/ml.

After preparation of the polymer solutions a Z. rouxii suspension in PBS (7.3 ml) was added to the PEI- PEO-PEI solution (8.6 ml) and the resulting suspension was gently mixed to obtain a homogeneous suspension.

Subsequently the carboxylic acid groups of the CMC were (partially) activated by the addition of 2.3 ml N- (3- dimethylaminopropyl)-N'-ethylcarbodiimide hydrochloride (EDC) in water (100 mg/ml) to 82 ml of the CMC solution.

The solution was gently mixed for 5 minutes and then added to the yeast suspension which also contained the PEI-PEO-PEI polymer. After gently mixing of the resulting suspension, this suspension was poured into a baffeled reaction vessel containing 700 ml n-decane which was stirred at 800 RPM by a mixer equipped with an axial flow impeller.

After 1 hour of mixing during which gelation of the suspension had occurred, 200 ml PBS was added to the reaction vessel en mixing was continued for several minutes. Then the decane was decanted and the spheres were further washed with PBS and sieved over a 212 um sieve. Spheres with a diameter that was larger than 212 um were isolated and used for fermentation experiments.

Example 14 Yeast immobilization and subsequent particle preparation by blendinq For the immobilisation of a yeast, the same polymer solutions were prepared as mentioned in the example above. After preparation of the polymer solutions a yeast suspension in PBS (11.0 ml) was added to the PEI- PEO-PEI solution (12.9 ml) and the resulting suspension was gently mixed to obtain a homogeneous suspension.

Subsequently the carboxylic acid groups of the CMC were (partially) activated by the addition of 3.5 ml N- (3- dimethylaminopropyl)-N'-ethylcarbodiimide hydrochloride (EDC) in water (100 mg/ml) to 123 ml of the CMC solution.

The solution was gently mixed for 5 minutes and then added to the yeast suspension which also contained the PEI-PEO-PEI polymer. After gently mixing of the resulting suspension, this suspension was poured into a petri dish and the suspension was allowed to gelate. After 1 hour 200 ml PBS was poured on the gel and after another hour the gel and PBS were transferred to a beaker and the gel was processed into particles using a homogenizer/handblender (Braun MR 555 M CA) operated with the blade insert. Small particles were removed by sieving over a 212 um sieve and the remaining particles were isolated and used for fermentation experiments.

Example 15 Ethanol production with immobilised yeast.

A commercial grade wine yeast (S. ellipsoidus, 0.5 gram) was added to 250 ml of a phosphate buffered glucose solution (80 mg/ml) that also contained some lemon juice and vitamin B1, and was incubated at 20°C.

After 48 hrs part of the yeast was immobilised. First polymer solutions as described in example (see example 'Yeast immobilisation during microsphere preparation) were prepared and 10 ml of the yeast suspension was added to 11,7 ml of the PEI-PEO-PEI solution and the resulting suspension was gently mixed to obtain a homogeneous suspension. Subsequently carboxylic acid groups of the CMC were activated by the addition of 3,2 ml N- (3- dimethylaminopropyl)-N'-ethylcarbodiimide hydrochloride (EDC) in water (100 mg/ml) to 112 ml of the CMC solution.

The solution was gently mixed for 5 minutes and then added to the yeast suspension which also contained the PEI-PEO-PEI polymer. After gently mixing of the resulting suspension, this suspension was poured into a petri dish and was allowed to gelate. After 1 hr 200 ml of the above described glucose solution was poured on the gel and after another hour the gel and glucose solution were transferred to a beaker and the gel was processed into particles using a homogenizer/handblender (Braun MR 555 M

CA) operated with the blade insert. This immobilised yeast suspension was transferred to a fermentor which contained 1 litre of a glucose solution as described above, however now with a glucose concentration of 200 mg/ml. The fermentor was incubated at 20°C and within 48 hrs C02 production was observed as gas bubbles leaving the fermentor through the water lock. After 21 days the fermentation was stopped by sieving and filtration of the suspension to remove non soluble particles. Subsequently ethanol was removed from the remaining solution by distillation and was identifie by its smell.

The invention is not limited to the above description; the requested rights are rather determined by the following claims.