FR2628751A1 - Continuous prodn. of L-alpha-amino acid from alpha-keto-acid - by reaction with ammonium ion and reducing agent in presence of Bacillus cells contg. specific enzymes and NAD coenzyme - Google Patents

Abstract

Continuous prodn. of L-alpha-amino acids (I) comprises (1) culturing a Bacillus strain so that the cells contain amino acid dehydrogenase and an enzyme able to catalyse regeneration of NAD; (2) keeping these cells in a reaction zone, then (3) supplying this zone with at least one alpha-keto acid (II), or salt, an NH4 ion source and reducing agent, and reaction with the cells in presence of NAD-type coenzyme (oxidised or reduced). The new feature is that (I)-contg. effluent is withdrawn and the (II) supply rate, (II) concn. and rate of withdrawal are controlled so that, in the reactin zone, the (I) concn. is 1-200 mM and the (II) concn. at least 150 mM. More specifically the cells are esp. of B. megaterium ATCC 39118 and the (II) feed rate is 0.01-25 (esp. 0.1-10) mmole/hr/l of reaction vol. USE/ADVANTAGE - Compared with known procedures, this method provides higher (I) yields for a given wt. of cells; does not require use of polymer-fixed coenzyme and uses only one type of cell culture. Long-term continuous operation is possible since coenzyme is not lost from the cells and is repeatedly reused. (I) are useful in the chemical and pharmaceutical industries.

Description

The present invention relates to a process for the continuous production of L amino acids by enzymatic reduction of corresponding ketoacids in the presence of a source of ammonium ions, a coenzyme of the nicotinamide adenine dinucleotide (NAD) type and a reducing agent. .

It has been described in the patent application EP-206904 a "batch" or "fed batch" embodiment of this reaction using, as sole source of enzyme, cells of a strain of Bacillus, these cells possessing both an amino acid dehydrogenase activity and a glucose dehydrogenase activity considered essential.

By operation in "batch" means a completely discontinuous operation: the reactants are introduced entirely at the start. In a "fed batch" operation, at least one of the reactants is introduced entirely at the start, the other (or the other) reagents being introduced progressively, as the course of the reaction unfolds, without racking. In both cases, the medium remains confined in the reactor.

After each operation according to EP-206904, it could be thought that the cells could be reused as the source of the two enzymes in a subsequent operation, but it would be necessary to add coenzyme, which is an expensive material, in this fed batch operation.

Furthermore, because of the solubility of NAD in the medium, it was not possible to envisage a continuous operation except to continually give NAD, which proved a priori uneconomical.

In addition, a continuous enzymatic process for the synthesis of L-α-amino acid from α'-keto acid in a membrane reactor with coenzyme nicotinamide adenine dinucleotide in the form of a high-weight derivative has been described in US Pat. soluble molecule in an aqueous medium and with two purified enzymes, an L-amino acid dehydrogenase for the reduction of o'-keto acid and a formate dehydrogenase for the regeneration of coenzyme.

This method has several disadvantages - two intracellular enzymes from two
different microorganisms Candida boidinii or Pseudomonas
oxalaticus for formate dehydrogenase and Bacillus subtilis for
L-amino acid dehydrogenase, which involves two cultures
different and breaking the cells to get the extracts
containing the intracellular enzymes and two stages of
purification of these enzymes.

- On the other hand, it is necessary to proceed with the grafting of coenzyme
on the polymer (polyoxyethylene), a step whose performance is close to
about 50%, this grafted NAD having subsequently a less reactivity
good as the native NAD.

The method of the present invention, which overcomes the aforementioned drawbacks, allows to obtain an increased yield of L-amino acids for the same amount of Bacillus cells, by a continuous process while simultaneously catalyzing the regeneration of nicotinamide adenine dinucleotide under reduced form necessary for the reaction.

Another object is the use of Bacillus cells in a continuous process after these cells have been pretreated ex situ in the presence of coenzyme.

More specifically, the invention relates to a process for the continuous production of L-Wamino acids in which a strain of Bacillus is cultured under conditions such that cells containing an enzymatic composition comprising dehydrogenase amino acid are produced. and an enzyme capable of catalyzing the regeneration of nicotinamide adenine dinucleotide, said cells are confined in a reaction zone, and a feed medium comprising at least one ketoacid or at least one of its members is subjected to said reaction zone; salts, a source of ammonium ions and a reducing agent to the action of said cells in the presence of a nicotinamide adenine dinucleotide-type coenzyme in reduced or oxidized form.This process is characterized in that a tapping of a effluent containing said LS-amino acid product, the flow rate of the O-keto acid of said feed medium is adjusted in the reaction zone, the concentration of keto acid in the feed medium and the withdrawal rate of said effluent so as to maintain an L-amino acid concentration of between 1 and 200 mmol. per liter in the reaction zone and so as to maintain a concentration of ε-keto acid in the reaction zone at most equal to 150 mmol per liter.

A higher productivity of L-o <amino acids is obtained when the concentration of L-amino acids in the reaction zone (reaction medium) is maintained at a value of between 50 and 200 mmol per liter.

It has been found that when the concentration of L-gamino acids exceeds 200 mmol / l in the reaction zone and when the concentration of α-keto acid in the reaction zone exceeds 150 mmol / l, there is a drop in the productivity in L-daminoacids, probably due to a partial or complete inhibition of enzymatic activities by the excessive presence of substrate or product in the reaction medium.

One of the advantages of the process according to the invention relates to its lifetime, which can be concretized by the notion of number of coenzyme recycling, which is the ratio of the number of moles of L-Kamino acid form per mole of NAD introduced.

The operation of the continuous process makes it possible to substantially increase the productivity of the process while using a quantity of NAD less than in "fed batch" or "batch".

We can introduce 1. ' -acetoacid and withdraw L-X amino acid continuously or periodically.

The feed medium may be composed of a single, generally aqueous solution, containing all three substrates of the reaction (α-ketoacid, reductor and ammonium ions), or by mixing two or three solutions depending on whether the one of the substrates is alone or mixed with one of the other two.

The product of culture of a strain of
Bacillus megaterium and even more preferably the culture product of Bacillus megaterium strain ATCC 39118 because of the particularly high yields of L-α-amino acids.

This strain is a sporulation mutant of the strain of the Bacillus megaterium IFP 188 laboratory collection described in patent FR 2526442.

According to a preferred embodiment of the method, the cells may be subjected to pretreatment ex situ comprising incubation of said cells in the presence of the coenzyme in oxidized or reduced form under conditions such that said cells contain a substantial amount of said coenzyme in intracellular form, sufficient for the production of amino acids. These pre-treated cells are then subjected to the actual bioconversion, thus providing the reaction medium with the coenzyme necessary for the reaction. By a substantial amount of coenzyme is meant a quantity of coenzyme retained for a long time inside the cell in order to be able to achieve , especially continuously, product synthesis reactions from corresponding substrate without addition of coenzyme, for example from 0.01 to 2 mmol of intracellular NAD per gram of cell dry weight.

The cell thus pretreated contains in its interior, the two enzymes, one serving for the conversion of the precursor the other for the simultaneous regeneration of the coenzyme. It also contains, and this is unexpected, the coenzyme which once introduced, and surprisingly, does not stand out substantially. The cell can then function as a membrane reactor containing the enzymes necessary for the continuous reaction and the coenzyme and not allowing them to escape while allowing the free diffusion of the substrates and products of the two reactions in the cell.

The process is then made even more economical because we can recover the unincorporated NAD in the cells during incubation and achieve a better recovery of coenzyme (number of recycling).

It is then possible to operate continuously for a substantially long period of time without having to add enzyme and coenzyme as long as it is known to keep these cells confined in a reaction space, which can be achieved by using either a membrane reactor with free microorganisms, or a membrane reactor or a column with cells fixed or adsorbed according to techniques well known to those skilled in the art in particular - the adsorption techniques microorganisms on supports such
only kieselguhr, crushed brick, exchange resins
ions, glass beads, wood chips, ceramics,
etc., - the techniques of inclusion of microorganisms in gels of
polymers such as agar, calcium alginate, carrageenan,
polyacrylamide, etc., - the coupling techniques of microorganisms on different
supports, silica beads, sintered glasses, by covalent bond
with agents such as glutaraldehyde, isocyanates, etc.

According to a first embodiment, it is possible, after the cell culture harvesting step, to separate the cells from the fermentation must and then to proceed with the incubation step with the coenzyme and then to harvest these cells again. impregnated with coenzyme. This makes it possible to adjust substantially, following the first harvest, the amount of coenzyme with respect to the cells and following the second harvest, firstly to adjust the quantity of impregnated cells which will be used for the conversion, and on the other hand to recover the excess coenzyme in the incubation solution.

According to another embodiment, it is possible, after the culturing step, to bring the coenzyme into contact with the fermentation must without first harvesting the cells, these cells being then harvested after incubation in the presence of coenzyme. This avoids the first cell harvesting step following the culture.

According to yet another embodiment, after the culture step, it is also possible to incubate the coenzyme with the fermentation must without harvesting the cells, and to use this solution directly for the desired conversion.

The method of the invention comprises - the cultivation of microorganisms by fermentation, the conditions of
culture being such that the microbial cells contain the
desired enzymatic activities and permeabilization
spontaneous membrane without the need for treatment of
permeabilization of the cells. These conditions are described in the
patent application published under No. EP 206904.

The state of permeabilization of the cells is largely conditioned by the culture conditions, in particular the composition of the medium, the temperature and especially the culture time. These culture conditions are generally as follows: temperature at 20 ° C., pH 5.5 at 8.0.

An excellent level of activity was obtained at a temperature between 27 and 320C and in a pH range of 6.5 to 7.5, for example an L-amino acid dehydrogenase and glucose dehydrogenase activity in a ratio of 0.1: 1 to 10: 1. The culture times are generally at least 15 hours, preferably 20 to 50 hours and preferably 25 to 40 hours. The medium can be agitated and aerated so as to have a dissolved oxygen level which is never zero.

From the cultures of microorganisms, the cells can advantageously be harvested by centrifugation, filtration or decantation. For verification, intracellular enzymatic activities are measured from an extract at optimum temperatures, ionic strength and pH for each enzyme with saturating substrate concentrations.

The cells thus harvested can be added to the reaction medium with the substrate necessary for the bioconversion. The NAD + or NADH coenzyme is introduced into this reaction medium at a concentration of at least 0.05 mmol / ml of solution of the reaction medium while the Bacillus cells are generally introduced into the reaction medium at a rate of 1 to 200 grams. , counted in dry weight, per liter of reaction mixture and preferably in a proportion of 2.5 to 50 g / l. In contrast, according to the preferred embodiment described above, the cells pretreated in the presence of coenzyme can be advantageously used as such for the bioconversion, without addition thereafter of coenzyme in the reaction medium.

The impregnation of the cells with coenzyme is done by simply contacting the cells and NAD or NADH equivalent. The cells prepared and harvested under the conditions described above are suspended in a buffered solution at a pH of between 5 and 10, preferably between 7.5 and 9, which contains coenzyme, in a proportion of 0.5 to 50. g dry cell weight per mmol of coenzyme, preferably from 1 to 50 g / mmol and at a temperature between 4 and 700C, preferably between 10 and 400C. The incubation time generally varies from 0.5 to 100 hours.

After this incubation, the solution may advantageously be centrifuged or filtered in order to recover, on the one hand, the cells impregnated with the coenzyme and, on the other hand, the coenzyme remaining in the supernatant or the filtrate, a coenzyme that can be reused.

The quantity of pretreated cells involved during the bioconversion is between 1 g of dry weight of cells per liter and 200 g / l of solution necessary for the bioconversion, preferably between 2.5 g / l and 50 g / l .

The actual synthesis is generally carried out in a reactor maintained at a temperature of 10 to 60 ° C, preferably 20 to 450 ° C, whether the cells are pre-treated or not.

The pH is chosen according to the optimum enzyme activity and stability, and the synthesis is generally in a buffered medium. It may be necessary to regulate the pH by the addition of acid or base.

A pH between 6.5 and 8.5 is particularly favorable for the stability of the enzymes used and NAD. The choice of pH is indeed important for the latter compound whose reduced form is unstable in acid medium and the oxidized form unstable in alkaline medium. The presence in the middle of a buffer containing ions
NH4 + is generally recommended because the ammonium ion is one of the substrates of the amino acid synthesis reaction. However, it is quite possible to use a buffer based for example on a soluble salt of phosphoric acid which is useful for maintaining the pH and stabilizing the enzymes. The pH is maintained at the chosen value by preferably adding ammonia. This one intervenes at the same time to maintain the pH and as reagent of the reductive amination.

A preferred procedure is described below: the selected elements, namely the buffer and the Bacillus megaterium cells which have been incubated in the presence of coenzyme and which have been immobilized beforehand, are introduced into the reactor. Only part or all of the reducing agent, for example glucose, necessary for the reaction is introduced at the beginning of the reaction. In some cases, it is preferred to avoid the presence in the reactor of too high concentrations of glucose which may have an inhibitory action on the enzymatic activity L-amino acid dehydrogenase. In the same way, it is preferred to avoid the presence in the reactor of high end-keto acid concentrations which may have an inhibitory action on the enzymatic activity L-amino acid dehydrogenase. Therefore, it is often advantageous to inject gradually and Preferably, these two compounds are continuously in the reactor in the form of an aqueous solution whose amount of reducing agent (glucose, for example) is 1 to 2 moles per 1 mole of ketoacid. The -keto acid can be introduced in the form of acid or of one of its salts, in particular a salt of Na
K +, NH4 +, Li +. Its molar concentration is chosen so as to avoid excessive inhibition by the amino acid formed knowing that the reaction is stoichiometric. The intake flow rate is preferably adjusted so that the limiting parameter of the reaction is the input of the ketoacid, the latter compound then being converted into L-amino acid substantially as it is introduced.

It may for example be between 0.01 and 25 mmol.h-1. (liter of reaction volume) -1 and preferably between 0.1 and 10 mmol.h-1.

(liter of reaction volume) -l so as to obtain a sufficiently high productivity and not to have an injection speed too fast which would cause a build up of substrate and by the same, an almost total stop of the reaction from a ketoacid concentration in the reaction zone of greater than 150 mmol per liter.

It is therefore maintained at a K-ketoacid concentration at most equal to 150 mmol / l of reaction volume, advantageously between
0 and 100 mmol / l of reaction volume and preferably of between 0 and 50 mmol / l and a concentration of L-vinyl amino acid of between 1 and 200 mmol / l of reaction volume. It is possible to modify the feed rate or its composition as a function of the evolution of the concentrations of products and substrates in the reaction medium. Thus, in the case of feed solutions with a high concentration of ketoacid, it is often necessary to lowering the intake flow rate or diluting the feed solution so as to avoid excessive inhibitions, that is to say, to maintain concentrations of -ceto acid and amino acid as defined above. In general, the withdrawal rate will be adjusted to be substantially equal to the feed rate to maintain a reaction volume compatible with the reactor volume.

It is possible to operate continuously in a membrane reactor.

The membrane may be an ultrafiltration membrane or microfiltration, as long as the pore diameter is less than 0.8 micrometer. Membranes with a pore diameter of between 0.2 and 0.8 microns which clog less easily than the ultrafiltration membranes will preferably be used. In general, the feed medium comprises glucose and keto acid or its salt. The respective concentration of the two substrates is from 1 to 2 moles of glucose per 1 mole of -acetoacid. In general, the o'-keto acid concentration in the feed medium is chosen such that if the reaction yield equals 100%, it remains below the inhibition threshold by the corresponding amino acid. in ammonium ions is generally ensured by the regulation of pH by ammonia.

If it is desired to work continuously with immobilized or fixed cells, it is easier to confine the catalyst inside the reactor (conventional filtration membranes, grids) and it is quite possible to operate in a column reactor.

According to another embodiment, it is possible to operate first in fed batch mode at the beginning of the synthesis by supplying at least once a cr-keto acid or one of its salts, so as to progressively bring the reactor into operation. until the concentration of between 1 and 200 mmol is reached in L-amino acid. The medium is then withdrawn according to the process of the invention.

This system is preferred in the case where cells are used without pretreatment, the duration of the fedbatch allowing the cells to be loaded with coenzyme while bringing the medium to the desired concentration. This avoids some of the losses of coenzyme that are inevitable in the case where one operates with cells that have not been incubated in the coenzyme prior to conversion.

The L-amino acid formed during the reaction can be assayed by techniques known to those skilled in the art, for example by high performance liquid chromatography after derivatization or enzymatically. Similarly, it is possible to determine the levels of
NAD and NADH by enzymatic methods.

The products of the reactions are harvested then extracted and purified according to conventional techniques which depend on the type of product manufactured. For example, the L-amino acid can be extracted by passage over a cation exchange resin, and then the eluate can be evaporated in vacuo.

Bacillus, and in particular Bacillus megaterium ATCC 39118, is capable of producing both L- <amino acid dehydrogenase and glucose dehydrogenase for the regeneration of NADH. With such a system, it is possible to produce either L-alanine from pyruvate, L-serine from ss-hydroxypyruvate, L-valine from l-ketoisovalerate, L-valine from leucine from 1'-ketoisocaproate, and L-isoleucine from α-ketossmethylvalerate.

The products thus synthesized are particularly interesting for the chemical and pharmaceutical industries.

The following examples which illustrate the invention are given without limitation.

EXAMPLE 1 Synthesis of L-valine continuously by cells
impregnated with NAD
The strain Bacillus megaterium ATCC 39118 is grown in a vial of
Fernbach in a medium with the following composition - trypsic casein hydrolyzate 17 g - papalnic soy peptone 3 g - sodium chloride 5 g - bipotassium phosphate 2.5 g - glucose 2.5 g - distilled water 1 liter
This medium is sterilized before inoculation, 30 minutes at 1150C. The culture is carried out at 300C, pH = 7.0, for 30 hours, the flasks being placed on a reciprocating stirring table.

Then the cells are harvested by centrifugation. From 300 ml of culture, 4. 5 g of wet cells (ie about 1.0 g dry weight) are collected and washed once with 0.1 M phosphate buffer, pH 7, and placed in suspended in 50 ml of NH4Cl buffer
NH 4 OH, 1 M, pH 8.2, containing 6 mmol of NAD. The enzymatic activity of the cells can be verified by taking a sample of the harvested cells; these cells are broken during ultrasonic treatment to determine the enzymatic activities of amino acid dehydrogenases and glucose dehydrogenase.

The enzymatic activity of valine dehydrogenase is found to be 0.47 units per mg of protein extract, that of glucose dehydrogenase is 0.82 units per mg of protein extract. An enzymatic unit is defined as the amount of enzymes allowing the transformation of one micromole of substrate (glucose in the case of glucose dehydrogenase, cr-ketoisovaleric acid or one of its salts in the case of valine dehydrogenase) per minute under the conditions of the test. We can then estimate the enzymatic activity per gram of wet cells which is in this case 33.8 U / g for valine dehydrogenase and 50.9 U / g in the case of glucose dehydrogenase.

The cells suspended in the NAD + solution are left under gentle stirring for 24 hours at room temperature and are then recovered by centrifugation. In the supernatant no glucose dehydrogenase or valine dehydrogenase activity was detected and 3.9 mmol / l of NAD + was found, indicating that 2.1 mmol / l was passed into the cells.

The entirety of the recovered cells is used for the synthesis of
L-valine which is carried out in a reactor containing 50 ml of initial reaction medium. This medium consists of 4.5 g of wet cells (impregnated in such a way that 2.1 mmol of NAD per liter of medium) are suspended in 50 ml of 1M NH4C1-NH40H buffer, pH 8.2. containing 0.6 mole l-1 of glucose. The pH is adjusted to 8.2 with 2N ammonia solution. The temperature is maintained at 300C and the medium is stirred moderately.

The reactor is an AMICON ultrafiltration cell with a volume of 200 ml, equipped here with a microfiltration membrane with a diameter of 62 mm, with a porosity of 0.2 microns, the outlet of the effluent being effected downstream of the membrane. .

During the first 24 hours, a 0.3 mole.l solution of sodium cetovalerate is injected via a peristaltic pump at a flow rate of 0.5 ml.h 1 without being drawn off. medium, which corresponds to a feed called "fed batch". Subsequently, it operates continuously by withdrawing an effluent downstream of the membrane using a peristaltic pump. The feed solution is then composed of sodium i-ketoisovalerate at 35 mmol / l and glucose at 100 mmol / l. The feed and withdrawal rates are equal to 3 ml.h1.

At the end of the first 24 hours corresponding to the "fed batch", the volume of medium in the reactor is 68 ml, the concentration of
L-valine is 44 mmol.l, i.e., a molar conversion level of 80%. The molar conversion ratio is defined as the ratio of the amount of amino acid formed to the amount of o'-keto acid injected. Over the next 24 hours (continuous), the L-valine concentration rose to 38.5 mmol / l, which corresponds to a molar conversion rate of 94% relative to the injected precursor. No trace of NAD or NADH was found in the bioconversion supernatants, which shows that the coenzyme remains intracellular. On the other hand, no lysis of the cells was observed. This shows that the NAD which has passed inside the cells, does not stand out and that it functions according to the L-amino acid synthesis scheme by the amino acid dehydrogenase with regeneration of coenzyme by glucose dehydrogenase.

The recovery of L-valine from the reactor effluent is carried out in the following manner the reaction medium is poured at the top of an Amberlite ER 151 cation exchange resin column previously formed by washing with water. 1N hydrochloric acid followed by washing with water. After introduction of the sample, the column is washed again with water and then eluted with 2N ammonia. After the passage of a volume of the elution solution corresponding to the "dead" volume of the column, the eluted fractions are started to recover until their total volume is 1.5 times the dead volume of the column. The elut is then evaporated under vacuum, redissolved in water and evaporated again. The purification yield is 98%.

Example 2: Continuous synthesis of L-valine with different
concentrations of α-ketoisovalerate in the feed solution
The Bacillus megaterium ATCC 39118 cells are cultured, harvested and pretreated before conversion under conditions identical to those of Example 1. The first 24 hours of synthesis are carried out in "fed batch" under the same conditions as those of the example 1. Then, the operation is continuous (feeding and racking) for ten days, the installation used for the conversion being the same as that of Example 1. The feed medium then contains 600 mmol.l-1 of glucose and varying concentrations of sodium-ketoisovalerate (Table 1) .The volume of reaction medium is kept constant at 70 ml by adjusting the feed and withdrawal pumps at the same rate. The concentration of L-valine in the reaction volume at the end of the fed batch is about 50 mmol.l. In the case of feeds at 35 mmol / l and 100 mmol / l, the concentrations of
L-valine pass gradually from the concentration obtained at the end of fedbatch to the indicated concentration. In the case of feeding at 200 mmol.l, the concentration of L-valine reaches 125 mmol.l after 72 hours of continuous and then decreases. progressively up to 95 mmol / l, the enzymes operating slower initially because of inhibition by L-valine and then also because of the accumulation of ketoisovalerate in the feed medium. These effects are even more marked in the case of feed solutions with 250 mmol of ketoacid where the residual concentration of ketoisovalerates causes a premature slowing down of the reaction From the moment when the residual ketoacid concentration reaches about 130 mmol.l, the conversion rate drops sharply (see Table 1, column for the last 50 hours).

In the case where the concentration of ketoisovalerate in the medium
-1 feed is 400 mmol.l, the L-valine content reaches a TABLE 1

Concentration <SEP> Flow rate <SEP> Concentration <SEP> Concentration <SEP> Rate <SEP> of <SEP> conversion <SEP> Productivity <SEP> Number <SEP> of
<tb> Acetoisovialerate <SEP> feeding <SEP> in <SEP> L-valine <SEP> residual <SEP> in <SEP> molar <SEP> in <SEP> L-valine <SEP> retraining
<tb> in <SEP><SEP> medium <SEP> in <SEP><SEP> milleus <SEP> acetolsovalerate <SEP> (calculated <SEP> on <SEP> of <SEP> coenzyme
<tb> Supply <SEP> Reaction <SEP> in <SEP><SEP> Middle <SEP> to <SEP> Tip <SEP> of <SEP> in <SEP><SEP> Course <SEP><SEP> 300 <SEP> hours
<tb> at <SEP> last <SEP> of <SEP> 300 <SEP> h <SEP> reaction <SEP> 300 <SEP> h <SEP> 50 <SEP> last <SEP> of <SEP> continuous)
<tb> of <SEP> continuous <SEP> (at <SEP> last <SEP> of <SEP> 300 <SEP> h) <SEP> hours
<tb> mM / 1 <SEP> m1 / h <SEP> mM / 1 <SEP> mM / 1 <SEP>% <SEP>% <SEP> mM / 1.h
<tb> 35 <SEP> 3 <SEP> 33 <SEP> 2 <SEP> 94 <SEP> 94 <SEP> 1.4 <SEP> 305
<tb> 100 <SEP> 3 <SEP> 90 <SEP> 5 <SEP> 94 <SEP> 93 <SEP> 2,6 <SEP> 574
<tb> 200 <SEP> 2 <SEP> 95 <SEP> 89 <SEP> 52 <SEP> 48 <SEP> 2.7 <SEP> 606
<tb> 250 <SEP> 2 <SEP> 50 <SEP> 180 <SEP> 22 <SEP> 6.5 <SEP> 1.4 <SEP> 320
<tb> 400 <SEP> 2 <SEP> 20 <SEP> 360 <SEP> 5 <SEP> 0 <SEP> 0.6 <SEP> 128
<tb> TABLE 2

Concentration <SEP> in <SEP> Throughput <SEP> Concentration <SEP> Concentation <SEP> Rate <SEP> of <SEP> conversion <SEP> Productivity <SEP> Number <SEP> of
<tb> asetoisovali6rate <SEP> feeding <SEP> in <SEP> L-valine <SEP> residual <SEP> in <SEP> molar <SEP> in <SEP> L-valine <SEP> recycling
<tb> in <SEP> the <SEP> middle <SEP> in <SEP> the <SEP> middle <SEP>&alpha; ketoisovaliérste <SEP> of <SEP> NAD
<tb> Supply <SEP> Reaction <SEP> in <SEP><SEP> Middle <SEP> to the <SEP><SEP> End of <SEP><SEP><SEP> Courses
<tb> at <SEP> last <SEP> of <SEP> 400 <SEP> h <SEP> reaction <SEP> 400 <SEP> h <SEP> 50 <SEP> last
<tb> of <SEP> continuous <SEP> (at <SEP> last <SEP> of <SEP> 400 <SEP> h) <SEP> hours
<tb> mM / 1 <SEP> m1 / h <SEP> mM / 1 <SEP> mM / 1 <SEP>% <SEP>% <SEP> mM / 1.h
<tb> 35 <SEP> 1.0 <SEP> 30 <SEP> 3.5 <SEP> 90 <SEP> 88 <SEP> 1.80 <SEP> 357
<tb> 100 <SEP> 2.5 <SEP> 90 <SEP> 5.0 <SEP> 94 <SEP> 92 <SEP> 4.50 <SEP> 900
<tb> 200 <SEP> 0.5 <SEP> 150 <SEP> 10.0 <SEP> 94 <SEP> 93 <SEP> 1.50 <SEP> 355
<tb> 250 <SEP> 0.4 <SEP> 170 <SEP> 20.0 <SEP> 89 <SEP> 80 <SE> 1.36 <SEP> 340
<tb> 400 <SEP> 0.2 <SEP> 180 <SEP> 70.0 <SEP> 72 <SEP> 65 <SEP> 0.72 <SEP> 220
<tb> maximum equal to 125 mmol.l after 12 hours of continuous. This value then decreases, any synthesis stopping after 35 hours of continuous, the L-valine previously formed then being progressively leached; here again, the inhibitory effect of L-valine decreases the activity of the enzymes, which causes a -cétoacide accumulation which generates a total inhibition of enzymatic activities.

In a second series of experiments, it was sought to adjust the feed rate so as to avoid inhibitions related to substrate accumulation (Table 2). The flow rate is adjusted to a value such that there is virtually no residual d-keto acid in the reaction medium. The values indicated correspond to those obtained after 400 hours of continuous operation, the reaction volume being kept constant at 50 ml and the continuous being started without preliminary batch fed. It is found that, although the conversion efficiency remains high, the productivity drops substantially as well as the number of NAD recycles, this being mainly due to the high levels of L-valine which are close to the inhibitory concentration of the process according to the invention. .

Example 3: Continuous synthesis of L-valine by cells
impregnated with NAD (without pre-fed batch)
The conditions of culture, harvest and incubation in the presence of NAD + are identical to those of Example I. The cells are converted in the same manner as in Example 1, except for the fed batch period (the conversion is started directly continuous). After 24 hours, the conversion is 85% and no coenzyme is found in the effluent. After 200 hours of continuous, the conversion rate is equal to 80%.

Example 4: Continuous synthesis of L-valine by whole cells
not pre-treated with a pre-fed batch
The culture conditions of Bacillus megaterium and harvest are identical to those of Example 1. The cells are not subjected to incubation in the presence of NAD + before being used in bioconversion. The activity of the harvested cells is similar to that of Example 1 and the reaction medium is identical to that of Example 1 except that it contains NAD + at a concentration of 3 mmol / l. The conversion is then carried out under conditions identical to those of Example 1 (24 hours of fed batch, then continuous with a feed medium free of coenzyme).

After the 24 hours of fedbatch, there is an extracellular concentration of NAD + which is only 1 mmol.l. NAD + is then diluted and leached into the effluent. The conversion efficiency is 90% during the fed batch and 80% during the 24 hours of the continuous process. It is 72% after 200 hours of continuous operation.

Example 5: Continuous synthesis of L-valine by whole cells
not pre-treated and without pre-fed batch
The conditions of culture and harvest as well as the initial medium are identical to those of Example 3. The conversion is started without batch or fed batch initial.- The feed medium contains 50 mM.l1-kétoisovalerate and 100 mmol.l -1 glucose. Injection and withdrawal rates are equal to 3 ml.h1. After 24 hours of continuous, the concentration of L-valine in the reactor corresponds to a conversion rate equal to 70%. In the effluent, there is a quantity of NAD corresponding to an extracellular content of 1.5 mmole.1. There is therefore a loss of coenzyme greater than in Example 4. The conversion rate after 200 hours of continuous operation is 60%.

Example 6 Synthesis of L-Valine by Whole Cells Without
preprocessing
The procedure is as in Example 4 and the reaction medium contains, as in Example 4, NAD + at a concentration of 3 mmol / l. The substrate concentration is varied as in the first series of experiments of Example 2. The results, presented in Table 3, are substantially identical to those of Example 2, but the number of coenzyme recycles is inferior.

Example 7: L-valine synthesis assay by cells without bet
in contact with these cells with NAD +
The conditions for culturing and harvesting the cells are identical to those described in Example 1. The valine dehydrogenase and glucose dehydrogenase activities of the cells are respectively 32.1 U per g of wet cell and 79.1 U per g of wet cell. All the cells harvested from a Fernbach flask (2.88 g wet) are integrated into the usual reaction medium (buffer NH40H-NH4Cl 1M, pH 8.2, glucose 0.6 mole.l1) without implementation. prior incubation with NAD. The conversion is carried out as in Example 1. The production of L-valine is then only
-1 2 mmol.l in 24 hours, a conversion rate virtually zero.

This shows that the initial intracellular NAD (NADH) is not sufficient to ensure the functioning of the enzymes involved in the synthesis of amino acids and that there is good passage of the coenzyme inside the cell and retention of it .

TABLE 3

Concentration <SEP> Flow rate <SEP> Concentration <SEP> Concentration <SEP> Rate <SEP> of <SEP> conversion <SEP> Productivity <SEP> in <SEP> Number <SEP> of
<tb><SEP> Acetoisovalerate <SEP> Feed <SEP> into <SEP> L-Valine <SEP> Residual <SEP> into <SEP> Molar <SEP> L-Valine <SEP> Recycling
<tb> in <SEP> the <SEP> medium <SEP> in <SEP> the <SEP> medium <SEP>&alpha;; ketoisovalerate <SEP> (calculated <SEP> on <SEP> of <SEP> coenzymes
<tb> Supply <SEP> Reaction <SEP> in <SEP><SEP> Medium <SEP> at <SEP> Tip <SEP> of <SEP> at <SEP><SEP> Course <SEP><SEP> 300 <SEP> hours
<tb> (at <SEP> last <SEP> of <SEP> 300 <SEP> h <SEP> reaction <SEP> 300 <SEP> h <SEP> 50 <SEP> last <SEP> of <SEP> continuous)
<tb> of <SEP> continuous) <SEP> (at <SEP> last <SEP> of <SEP> 300 <SEP> h) <SEP> hours
<tb> mM / 1 <SEP> m1 / h <SEP> mM / 1 <SEP> mM / 1 <SEP>% <SEP>% <SEP> mM / 1.h
<tb> 100 <SEP> 2 <SEP> 89 <SEP> 5 <SEP> 94 <SEP> 93 <SEP> 2.5 <SEP> 402
<tb> 200 <SEP> 2 <SEP> 94 <SEP> 90 <SEP> 51 <SEP> 46 <SEP> 2.6 <SEP> 420
<tb> 250 <SEP> 2 <SEP> 48 <SEP> 182 <SEP> 21 <SEP> 5 <SEP> 1.4 <SEP> 220
<tb> 400 <SEP> 2 <SEP> 20 <SEP> 360 <SEP> 5 <SEP> 0 <SEP> 0.5 <SEP> 90
<Tb>
Example 8 Synthesis of L-Valine by NAD Impregnated Cells
then immobilized
The conditions for culturing and harvesting the cells are identical to those described in Example 1. From 300 ml of culture medium, 2.96 g of cells (wet weight) having activities valine dehydrogenase and glucose dehydrogenase are collected. of 54 U and 102 U per gram of wet cell. These cells are stirred for 24 hours in 50 ml of 1 M NH4Cl-NH4OH buffer, pH = 8.2, containing 3 mmol / l NAD. The supernatant remaining after centrifugation of this solution is 2.48 mmol / l in NAD.

The cells thus prepared are immobilized in calcium alginate gel beads. The method consists of preparing a 20% cell suspension in 2% sodium alginate. The mixture is extruded drop by drop using a syringe pump into a stirred 0.1 M CaCl 2 solution. The beads are left in CaCl 2 for about 2 hours and then recovered by filtration on a filter crucible.

The L-valine synthesis is carried out under conditions identical to those described in Example 1 with the exception of the catalyst which here corresponds to the calcium alginate beads containing the NAD-impregnated cells. After 30 hours of continuous ketoacid injection, valine production is 10 mmol / l and the molar conversion ratio is 15% continuous. After 200 hours of continuous operation, this conversion rate is equal to 12%.

Example 9 Synthesis of L-valine by immobilized cells and then
impregnated with NAD
The conditions for culturing and harvesting the cells are identical to those of Example 1. Here 4 g of wet cells having 43.7 U valine dehydrogenase and 117.3 U glucose dehydrogenase per gram of wet cells are harvested. These cells are then immobilized in calcium alginate beads using the same procedure as described in Example 8.

After immobilization, the gel beads containing the
Bacillus megaterium are stirred in 50 ml of 1M NH4Cl-NH4OH buffer solution, pH = 8.2, containing 3 mmol of NAD + for 24 hours. The suspension is then filtered in order to recover the immobilized cells and then incubated in the presence of NAD +. The filtrate contains only 0.525 mmol / l of NAD +.

These beads are bioconverted using conditions identical to those described in Example 1 with the exception of the precursor concentration in the feed which is equal to 50 mmol.l. After 48 hours of conversion (24 h fed batch and 24 h continuous), the concentration of L-valine is 45 mmol.l, which corresponds to a molar conversion yield of 90% relative to the precursor injected during the continuous. No trace of NAD + or NADH is detected in the extracellular medium throughout the conversion which is continued for a further 200 hours. The final conversion rate is 80%.

Example 10: Highlighting the entry of NAD and its retention by
cells
After culturing and harvesting Bacillus megaterium cells
ATCC 39118 under conditions identical to those described in Example 1, these cells are resuspended at a rate of 0.094 g of wet cells per ml of solution in a buffer NH4Cl-NH4OH 1
M, pH = 8.2 containing different concentrations of NAD +. The NAD concentrations found in the supernatant after 22 and 45 hours of incubation are shown in Table 4.

TABLE 4

<tb><SEP> Time <SEP> NAD + <SEP> NAD + <SEP> NAD +
<tb><SEP> 3 <SEP> mM / l <SEP> 6 <SEP> mM / l <SEP> 10 <SEP> mM / 1 <SEP>
<tb> O <SEP> hour <SEP> 2.70 <SEP> 5, S2 <SEP> 9.79
<tb> 22 <SEP> hours <SEP> 1.75 <SEP> 4.08 <SEP> 7.96
<tb> 45 <SEP> hours <SEP> 1.41 <SEP> 3.44 <SEP> 7.74
<Tb>
RR
Example 11: Synthesis of L-valine continuously with Bacillus subtilis
impregnated with NAD +
The Bacillus subtilis 168M strain is cultured, harvested and incubated with NAD under the same conditions as those of Example 1. The activities per gram of wet cells are 3.2 units / g for glucose dehydrogenase and 3.7 u / g for valine dehydrogenase and 3 g of wet cells are introduced into the 50 ml of initial reaction medium. The procedure is then followed in the same manner as in Example 1 except for the concentration of O In a 24 hour conversion period, the concentration of L-valine found corresponds to a conversion of 23% in the feed medium which is fixed at 50 mmol / l. After 100 hours of continuous operation, the conversion rate is 15%.

Example 12: Continuous synthesis of L-leucine with Bacillus megaterium
ATCC 39118 impregnated with NAD
The preparation of Bacillus megaterium ATCC 39118 cells is carried out under conditions identical to those of Example 1.

After incubation in a solution of NAD + at 3 mmol-1, the cells are harvested and then put into bioconversion under conditions identical to those of Example 1, with the exception of the precursor which is a solution of o (cetoisocaproate sodium at 50 mmol -1 for the continuous phase The conversion efficiency of α-ketoisocaproate to L-leucine is equal to 80% after 48 hours and 70% after 100 hours of operation. keep on going.

Example 13 Synthesis of various L -, - amino acids
Three continuous operations are carried out as in Example 1, but instead of using the precursor of the α-ketoisovalerate, one is used in the case of sodium pyruvate, in the second case lithium hydroxypyruvate and in one case third case of sodium alpha-ketossmethylvalerate.
L-alanine, L-serine and L-isoleucine with a molar conversion rate substantially identical to that of Example 1.

Example 14 Use of cells derived from the fermentation must
completed in NAD +
Bacillus megaterium ATCC 39118 cells were cultured under conditions identical to those of Example 1. After 30 hours, NAD + was added to the fermentation broth containing 1 g of NAD + in dry weight in order to obtain a final concentration of NAD + equal to 5 mmol. 1- The medium is left at ambient temperature with stirring for 24 hours, the pH of the medium being 6.5, then the cells are harvested by centrifugation and placed in bioconversion under identical conditions to those of Example 1.

The molar conversion yield of L-valine is 61% after 24 hours of continuous and 52% after 150 hours of continuous.

Example 15 Comparison of the recycling rate of NAD fed
batch and continuously
The cells of Bacillus megaterium ATCC 39118 are cultured and harvested before conversion under conditions identical to those described in Example 1. On the one hand, a continuous synthesis of L-valine is carried out under the conditions described in Example 2 with a sodium dPTcetoisetrate concentration in the feed medium equal to 100 mmol / l for 350 hours. On the other hand, a fed batch conversion is carried out for 180 hours under conditions identical to those of the fed batch period described in Example 1. For the continuous, after 350 hours, the L-valine concentration is 90.degree. mmoles.l and the number of recycling of NAD is 660.

In fed batch, there is an L-valine concentration equal to 110.mmoles 1 after 140 hours of synthesis, ie a number of NAD recycling equal to 128, this recycling rate remaining around this value, even while continuing the feeding, the L-valine synthesis is then impaired due to the combined inhibitory effects of ketoacid and L-valine concentrations.

It has furthermore been observed that with cells which have not been pretreated, according to Example 4, the same ratio between the numbers of continuous NAD recycling and NAD in fed batch as obtained with the pretreated cells is obtained.

Claims (10)

1. A process for the continuous production of L-amino acids in which  cultures a strain of Bacillus under conditions such  that cells containing a composition are produced  enzymatic composition comprising amino acid dehydrogenase and a  enzyme capable of catalyzing the regeneration of nicotinamide  adenine dinucleotide, said cells are kept confined  in a reaction zone is subjected to said reaction zone  a feed medium comprising at least one i-keto acid or at least one  at least one of its salts, a source of ammonium ions and a reducing agent.  the action of said cells in the presence of a type coenzyme  nicotinamide adenine dinucleotide in reduced or oxidized form,  characterized in that a withdrawal of an effluent is carried out  containing said L-Kamino acid product, the flow rate is regulated  for admitting the d-keto acid from said feed medium into the  reaction zone, the concentration of i-keto acid in the medium  supply and the withdrawal rate of said effluent so  to maintain an L-amino acid concentration of between 1 and  200 mmol per liter in the reaction zone and so as to  maintain an i-keto acid concentration in the  reaction at most equal to 150 mmol per liter. 2. The method of claim 1 wherein the flow rate  of -ceto acid of said feed medium in the reaction zone  is between 0.01 and 25 mmol per hour and per liter of  reaction volume and preferably between 0.1 and 10 mmol per  hour and per liter of reaction volume. 3. Method according to one of claims 1 to 2 wherein said  coenzyme is brought into said reaction zone by a  ex-situ pretreatment of cells including  incubation of these cells in the presence of said coenzyme in form  oxidized or reduced under conditions such that said cells  contain a substantial amount of said coenzyme in the form  intracellular, sufficient for the production of those  amino acids. 4. Method according to one of claims 1 to 3 wherein the strain  is that of Bacillus megaterium and preferably that of  Bacillus megaterium ATCC 39118. 5. Method according to one of claims 3 to 4 wherein  incubation in the presence of said coenzyme is carried out in a  buffered solution at a pH between 5 and 10, at a rate of 0.02  at 2 mmol of coenzyme per gram of dry weight of cells during  a time of 0.5 to 100 hours and a temperature of 4 to 700C. 6. Method according to one of claims 1 to 5 wherein the  amount of cells involved is between 1 g and 200 g  dry weight per liter of reaction volume. 7. Method according to one of claims 1, 2, 4 and 6 wherein the  amount of coenzyme per liter of reaction volume is from  minus 0.05 mmol. 8. Method according to one of claims 1 to 7 wherein  performs a preliminary step of feeding into the area  reaction at least once ena-keto acid or a salt thereof  and withdrawing said effluent when obtained in the zone of  reaction of the molar concentrations of L-amino-acid defined  in claim 1. 9. Method according to one of claims 1 to 8 wherein said  production of L-amino acid is carried out at a temperature  from 10 to 700 ° C., at a pH of 6 to 10 and in which use is made of  1 to 2 moles of reducing agent and 1 to 3 moles expressed as ammonia  the source of ammonium ions per 1 mole of α-ketoacid. 10. Method according to one of claims 1 to 9 wherein the  reducer is glucose and in which the alpha-keto acid is either  pyruvic acid, either hydroxypyruvic acid or acid  α-ketoisocaproic acid, either methylvaleric acid or alpha-ketones;  alpha-ketoisovaleric acid or a salt thereof.