The invention relates to the field of producing aniline from raw material of renewable resources, such as e.g. biomass via a suitable microbial host followed by chemical conversion of an intermediate product to aniline.

Aniline is currently produced at several million tonnes per year from fossil raw materials, e.g. to produce polyurethanes. An aniline source based on renewable resources, also called "bioaniline", is strongly desired for the chemical industry in order to become independent from fossil resources. More importantly, there is a strong desire to reduce carbon dioxide (CO 2) emissions both for the chemical processes as well as by increasing the use of renewable resources in the raw materials. Bioaniline has a high potential of saving CO2 emissions. The invention further relates to engineering of microorganisms and production of aromatic compounds therefrom. In particular, the invention relates to the field of producing o-aminobenzoate (oAB) from renewable sources, such as e.g. biomass in a suitable recombinant microbial host. Typically a source containing a significant proportion of fermentable sugars is used. These sugars may include polysaccharides such as disaccharides, e.g. sucrose, or trisaccharides, e.g. kestose, as well as C-6 monosaccharides such as glucose, fructose or mannose and C-5 monosaccharides such as xylose and arabinose. A recombinant microbial strain capable of converting sugar to o-aminobenzoate (2- aminobenzoate, ortho -aminobenzoate, o -aminobenzoate, oAB) would enable the production of o- aminobenzoate from a wide range of renewable resources including sugar beet and sugar cane, starch- containing plants such as corn, wheat and rye, as well as lignocellulose e.g. from straw, wood or bagasse.

Currently, there is no renewable or biologically derived source of o-aminobenzoate or the corresponding acid available commercially and no known example of the large-scale biological production of o- aminobenzoate has been described, o - Aminobenzoate is a natural

pathway and a precursor for the biosynthesis of the aromatic amino acid L-tryptophane. The biosynthetic pathway to o-aminobenzoate is relatively well understood in both prokaryotes and eukaryotes. A chemical conversion of o-aminobenzoate to aniline can be achieved. Current production methods of aniline rely on chemical synthesis from petroleum-derived raw-materials. Such petroleum-derived raw materials are not renewable as opposed to raw materials which are renewable, such as the renewable resource "biomass". Several chemical steps involved in the chemical synthesis result in high production costs of the chemicals. The conventional chemical synthesis of aniline can be associated with hazardous intermediates, solvents, and waste products which can have substantial impacts on the environment. Non-specific side-reactions on the aromatic -ring result in the reduction of the product yield. Petroleum-derived raw materials are influenced by cost fluctuations resulting from the global petroleum price.

WO 2013/103894 Al discloses a method of producing aromatic amines via biologically-derived p- aminobenzoic acid (4 -aminobenzoate) . However, this document discloses to produce the />-aminobenzoic acid in either E. coli or in S. cerevisiae and fails to recognize the advantages of Corynebacterium glutamicum as a host. In addition, this document does also not disclose how to successfully combine the fermentation process with the downstream chemical process of converting the b io logically -derived p- aminobenzoic acid to aromatic amines.

A direct fermentation of sugar to aniline as a one-step conversion was thought to be most cost efficient if based on a biosynthesis pathway including an enzymatic, in vivo, decarboxylation of anthranilate to aniline as the final reaction step. Since an aminobenzoate decarboxylase could not successfully be identified or developed through protein engineering, the decarboxylation reaction of anthranilate to aniline could not be carried out by pure enzymatic means. Since such a one- step process was not technically feasible, process alternatives to perform the final reaction step of decarboxylating anthranilate to aniline as the final reaction step were taken into consideration, e.g. by a chemical step, as opposed to an enzymatic step. Therefore, it has been the technical problem of the invention to provide a method of producing aniline based on renewable resources that is superior to existing fermentation and chemical methods and that achieves a large reduction in carbon dioxide emissions, independence from fossil resources, and similar or lower production cost compared to the established petroleum-based production processes. The invention has solved said problem by providing a recombinant microbial host cell capable of converting a raw material comprising a fermentable carbon substrate to o -aminobenzoate biologically.

The invention has further solved said problem by providing a method for producing aniline, comprising the steps of:

a) producing o-aminobenzoate by fermentation of a raw material comprising at least one fermentable carbon substrate using the recombinant microbial host cell of the invention as described herein and as claimed in the claims that is capable of converting said raw material comprising at least one fermentable carbon substrate to o -aminobenzoate biologically, wherein said o-aminobenzoate comprises anthranilate anion, b) converting said o-aminobenzoate from said anthranilate anion to anthranilic acid by acid protonation,

c) recovering said anthranilic acid by precipitation or by dissolving in an organic solvent, and d) converting said anthranilic acid to aniline by thermal decarboxylation in an organic solvent.

The change to aniline production based on renewable resources, e.g. biomass or fermentable carbon sources, offers the advantages of reducing CO2 emissions significantly, allows for independence from fossil resources, and enables a possible reduction in production cost. A further advantage of the invention is that the use of hazardous chemicals and the resulting waste are kept to a minimum. Further, biologically derived o-aminobenzoate can be produced and converted to aniline in a process with much less overall impact on the environment.

In particular, the invention provides a recombinant microbial host cell capable of converting a raw material comprising a fermentable carbon substrate to o -aminobenzoate biologically. Such a recombinant microbial host according to the invention can be a genetically engineered microorganism for fermentative production of o-aminobenzoate from renewable sources, such as e.g. biomass. For example, the invention provides genetically engineered strains of Corynebacterium glutamicum as biocatalysts that are suitable for efficient fermentative production of o-aminobenzoate from fermentable carbon sources. Corynebacterium glutamicum is a soil-dwelling Gram-positive bacterium. It belongs to the high GC content Gram-positives actinobacteria. C. glutamicum is one of the biotechnologically most important bacterial species with an annual production of more than two million tons of amino acids, mainly L- glutamate and L-lysine. Due to this immense industrial importance, C. glutamicum has been studied extensively starting already shortly after its discovery in 1956. The genome of C. glutamicum was sequenced and published in 2003 (Ikeda M, Nakagawa S. Appl Microbiol Biotechnol, 2003, 62:99.; Kalinowski J, Bathe B, B artels D, Bischoff N, Bolt M, Burkovski A, Dusch N, Eggeling L, Eikmanns BJ. Gaigalat L, Goesmann A, Hartmann M, et al., J Biotechnol, 2003, 104: 5). C. glutamicum exhibits numerous ideal intrinsic attributes as a microbial factory to produce not only amino acids but also other chemicals. The deep fundamental knowledge of the corynebacteria physiology and the postgenomic tools to model processes or design synthetic pathways in combinatorial approaches constitute a foundational basis to design efficient and versatile corynebacterial biorefineries. The biosynthetic pathways of C. glutamicum leading to oAB have been extensively studied and can be divided into three main sections: the central metabolism, the common aromatic pathway, and the L-tryptophan branch pathway. oAB is an intermediate of L-tryptophan biosynthesis (Figure 1). However, previous attempts at strain improvement for amino acid production relied mainly on classical random mutagenesis and screening procedures, aiming at deleting competing pathways and eliminating feedback regulations in the biosynthetic pathways. The classical approach has limited usefulness since complete deregulation of regulatory steps and enhancement of an appropriate biosynthetic enzyme activity are difficult to achieve. Furthermore random mutagenesis often produces unexpected mutations at some locations in the genome together with desirable ones. Since it is difficult to ascertain the influence of these unidentified mutations, further strain improvement would be affected. These problems widely exist in industrial strains constructed by random mutagenesis. On the basis of literature reported mainly during the last decade, the current production with strains generated by the classical approach yields towards sugar (wt%) approximately 20-23 for L- tryptophan and around 25 for L-phenylalanine. In contrast, far higher production yields towards sugar (wt%) have been reported for many other amino acids, e.g., L- lysine, 40-50; L-glutamate, 45-55; L- arginine, 30-40; L-threonine, 40-50, L-valine, 30-40; and L- alanine, 45-55 (Leuchtenberger W, Huthmacher K. Drauz K. Appl Microbiol Biotechnol, 2005, 69: 1.; Ikeda M, Nakagawa S, Appl Microbiol Biotechnol, 2003, 62:99).

Based on the recent advantages made in molecular biology and in the understanding of the functionality of the biosynthetic pathways leading from sugars to L -tryptophan (and oAB), the invention provides a recombinant microbial host cell capable of converting a raw material comprising a fermentable carbon substrate to o-aminobenzoate biologically, and more specifically, an oAB producer from Corynebacterium glutamicum ATCC13032 based on directed mutagenesis and the metabolic engineering approach. Certain gene targets for metabolic engineering to generate an oAB producer are located in the aromatic biosynthesis pathway leading to oAB and subsequent to L-tryptophan (Figure 1). The biosynthesis of L-tryptophan is strictly controlled at several steps in C. glutamicum. Therefore, overproduction of oAB required the genetic removal of the metabolic controls existing both in the common aromatic pathway and in the L-tryptophan branch. In addition, amplification of the DA H synthase, which initiates the aromatic pathway, was an important strategy to increase net carbon flow down the common pathway as a further embodiment of the invention. These mutations were combined with the generation of aromatic amino acid bradytroph strains to circumvent the constant need for aromatic amino acid supplementation. A balanced supply of precursors was addressed to achieve efficient production of the product, considering that biosynthesis of 1 mol of oAB from glucose requires 1 mol of erythrose-4-phosphate (E4P) and 2 mol phosphoenolpyruvate (PEP) as starting precursors and, in addition, consumes 1 mol of L-glutamine and releases 1 mol of pyruvate on its pathway.

In the following, the inventors focus sed on Corynebacterium glutamicum as a recombinant microbial strain for producing o-aminobenzoate biologically. Therefore, the invention provides a recombinant microbial host cell capable of converting a raw material comprising a fermentable carbon substrate to o-aminobenzoate biologically. Said Corynebacterium glutamicum preferably is Corynebacterium glutamicum ATCC 13032, most preferably a recombinant Corynebacterium glutamicum ATCC 13032.

In further embodiments of the invention, the Corynebacterium glutamicum strain, preferably Corynebacterium glutamicum ATCC 13032 can comprise a genetic modification of the trpD gene (Cgl3032, SEQ ID NO: 1) encoding anthranilate phosphoribosyl transferase, wherein said genetic modification preferably is selected from the group consisting of SEQ I D NO: 2, SEQ ID NO: 3, SEQ I D NO: 4, SEQ ID NO: 5, SEQ I D NO: 6, SEQ I D NO: 7, and SEQ ID NO:8, which all have the effect of a reduced expression of the trpD gene to yield a tryptophan auxotrophic strain. A single one of the aforementioned genetic modifications already results in a recombinant Co rynebacterium glutamicum ATCC 13032 strain that produces o-aminobenzoate (oAB).

In the following, a few terms used to describe the invention are defined. The term "bioaniline" according to the invention refers to aniline that is based on raw material from renewable resources, such as sugar beet, sugar cane, starch-containing plants, preferably corn, wheat and rye, and lignocellulose, preferably straw, wood and bagasse, glycerol and CI -compounds, preferably CO. or such as fermentable sugars, preferably C-5 monosaccharides, C-6 monosaccharides, disaccharides, and tri-saccharides, wherein the C-5 monosaccharides preferably are xylose and arabinose, and wherein the C-6 monosaccharides preferably are glucose, fructose or mannose, and wherein the disaccharide preferably is saccharose, and wherein the trisaccharide preferably is kestose.

"o-aminobenzoate" according to the invention refers to ortAo-aminobenzoate (o-aminobenzoate, "oAB"). o-aminobenzoate can be present in the form of the anthranilate salt comprising the anthranilate anion, C VHJCOX) , and a suitable cation, such as NH4 or Na ^÷ , or as anthranilic acid, which is the zwitter ion C. IU ^' OO NH ₃ ⁺ and ( i U ^' OO NH ₂ . "o-aminobenzoate" ("oAB") is different from "4-aminobenzoate" ("pAB") in that the amino group is attached to the benzene ring at the exposition (para) as opposed to the C ₂ -position (ortho) in the case of o -aminobenzoate ("oAB"). The term "host" within the meaning of the invention can comprise any host that is capable of producing o- aminobenzoate by fermentation, either naturally, or only after transformation as a "recombinant microbial host", or in addition to the naturally present o-aminobenzoate, either in the form of the anthranilate anion or as anthranilic acid, following transformation. A "microbial host" according to the invention can be selected from the group consisting of bacteria, yeast and fungi. Said host can be selected from the group consisting of bacteria, yeast and fungi, wherein said bacterium preferably is an Escherichia coli strain, a Corynebacterium strain or a Pseudomonas strain, wherein said Corynebacterium strain preferably is Corynebacterium glutamicum and wherein said Pseudomonas strain preferably is Pseudomonas putida. Preferably, said microbial host can be a recombinant microbial host. Such a recombinant microbial host can be E. coli W31 10 trpD9923, as shown in Example 1. Such a recombinant host can be a Pseudomonas putida KT2440, as shown in Example 4. Such a recombinant host can also be a Corynebacterium glutamicum ATCC 13032, as shown in Example 3.

The term "genetic modification" within the meaning of the invention refers to changes in nucleic acid sequence of a given gene of a microbial host as compared to the wild-type sequence. Such a genetic modification can comprise deletions as well as insertions of one or more deoxy-ribo nucleic acids. Such a genetic modification can comprise partial or complete deletions as well as insertions introduced by transformations into the genome of a microbial host Such a genetic modification can produce a recombinant microbial host, wherein said genetic modification can comprise changes of at least one, two, three, four or more single nucleotides as compared to the wild type sequence of the respective microbial host. For example, a genetic modification can be a deletion or insertion of at least one, two, three, four or more single nucleotides or a transformation of at least one, two, three, four or more single nucleotides. A genetic modification according to the invention can have the effect of e.g. a reduced expression of the respective gene or of e.g. an enhanced expression of the respective gene. In one example of such a genetic modification according to the invention, a recombinant microbial host, e.g. Corynebacterium glutamicum, can comprises a genetic modification of the trpD gene (Cgl3032, SEQ ID NO: 1) encoding the enzyme anthranilate phosphoribosyl transferase, wherein said genetic modification can have the effect of a reduced expression of the modified trpD gene. Such modified trpD gene can be selected from the group consisting of SEQ ID NO: 2, SEQ ID NO: 3, SEQ ID NO: 4, SEQ ID NO: 5, SEQ ID NO: 6, SEQ ID NO: 7, and SEQ ID NO: 8. A single one of the aforementioned genetic modifications already results in a recombinant Co rynebacterium glutamicum ATCC 13032 strain showing a reduced expression of the modified trpD gene that produces o-aminobenzoate (oAB). In a further embodiment of the invention, such genetic modification can also be a deletion of the trpD gene, e.g. as in Corynebacterium glutamicum AtrpD. The term "batch fermentation" within the meaning of the invention refers to a single fermentation reaction having a defined starting point and a defined end point. Batch fermentation can be used in step a) of the method according to the invention in cases where the production rates of the microorganisms cannot be maintained at a high rate in continuous fermentation mode. The term "fed-batch fermentation" within the meaning of the invention Fed-batch fermentation refers to an operational technique in biotechnological processes where one or more nutrients (substrates) are fed (supplied) to the bioreactor during cultivation and in which the product(s) remain in the bioreactor until the end of the run. "Fed-batch fermentation" can be used in step a) of the method according to the invention in cases where the production rates of the microorganisms cannot be maintained at a high rate in continuous fermentation mode.

The term "continuous fermentation" within the meaning of the invention, refers to a fermentation method in which substrate is added and the product (i.e. o-aminobenzoate, oAB) is removed continuously during the fermentation in step a) of the method according to the invention.

I n the following, the invention is described in more detail.

The invention provides a recombinant microbial host cell capable of converting a raw material comprising a fermentable carbon substrate to o-aminobenzoate (oAB) biologically.

Said microbial host cell can be selected from the group consisting of bacteria, yeast and fungi, wherein said bacterium preferably is an Escherichia coli strain, a Corynebacterium strain or a Pseudomonas strain, and wherein said Corynebacterium strain preferably is Corynebacterium glutamicum, more preferably Corynebacterium glutamicum ATCC 13032, and wherein said Pseudomonas strain preferably is Pseudomonas putida, more preferably Pseudomonas putida KT2440.

In a further embodiment of the recombinant microbial host cell of the invention, the recombinant microbial host cell can be Corynebacterium glutamicum. In more specific embodiments of the invention, Corynebacterium glutamicum ATCC 13032 is the preferred recombinant microbial host for the production of o-aminobenzoate, since the inventors observed that the organism has a high tolerance for o- aminobenzoate, whereas for other microorganisms, e.g. Escherichia coli K 12. already low concentrations of o- aminobenzoate are toxic.

I n a further embodiment of the recombinant microbial host cell of the invention, said Corynebacterium glutamicum, preferably Corynebacterium glutamicum ATC 13032, can comprise a genetic modification of the trpD gene encoding anthranilate phosphoribosyl transferase (Figure 3). Said genetic modification of the trpD gene (Cgl3032, SEQ II ^" ) NO: 1) encoding anthranilate phosphoribosyl transferase can be selected from the group consisting of SEQ ID NO: 2, SEQ ID NO: 3, SEQ ID NO: 4, SEQ I NO: 5, SEQ I D NO: 6, SEQ ID NO: 7, and SEQ ID NO: 8, and wherein said genetic modification has the effect of a reduced expression of the trpD gene. More specifically, the mutations in the aforementioned sequences can change the spacer region between the ribosomal binding site and the start codon of the trpD gene as. In addition, an exchange of the natural start codon of the trpD gene can lead to a reduced translation of trpD and thus to a bradytroph strain towards L-tryptophan that produces o-aminobenzoate in a multi-gram scale without supplementation with L-tryptophan. It follows that such type of strain is particularly preferred as a recombinant microbial hast according to the invention. in further embodiments of the recombinant microbial host cell of invention, the microbial host cell can be Corynebacterium glutamicum, preferably Corynebacierium glutamicum ATCC 13032, wherein said Corynebacterium glutamicum or Corynebacterium glutamicum ATCC 13032 can comprise a genetic modification of the csm gene (Cgl0853, SEQ ID NO: 9) encoding chorismate mutase (Figure 3), wherein said genetic modification preferably is selected from the group consisting of SEQ I D NO: 10, SEQ I D NO:l l , SEQ ID NO: 12, SEQ ID NO: 13, SEQ ID NO: 14, SEQ ID NO: 15 and SEQ ID NO: 16, and wherein said genetic modification has the effect of a reduced expression of the csm gene. In particular, the csm gene can be manipulated to further improve o-aminobenzoate production rates by reduction of L -phenylalanine and L -tyrosine production rates to a minimum.

The effect of reducing the expression of the csm gene is that a bradytroph strain towards L -tyrosine and L- phenylalanine results, that produces o-aminobenzoate in a multi-gram scale without supplementation with L-tyrosine and L -phenylalanine. This circumvents the reduction of o-aminobenzoate production due to feedback-inhibition of enzymes catalyzing reactions in the aromatic acid pathway by L-tyrosine and/or L- phenylalanine, as possible in L-tyrosine and L-phenylalanine auxotrophic strains supplemented with L- tyrosine and/or L-phenylalanine. It follows that such type of strain is particularly preferred as a recombinant microbial hast according to the invention. This effect can result in a higher production of o- aminobenzoate. In further embodiments of the recombinant microbial host cell of invention, the Corynebacterium glutamicum microbial host cell, preferably Corynebacterium glutamicum ATCC 13032, can further comprise one or more deletions selected from the group consisting of hpr (Cgll937 - SEQ I D NO: 17 or SEQ ID NO: \ K), ptsG (Cgll537 - SEQ ID NO:19 or SEQ ID NO:20), pepco (Cgll523 - SEQ ID NO:21 or SEQ ID NO:22), /7>vfc (Cgl2089 - SEQ I D NO: 23 or SEQ I D NO: 24), and gpi (Cgl0851 - SEQ I D NO: 27 or SEQ I D NO: 28).

The gene hpr (Cgll937 - SEQ I D NO: 17 or SEQ I D NO: 18) and the gene ptsG ( SEQ ID NO: 19 or SEQ ID NO:20): each one of these genes encodes a unit of the multi-enzyme complex PEP- phosphotransferase system (PTS) (Figures I . 2). Both units of PTS, Hpr and PtsG, of C. glutamicum - Si - were manipulated and compared with regard to the effects on growth, cell viability, as well as oAB production, as shown in Example 3.

The gene pepco (Cgll523 - SEQ ID NO:21 or SEQ ID NO:22) encodes the enzyme phosphoenolpyruvate-carboxylase. Phosphoenolpyruvate-carboxylase consumes PEP (phosphoenolpyruvate) by transforming it into oxaloacetate (Figures 1, 2). This reaction is the major anaplerotic reaction under oxygen deprivation, whereas under standard sufficient oxygen supply the pyruvate kinase consumes most of the PEP to generate pyruvate under gain of one ATP. The gene pepco of C. glutamicum was manipulated and compared with regard to the effects on growth, cell viability, as well as oAB production, with other C. glutamicum strains, as shown in Example 3.

The gene pyk (Cgl2089 - SEQ I NO: 23 or SEQ I D NO: 24) encode the enzyme pyruvate kinase. Pyruvate kinase consumes PEP (phosphoenolpyruvate) by production of pyruvate and ATP (Figures 1, 2). This reaction is part of the glycolysis in C. glutamicum. The Pyk-encoding gene, annotated in the genome of C. glutamicum, was deleted, as described in Example 3.

The gene gpi (Cgl0851 - SEQ ID NO: 27 or SEQ I D NO: 28) encodes the enzyme glucose-6- phosphate isomerase. The enzyme catalyses the first step of the glycolysis transforming glucose-6- phosphat into fructose-6-phosphat (Figures 1, 2).

In more specific embodiments of the recombinant microbial host cell of invention, the Corynebacterium glutamicum microbial host cell, preferably Corynebacterium glutamicum ATCC 13032, can further overexpress one or more of the genes selected from the group consisting of galP (SEQ ID NO: 30), iolT2 (SEQ ID NO: 31), ppgk (SEQ I D NO: 32), pps (SEQ ID NO: 33-35), ppk (SEQ ID NO: 36), zwfi ( SEQ I NO: 37), opcA (SEQ ID NO: 38-39), tktCg ( SEQ ID NO: 40), tktEc (SEQ ID NO: 41), talCg (S EQ ID NO: 42), to/Ec (SEQ I NO: 43), a gene encoding TrpEGS38F (trpEG ^S38F , SEQ I D NO: 50), a gene encoding TrpEGS38R (trpEG ^S38R , SEQ ID NO: 51), a gene encoding TrpEGS40R (trpEG ^S4m , SEQ ID NO: 52), a gene encoding TrpEGS40F {trpEG ^S40F , SEQ ID NO: 53), the gene encoding AroGD146N (aroG ^D146N , SEQ ID NO: 55), aroL of Escherichia coli LJ110 (S EQ ID NO: 93), aroK (SEQ I D NO: 94), glnA (SEQ ID NO: 95), and qsuA (SEQ I D NO: 96).

The expression of feedback-resistant trpEG genes (anthranilate synthase genes), a gene encoding TrpEGS38F (trpEG ^S3SF - SEQ I D NO: 50), a gene encoding TrpEGS38R (trpEG ^S3M - SEQ ID NO: 51), a gene encoding TrpEGS40R (trpEG ^S40R - SEQ ID NO: 52), a gene encoding TrpEGS40F (trpEG ^S40F - SEQ ID NO: 53), in the Corynebacterium glutamicum microbial host, preferably Corynebacterium glutamicum ATCC 13032, is a preferred embodiment of the invention, alone or in combination with any one of the other genetic modifications of the invention. The genes all share the common feature that they encode mutated versions of the anthranilate synthase unit, which releases a pyruvate molecule under formation of o-aminobenzoate and L-glutamate (Figure 3). These are feedback -resistant versions of Tr EG, which is why the corresponding mutated genes are referred to as trpEG'^.The overexpression of feedback resistant aroG gene (3-deoxy-D-arabinoheptulosonate 7- phosphate synthase gene) in the Corynebacterium glutamicum microbial host, preferably Corynebacterium glutamicum ATCC 13032, such as the gene encoding AroGD146N {aroG ^D146N — SEQ I D NO: 55), is another preferred embodiment of the invention, alone or in combination with any one of the other genetic modifications of the invention. The gene encoding AroGD146N (aroG ^D!46N - SEQ ID NO: 55) encodes a mutated version of AroG, which is an enzyme that catalyses the reaction from erythrose-4-phosphate (E4P) to 3 -deoxy-D-arabino-heptulosonate-7 -phosphate (DAHP) in Escherichia coli (Figure 3).

The gene gal? (Cgi2409 - SEQ ID NO: 30) encoding the enzyme galactopermease. The glucose uptake can be restored after PTS disruption by expression of a galactopermease (Figures 1, 2). in C. glutamicum a gene was identified and annotated as galactopermease and is therefore a good candidates for such an expression. The candidate gene is expressed in C. glutamicum strains with disrupted PTS, as shown in Example 3.

The gene iolT2 (Cgl3058 - SEQ ID NO: 31) encodes the inositolpermease T2 unit. The glucose uptake can be restored after PTS disruption by expression of a inositolpermease T2 unit (Figures 1, 2). As an alternative approach to the overexpression of a galactopermease to restore glucose-uptake after PTS disruption in C. glutamicum, the overexpression of the inositolpermease T2 unit of the inositolpermease is performed. The resulting strain characteristics are compared to that of strains overexpressing a galactopermease, as shown in Example 3.

The gene ppgk (Cgll910 - SEQ ID NO: 32) encodes the enzyme polyphosphoglucokinase. The galactopermease, as well as the inositolpermease T2 unit, can only take up glucose, but they cannot phosphorylate the sugar molecule (Figures 1, 2). This phosphorylation, however, is essential to enable the metabolism of the glucose molecule. Therefore, expression of the different permeases is combined with expression of a glucose phosphorylating enzyme. This Ppgk is co-expressed in the different permease- expressing PTS deficient strains of the invention, as described in Example 3. The genes pps (Cgl0551 and Cgl0552 - SEQ ID NO: 33-35) encode the enzyme PEP (phosphoenolpyruvate) synthase. Apart from the disruption of reactions consuming PEP, biosynthetic steps leading to the generation of PEP are promoted. PEP synthase catalyses the formation of PEP by pyruvate recycling and consumption of ATP to AMP, as part of the gluconeogenesis in C. glutamicum (Figures 1, 2). Overexpression of the pps gene results in an increased oAB pool in C. glutamicum strains, as described in Example 3.

The gene ppk (Cgl2862 - SEQ ID NO: 36) encodes the enzyme PEP carboxykinase. As part of the glyoxylate pathway oxaloacetate can be recycled into PEP by a PEP carboxykinase (Figures 1, 2). Overexpressing the gene can be another route to a larger PEP pool available for oAB biosynthesis, as described in Example 3.

The gene zvv (Cgll576 - SEQ ID NO: 37) and the gene (opcA (Cgl577 - SEQ ID NO: 38-39) encode the enzyme glucose-6-phosphat dehydrogenase (Figures 1, 2). The enhanced production of the enzyme catalysing the first reaction of the pentose phosphate pathway (PPP) (formation of ribulose-5 - phosphate from glucoce-6-phosphate) (Figures 1, 2) can result in an increased flux into the PPP leading to the production of higher amounts of E4P and via fructose-6-phosphate (Frc-6-P) production and can furthermore increase the cells PEP pool, finally resulting in an increased oAB pool in C. glutamicum strains. This effect is achieved when this manipulation is applied to C. glutamicum strains and an expression of the Zop enzyme is performed, as described in Example 3.

The gene tktCG (Cgll574 SEQ I D NO: 40) and the gene tktEC (ECDH10B_3110 - SEQ I D NO: 41 ) encode the enzyme transketolase. An enhanced flux through the PPP and an increased E4P pool, and even an increased production of oAB and aromatic amino acids is observed by the expression of a transketolase in C. glutamicum strains (Figures 1, 2). A transketolase gene from E. coli was overexpressed as well as the transketolase encoded in C. glutamicum, as described in Example 3.

The gene talCG (Cgll575 SEQ I D NO: 42) and the gene talEC ( ECDH ! OB 2629 - SEQ I D NO: 43) encode the enzyme transaldolase. A comparable favorable effect on E4P production as by transketolase overexpression can be observed for the overexpression of transaldolases in E. coli (Figures 1, 2). As the sequences of E. coli and C. glutamicum transaldolases differ significantly the inventors proceeded to express the described E. coli gene and the natural C. glutamicum gene in C. glutamicum strains of the invention, as described in Example 3. The transaldolase and the transketolase encoding genes are combined on one vector and co- expressed in C. glutamicum strains to further enhance oAB producer strain characteristics, as described in Example 3.

The gene qsuA (Cgr0492 - SEQ ID NO: 96) encodes the drug resistance transporter protein QsuA. Expressing the gene in C. glutamicum strains results in an optimized of oAB transport, as described in Example 3.

The gene aroL (B0388 - S EQ ID NO: 93) from Escherichia coli LJi l O encodes an enzyme that catalyses the reaction from shikimate (SHI) to shikimate-3 -phosphate (SHBP)) in Escherichia coli. The aroL gene from E. coli is expressed in C. glutamicum strains, as described in Example 3.

The gene aroK (Cgll622 - SEQ ID NO: 94) encodes an enzyme that catalyses the reaction from shikimate (SHI) to shikimate-3 -phosphate (SHBP) in Corynebacterium glutamicum ATCC 13032 (Figure 3). The aroK gene is expressed in C. glutamicum strains, as described in Example 3.

The gene glnA (Cgl2214 - SEQ ID NO: 95) encodes the enzyme glutamine synthethase in

Corynebacterium glutamicum ATCC 13032 (Figure 3). The glnA gene is expressed in C. glutamicum strains, as described in Example 3. In particularly preferred embodiments of the invention, the recombinant microbial host cell can be C. glutamicum AtrpD, C. glutamicum AtrpD::trpD5, C. glutamicum AtrpD ::trpD5Apepco, C. glutamicum AtrpD ::trpD5Agpi, C. glutamicum AtrpD ::trpD5Apyk, C. glutamicum AtrpD: : trpD5l pEKEx2 - trpEG^, C. glutamicum AtrpD::trpD5lpEKEx2- aroG ^D146N , C. glutamicum AtrpD: : trpD5/pEKEx2 - aroG ^D146N -trpEG ^S4(IF , C. glutamicum AtrpD::trpD5/pEKEx2-aroL, C. glutamicum AtrpD::trpD5/pSB072, C. glutamicum AtrpD::trpD5/pSB073, C. glutamicum AtrpD::trpD5/pSB074, C. glutamicum AtrpD: : trpD5l pSBO 75, C. glutamicum A trpD : : trpD5 lpSB076, C. glutamicum AtrpD::trpD5/pSB077, C. glutamicum AtrpD::trpD5lpSB078, C. glutamicum AtrpD::trpD5lpSB085, or C. glutamicum AtrpD::trpD5/pSB096, as further described in Example 3 and in Table 4. In the following further recombinant microbial hosts of the invention are described.

I n yet another embodiment of the recombinant microbial host cell of the invention, the microbial host cell can be Pseudomonas putida, preferably Pseudomonas putida KT2440. In a further embodiment of the recombinant microbial host cell of the invention, said Pseudomonas putida, preferably Pseudomonas putida KT2440, can comprise a deletion of the trpDC gene encoding anthranilate phosphoribosyl transferase and indole-3 -glycerol phosphate synthase ( PP 042 1 and PP 0422 - SEQ ID NO: 63-64), or a deletion of the pheA gene encoding chorismate mutase and prephenate dehydratase ( PP ! 769 SEQ I NO: 65-66), or both, as described in Example 4 (Figure 4).

I n a further embodiment of the recombinant microbial host cell of the invention, said Pseudomonas putida, preferably Pseudomonas putida KT2440. wherein said Pseudomonas putida or Pseudomonas putida KT2440 can express one or more of the genes selected from the group consisting of the gene encoding TrpEGS40F {trpEG ^S40F - SEQ ID NO: 53) and the gene encoding AroGD146N {aroG ^D146N

SEQ I D NO: 55) (Figure 4). In certain embodiments of the recombinant microbial host cell of invention, the one or more expressed genes mentioned above can be integrated into said P. putida microbial host cells by plasmid transformation or by chromosomal transformation, as described in Example 4.

The raw material comprising a fermentable carbon substrate that is converted by the recombinant microbial host cell biologically to o-aminobenzoate (oAB) can be selected from the group consisting of sugar beet, sugar cane, starch-containing plants, preferably corn, wheat and rye, and lignocellulose, preferably straw, wood and bagasse, glycerol and CI -compounds, preferably CO.

In further embodiments of the invention, the fermentable carbon substrate that is comprised in the raw material, can be selected from the group consisting of C-5 monosaccharides, C-6 monosaccharides, disaccharides, and tri-saccharides, wherein the C-5 monosaccharides preferably are xylose and arabinose, and wherein the C-6 monosaccharides preferably are glucose, fructose or mannose, and wherein the disaccharide preferably is saccharose, and wherein the trisaccharide preferably is kestose.

The invention has further solved the above problem by providing a method for producing aniline, comprising the steps of: a) producing o-aminobenzoate by fermentation of a raw material comprising at least one fermentable carbon substrate using the recombinant microbial host cell, as described herein and as claimed in the claims, capable of converting said raw material comprising at least one fermentable carbon substrate to o-aminobenzoate biologically, wherein said o- aminobenzoate comprises anthranilate anion, b) converting said o-aminobenzoate from said anthranilate anion to anthranilic acid by acid protonation,

c) recovering said anthranilic acid by precipitation or by dissolving in an organic solvent, and d) converting said anthranilic acid to aniline by thermal decarboxylation in an organic solvent. The method according to the invention provides the technical advantages of producing aniline whilst achieving a large reduction in CO2 emissions as compared to methods that are based on fossil resources, independence of such fossil resources, as well as achieving similar or lower production costs as compared to the established petroleum based production processes for aniline. The method according to the invention comprises two main parts, in which the first part, step a), is fermentation -b ased (biotechnological) and comprises a conversion of raw material comprising at least one fermentable carbon substrate to o-aminobenzoate by fermentation by a suitable recombinant microbial host, wherein said o-aminobenzoate comprises anthranilate anion (CeHsCOO ^" N¾). The second part is of a more chemical nature and follows downstream of step a), and comprises the following steps a) to d), and optionally e):

b) converting said o -aminobenzoate from said anthranilate anion to anthranilic acid by acid protonation,

c) recovering said anthranilic acid by precipitation or by dissolving in an organic solvent, and d) converting said anthranilic acid to aniline by thermal decarboxylation in an organic solvent.

In summary, the invention therefore provides a method for producing aniline, which combines a first biotechnological fermentation step a) with a subsequent downstream, chemical process comprising at least the steps b), c) and d), with an optional step e).

in the following, the invention is described in further detail. The overall concept of the method according to the invention is depicted in Figure 5, and a more detailed overview is presented in Figure 6.

The invention further provides a method for producing aniline, comprising the steps of: a) producing o-aminobenzoate by fermentation of a raw material comprising at least one fermentable carbon substrate using the recombinant microbial host cell, as described herein and as claimed in the claims, capable of converting said raw material comprising at least one fermentable carbon substrate to o -aminobenzoate biologically, wherein said o- aminobenzoate comprises anthranilate anion,

converting said o -aminobenzoate from said anthranilate anion to anthranilic acid by acid protonation,

recovering said anthranilic acid by precipitation or by dissolving in an organic solvent, and converting said anthranilic acid to aniline by thermal decarboxylation in an organic solvent.

The fermentation of step a) can be a batch fermentation, a fed-batch fermentation or a continuous fermentation. In the case of a continuous fermentation in step a), a cell retention device may be used to separate the host from a water phase used in fermentation step a) and to retain the host in a fermenter used for performing step a). The cell retention device may be incorporated in such a fermenter, or, as a preferred option, it can be located outside the fermenter as a separate device. Such a separation may be achieved by centrifugation, e.g. in a disc separator or a decanter, by gravity in a settling device, or by a filtration technique such as cross flow filtration. The separation can be designed with a limited retention time so that the host, e.g. cells of bacteria, yeast and fungi, spend a limited time outside of the fermenter and do not lose their ability to grow or produce anthranilate, while they are in the separation device. The continuous fermentation with cell retention allows the fermentation step a) to be run with a high cell density and thus a high space time yield. More importantly, the fermentation in step a) can be run with a very low growth rate. This results in a higher overall product yield (g of product per g of raw material) since less sugar is used for the production of biomass and more for the production of oAB.

A further advantage of the continuous fermentation over batch fermentation is that the so-called "down time" of the fermenter can be reduced to a minimum. A continuous fermentation has a much longer production phase than a batch fermentation so the time spent on cleaning, sterilization, filling and harvesting is proportionally much shorter. This aspect contributes to the increase the space time yield and therefore reduces the investment costs for a given plant capacity, which is used to run the method according to the invention.

If step a) of the method is run as a batch or a fed-batch fermentation the total fermentation capacity can be divided between several fermenters, and break tanks can be used to enable a continuous supply of fermentation material of step a) to the downstream section comprising steps b) to d), or even e). Batch or fed-batch fermentation can be used in step a) of the method according to the invention in cases where the production rates of the microorganisms cannot be maintained at a high rate in continuous fermentation mode. Step a) of the method according to the invention provides o-aminobenzoate comprising anthranilate anion by fermentation of a raw material comprising fermentable sugars using a host, preferably a recombinant host that is capable of converting sugars to o-aminobenzoate by fermentation. Preferably, a fermentation reactor comprising such a host is cultivated with the addition of a carbon source, for example corn syrup, sugar can juice, molasses, and the like and a nitrogen source, for example ammonia gas, ammonium hydroxide solution, ammonium sulphate, ammonium nitrate, corn steep liquor, and the like, as well as the respective micro-nutrients needed for survival of the microorganism. The pH in the fermentation of step a) can be in the range of 3 to 9, preferably between pH 4 to 8, most preferably kept at a value between 6.5 and 7.5. The pi I in step a) can be adjusted e.g. by the addition of a base, for example, ammonia gas, ammonium hydroxide or sodium hydroxide.

In a further embodiment of the method according to the invention, at least step a) of producing o- aminobenzoate by fermentation and step b) of converting said o-aminobenzoate from said anthranilate anion to anthranilic acid by acid protonation can be run continuously. In this embodiment, a fermenter in which step a) is performed is operated continuously, wherein a fermentation broth produced in step a) can be withdrawn continuously and processed through a device to separate the biomass comprising the host grown to a certain density, for example a filter, a centrifuge, membranes, etc. This biomass can be recycled to the fermenter used in the fermentation of step a) after purging a small portion the biomass comprising the recombinant microbial host. Such purge stream from the biomass can be useful in order to avoid biomass accumulation. A portion of microbial host cells that multiply in the fermenter and dead cells can thus be removed in order to keep the concentration of live host cells in the reactor of fermentation step a) within defined limits, most preferably constant. This can be different in the case of fed-batch fermentation, where the recombinant host cells and the fermentation product(s) remain in the bioreactor until the end of the run, which therefore is not a continuous fermentation but a fed-batch fermentation. Sufficient oxygen can be added to the reactor used in step a) either as pure oxygen, as air, or as enriched air. The cell free fermentation broth of the fermentation step a) can essentially be a solution of an anthranilate salt comprising the anthranilate anion and a suitable counter cation, such as ¾ ⁺ or Na ⁺ . The anthranilate solution produced in fermentation step a) can have a concentration between 5 g/L and 500 g/L, preferably between 20 g/L and 200 g/'L, most preferably between 50 g/L and 150 g/L of anthranilate salt.

In a further embodiment of the method according to the invention, the recombinant microbial host used in the fermentation of step a) is removed prior to performing step b) of converting said o-aminobenzoate from said anthranilate anion to anthranilic acid. Preferably, the removed recombinant microbial host is re- fed to the fermentation of step a). This has the technical advantage of allowing a continuous fermentation process, which can be particularly efficient and cost-effective. o-aminobenzoate (oAB) is an amino acid; thus its ionization state, and solubility in water, depends on the pH; with the zwitterion being the least soluble form. At pH 7, which prevails in the fermenter that can be used in fermentation step a), oAB exists as anions buffered with cations such as Na ⁺ or NELf . Adding an acid, such as HC1, until the pH drops to the isoelectric point therefore causes precipitation of oAB crystals. Accordingly, method step b) of the method according to the invention comprises converting the anthranilate anion to anthranilic acid by acid protonation thereby forming a precipitate comprising anthranilic acid crystals. For this reason, method step b) can also be referred to as "crystallisation". Specifically, the anthranilate solution of step a) can be mixed with an acid, preferably HQ. Thus, the pH can be reduced to a value between 2 and 4. preferably between 3.2 and 3.6. This causes a change in solubility and results in the precipitation of the anthranilate crystals. The anthranilate crystals can be recovered, preferably by filtration, in the subsequent step c) ("recovery"). The residual moisture and inorganic salt contents in the precipitated anthranilate crystals ("cake") depends on the soli liquid separation operation.

When performing method step b), the anthranilate salt in the cell free aqueous fermentation broth produced in fermentation step a) is first protonated to anthranilic acid by reaction with the stronger acid, preferably HC1, wherein the anthranilic acid precipitates and is subsequently recovered as a solid material in method step c) ("recovery"). The anthranilic acid is then converted to aniline in a subsequent method step d) by thermal decarboxylation.

In a preferred embodiment of the method according to the invention, the conversion of o-aminobenzoate from anthranilate anion to anthranilic acid by acid protonation in method step b) can be done by adding HC1, preferably to a pH of 2.5 to 4.5, more preferably to a pH of 3 to 4, most preferably between 3.2 and

3.6.

In a further embodiment of the method according to the invention, the recovering of the anthranilic acid by precipitation in method step c) comprises filtration, thereby generating a slurry comprising said recovered anthranilic acid, wherein said slurry comprising said recovered anthranilic acid is preferably dissolved in an organic solvent. Preferably, the organic solvent used at this stage is aniline or 1 -dodecanol or a mixture thereof.

In a further embodiment of the method according to the invention, the recovering of the anthranilic acid by dissolving said anthranilic acid in an organic solvent in method step c) comprises adding said organic - i n solvent, preferably aniline or 1 -dodecanol or a mixture thereof, to said anthranilic acid such that said anthranilic acid is recovered as a solute in the organic solvent.

Preferably, the recovery step c) of the method according to the invention can be followed by washing and drying the recovered anthranilic acid precipitate in advance of performing the thermal decarboxylation of step d).

In a further embodiment of the method according to the invention, thermal decarboxylation step d) can be performed in the presence of a catalyst. Such a catalyst that is preferred can be a zeolite catalyst, wherein said zeolite catalyst preferably is zeolite H-Y (e.g. as obtained from Zeolyst International, catalog number CBV600). The acid catalyst zeolite H-Y (S1O2/AI2O3 can be between 5 and 7, preferably it is 5.5) has a particularly high acidic character and has a wider pore size (0.7-0.8 nm) than e.g. ZSM5-27 (e.g. as obtained from Clariant SuedChemie, catalog number MFI-27, S1O2/AI2O3 = 27), which also possesses a strong acidic character, but which has smaller pore size (0.5 nm) so that A A molecules cannot effectively penetrate into them and consequently do not have easy access to the active sites of the acidic catalyst, thereby reducing its effectivity.

Method step d) comprises converting the anthranilic acid of step c) to aniline by thermal decarboxylation in an organic solvent. The anthranilic acid, e.g. in the form of anthranilate crystals, with or without residual moisture, is thermally decarboxylated, preferably by feeding the anthranilic acid to a decarboxylation reactor, in which step d) can be performed. The thermal decarboxylation of step d) can be operated at a temperature between 150 °C and 250 °C, preferably between 160 °C and 220 °C, more preferably between 180 °C and 200 °C. The thermal decarboxylation of step d) can be run for sufficient time to react the anthranilic acid to aniline. Preferably, step d) can be run for 0.5 hours to 3 hours. Thermal decarboxylation step d) can be performed in the presence of an acid catalyst in order to speed up the thermal decarboxylation.

Thermal decarboxylation step d) can be run in a solvent such as water, aniline, or in 1 -dodecanol, preferably in 1 -dodecanol, or in a mixture of 1 -dodecanol and aniline.

Further, thermal decarboxylation step d) can be performed in a reactor, wherein the pressure in the reactor can be selected as a function of how much of the liquid phase is allowed to evaporate during the reaction and leave the reactor with the carbon dioxide (CO2) that is a product of the decarboxylation. Further, thermal decarboxylation step d) can produce an aniline organic solvent mixture, which can then be distilled with aniline and any water entrained or dissolved in it. Any organic solvent used in recovery step c) can be recovered as "overhead product". The solvent can then be cooled and be recycled for distillation. The overhead stream that contains aniline can then be fed to a distillation step, e.g. a heteroazeotropic distillation. in a further embodiment of the method according to the invention, the thermal decarboxylation step d) can be followed by a further step e) of purifying the aniline, preferably by distillation. In one embodiment of the method according to the invention, following recovery step c) the solution ("mother liquor") can still have up to 3-10 g/1 anthranilate. Thus, in a further embodiment of the method according to the invention, following recovery step c) residual anthranilate anion can be recovered by adsorption to an ion exchange resin or an active carbon material or a zeolite, preferably a zeolite modified with Fe ^3* or Ca ^2* , preferably a Fe-Y zeolite that can be produced by ion exchange of commercially available zeolite H-Y (e.g. as obtained from Zeolyst International - CBV600), e.g. as described in Example 6. such as Fe-Y, followed by desorption of the recovered anthranilate anion, and re- feeding of said anthranilate anion to conversion step b) for acid protonation. The desorption of the recovered anthranilate anion can preferably be carried out at a pH of 5 to 10. For example, the desorption can be performed with water having a pH or 5 to 10, e.g. water with a neutral or alkaline pH can be used. The solution after desorption can be recycled upstream of the acid addition in conversion step b) and downstream of the biomass removal after fermentation step a).

Following recovery step c) residual anthranilate anion can be recovered by adsorption to an ion exchange resin or an active carbon material or a zeolite. A preferred zeolite is modified with Fe ^3* or Ca ²⁺ , preferably a Fe-Y zeolite. Such a Fe-Y can be prepared as described in Example 6. After adsorption of residual anthranilate anion, desorption of the recovered anthranilate anion into a liquid phase follows. Such desorption of the recovered anthranilate anion can be into an organic solvent phase or in a water phase. Figure 7 shows a mass profile of AA decomposition adsorbed on the zeolite Fe-Y. Figure 8 shows the desorption of A A from the zeolite Fe-Y into the organic solvent 1 -dodecanol, which is the preferred organic solvent due to high solubility of AA in it and its low miscibility in water (0.004 g/L). Figure 9 shows the desorption of anthranilic acid (AA) from the adsorbent into the liquid phase.

It is preferred that the recovered and desorbed residual anthranilate anion is at least partially re-fed to conversion step b) for converting said anthranilate anion to anthranilic acid by acid protonation. This has the technical advantage of improving the efficiency of the method even further leading to the associated economic benefit. in a particularly preferred embodiment of the method according to the invention, subsequent to the recovery of the residual anthranilate anion by adsorption following recovery step c), a water stream devoid of the adsorbed anthranilate anion can at least be partially re- fed to the fermentation of step a). Again, this has the technical advantage of improving the efficiency of the method even further. in a preferred embodiment of the invention, steps a) through d) can be run continuously as an integrated process. In a further preferred embodiment of the invention, steps a) through e) can be run continuously as an integrated process, in such an integrated and continuous process, the fermentation of step a) can be run continuously with cell retention and continuous removal of anthranilate from the fermentation broth in the fermentation of step a), followed by continuous processing to aniline by thermal decarboxylation in the subsequent steps b) through d), or b) through e). The method according to the invention can therefore be an integrated and continuous process, which is optimized with respect to production cost. The method according to the invention therefore offers a significant technical advantages leading to a higher yield of aniline with the corresponding economic benefit, as compared to more traditional state of the art methods that run in a non- continuous fashion.

In a further embodiment of the method according to the invention, the raw material of fermentation step a) can be selected from the group consisting of sugar beet, sugar cane, starch-containing plants, preferably corn, wheat and rye, and lignocellulose, preferably straw, wood and bagasse, glycerol and CI -compounds, preferably CO. in a further embodiment of the method according to the invention, said fermentable carbon substrate of fermentation step a) can be selected from the group consisting of C-5 monosaccharides, C-6 monosaccharides, disaccharides, and tri-saccharides, wherein the C-5 monosaccharides preferably are xylose and arabinose, and wherein the C-6 monosaccharides preferably are glucose, fructose or mannose, and wherein the disaccharide preferably is saccharose, and wherein the trisaccharide preferably is kestose. in a further embodiment of the method according to the invention, said recombinant microbial host of fermentation step a) can be selected from the group consisting of bacteria, yeast and fungi, wherein said bacterium preferably is an Escherichia coli strain, a Corynebacterium strain or a Pseudomonas strain, and wherein said Corynebacterium strain preferably is Corynebacterium glutamicum, more preferably Corynebacterium glutamicum ATCC 13032, and wherein said Pseudomonas strain preferably is Pseudomonas putida, more preferably Pseudomonas putida KT2440. In essence, any of the recombinant microbial host cells of the invention described above can be used in the fermentation step a) of the method according to the invention. In a different embodiment of the invention a further method for producing aniline is provided, comprising the steps of:

a) providing o-aminobenzoate, wherein said o -aminobenzoate comprises anthranilate anion and a suitable cation,

b) converting said anthranilate anion to aniline by thermal decarboxylation in the presence or absence of a catalyst,

c) extracting the aniline produced in step b) in an organic solvent at least once, and

d) purifying the aniline produced in steps b) and c) by distillation, wherein said distillation produces aniline and a water phase. In this embodiment, said o-aminobenzoate in step a) can be provided chemically or produced biologically, preferably it is produced biologically by fermentation of a raw material comprising at least one fermentable carbon substrate using a recombinant microbial host cell capable of converting said raw material comprising a fermentable carbon substrate to o-aminobenzoate by fermentation, wherein said o- aminobenzoate comprises anthranilate anion and a suitable cation. Said fermentation of step a) can be a batch fermentation, a fed-batch fermentation or a continuous fermentation. In a preferred embodiment, step a) to step d) can be run continuously. The suitable cation of step a) can be NHLT or Na ⁺ . The recombinant microbial host of step a) can be removed prior to the subsequent conversion of said anthranilate anion to aniline by thermal decarboxylation in step b), wherein said removed recombinant microbial host preferably can be re-fed to the fermentation of step a). The catalyst used in this embodiment of the invention can be a heterogeneous acid catalyst, preferably a zeolite, most preferably zeolite H-Y. However, said catalyst can also be a heterogeneous base catalyst, preferably a layered double hydroxide, most preferably Mg-Al hydrotalcite.

In a preferred embodiment of the further method according to the invention, the extraction of aniline in an organic solvent in step c) can be performed for more than one time for a further pre-concentration of aniline in advance of distillation.

In a preferred embodiment of the further method according to the invention, the method can comprise recovering the organic solvent used in the extraction of step c), wherein said recovering can preferably be done by distillation, wherein said recovered organic solvent preferably can be re-fed to step c) to be reused again for extracting the aniline produced in step b). Said organic solvent can be selected from the group consisting of alcohols, phenols, amides, ethers and aromatic hydrocarbons, wherein said alcohol preferably is 1 -dodecanol. In another embodiment, this further method of the invention can comprise a further step e) of re-feeding the water-phase of the extraction performed in step c) and/or of re- feeding the water-phase of the distillation performed in step d) to the fermentation of step a). The LLf cation can be recovered as NH-. subsequent to the distillation of step d) and can be re-fed to the fermentation of step a).

Lastly, the invention also provides the use of the aniline produced according to the method of the invention, wherein the aniline produced in step d) and/or e) is further converted to methy lenedi aniline (MDA) with formaldehyde in the presence of water and catalyst. The MDA produced in this way can be further converted to methylenediisocyanate (MDI) with phosgene.

It will be apparent to those skilled in the art that various modifications can be made to the methods and recombinant host strains of the invention. Thus, it is intended that the present invention covers such modifications and variations, provided they come within the scope of the appended claims and their equivalents.

Figures and Tables List of figures

Figure I shows a schematic overview display of biosynthetic pathway from glucose to oAB (anthranilate) in C. glutamicum showing the strategy to increase the precursor supply from the central metabolism (PTS: phosphotrans f eras e system; PEP: phosphoenolpyruvate; TCA: tricarbonic acid; IolT2: inositolpermease unit T2; PorA/H: porine complex of PorA and PorH; MFS: mayor facilating system) Thick arrows display biosynthetic steps that are enhanced and crosses mark biosynthetic steps that are interrupted.

Figure 2 Schematic overview of the central carbon metabolism in C. glutamicum. Shown are the pho sphotrans f erase system (PTS) for glucose uptake, the glycolysis, the pentose phosphate pathway and relevant anaplerotic reactions. The TCA (tricarbonic acid)-cycle and the shikimic acid pathway are indicated to explain the positioning and transitions into these pathways (GAPDH = glyceraldehyde-3 - phosphate dehydrogenase; G6PDH = glucose-6-phosphate dehydrogenase; 6PGD = phosphogluconate dehydrogenase; PEP = phosphoenolpyruvate).

Figure 3 Schematic overview of the shikimic acid pathway in C. glutamicum from precursor molecules to the aromatic amino acids. The structural formulas of central intermediates are given.

Figure 4 shows the genetic background of aromatic amino acid biosynthesis in P. putida KT2440. Performed genetic manipulations are highlighted in grey. Figure 5 shows the overall concept of the method according to the invention comprising the conversion of raw materials to anthranilate in the fermentation step followed by a chemical conversion and purification to aniline in the downstream processing.

Figure 6 shows the overall concept of the method according to the invention in more detail. Both of NaOH and N¾ can be used as a buffer in the fermenter.

Figure 7 shows a mass profile of anthranilic acid decomposition adsorbed on Fe-Y in the temperature range of 100-550 °C. Figure 8 shows the desorption of anthranilic acid from Fe-Y into 1 -dodecanol. The experiment was performed by suspending 0.2 g Fe-Y containing 10.8 mg anthranilic acid in 2 mL 1 -dodecanol. The slurry was stirred for 0.5 h at the temperature range of 25-120 °C. The desorption results are shown in Figure 2 in which maximum 27.8% anthranilic acid could be desorbed into 1 -dodecanol phase at 120 °C.

Figure 9 shows the desorption of anthranilic acid from adsorbent into the liquid phase. The desorption test of anthranilic acid from Fe-Y into water showed that the adsorption of anthranilic acid by metal- exchanged zeolite in aqueous solution is reversible. The desorption of anthranilic acid into an organic solvent was tested. 1 -dodecanol was selected as organic solvent due to high solubility of anthranilic acid in it and also its very low miscibility in water (0.004 g/L).

Figure 10 shows the decomposition of anthranilic acid in organic media with a catalyst. The kinetics of decarboxylation of anthranilic acid dissolved 3 wt% in 1 -dodecanol in the presence of zeolite Y at 160 °C and 180 °C is shown. Anthranilic Acid (3 wt%) was dissolved in 1 -dodecanol. At this concentration the acid was perfectly soluble in the organic solvent even at room temperature. 80 mL of the above solutions are then transferred into an autoclave of 160 mL, 1.5 g of zeolite catalyst are added and heated to 160 °C or 180 °C. For comparison a blank test without catalyst (Blank) and with an alkaline hydrotalcite (HTC), H-ZSM5, sodium doped zeolite Y (NaY), zeolite Y ammonium form (NFLt-Y) are also tested. Samples were taken at different time intervals and analyzed by HPLC methods to determine the rate of aniline formation.

Figure 11 shows the optical density at 600 nm (ODeoo) followed over 33 h (triple determination) of 100-mL-shake flask cultures at pH 5 (with MES buffer) of the strains Escherichia coli, Bacillus subtilis, Corynebacterium glutamicum, Pseudomonas putida, and Saccharomyces cerevisiae. C. glutamicum was additionally cultivated at pH 7 (supplemented with MOPS buffer).

Figure 12 shows the optical density at 600 nm (ΟΟβοο) followed over 25 h (double determination) of 100-mL-shake flask cultures at pH 7 (with MOPS buffer) of the strains Escherichia coli, Bacillus subtilis, Corynebacterium glutamicum, Pseudomonas putida, and Saccharomyces cerevisiae.

Figure 13 shows the optical density at 600 nm (ODeoo) followed over 25 h (double determination) of 100-mL-shake flask cultures at pH 7 (with MOPS buffer) supplemented with 3 g/L oAB of the strains Escherichia coli, Bacillus subtilis, Corynebacterium glutamicum, Pseudomonas putida, and Saccharomyces cerevisiae.

Figure 14 shows the optical density at 600 nm (ODeoo) followed over 25 h (double determination) of 100-mL-shake flask cultures at pll 7 (with MOPS buffer) not supplemented (black) or supplemented with 3 g/L oAB (grey) of the strain Corynebacterium glutamicum.

Figure 15 shows the optical density at 600 nm (ΟΟβοο) followed over 25 h (double determination) of 100-mL-shake flask cultures at pH 7 (with MOPS buffer) not supplemented (black) or supplemented with 3 g L oAB (grey) of the strain Pseudomonas putida. Figure 16 shows the optical density at 600 nm (ΟΟβοο) followed over 27 h (double determination) of 100-mL-shake flask cultures at pH 7 (with MOPS buffer) supplemented with 3 g/L pAB of the strains Escherichia coli, Bacillus subtilis, Corynebacterium glutamicum, Pseudomonas putida, and Saccharomyces cerevisiae. Fignre 17 shows the optical density at 600 nm (ΟΟβοο) followed over 25 h (double determination) of 100-mL-shake flask cultures at pH 7 (with MOPS buffer) supplemented with 0.1 g/L octanol (duplicates in black and grey) of the strain Corynebacterium glutamicum. Figure 18 shows the optical density at 600 nm (ODeoo) followed over 25 h (double determination) of 100-mL-shake flask cultures at pH 7 (with MOPS buffer) supplemented with 0.1 g/L octanol (duplicates in black and grey) of the strain Pseudomonas putida.

Figure 19 shows the optical density at 600 nm (ΟΟβοο) followed over 25 h of 100-mL-shake flask cultures at pH 7 of Corynebacterium glutamicum (supplemented as denoted in the graph; duplicates).

Figure 20 shows the optical density at 600 nm (ODeoo) followed over 25 h of 100-mL-shake flask cultures at pH 5 of Corynebacterium glutamicum (supplemented as denoted in the graph; duplicates). Figure 2 1 shows the optical density at 600 nm (ODeoo) followed over 28 h of 100-mL-shake flask cultures at pH 7 (MOPS) and pH 5 (MES) of Co rynebacterium glutamicum (supplemented as denoted in the graph; duplicates).

Figure 22 shows the optical density at 600 nm (ODeoo) followed over 50 h of fermentation cultures at pH 7 of Corynebacterium glutamicum supplemented after 3, 4, and 5 h to a final concentration of 0 g/L supplement (square), 7 g/L oAB (circle), 7 g/L pAB (cross), 40 mg/L dodecanol (triangle).

Figure 23 shows the HPLC - chromatogram at 254 nm of culture supernatants of fermentation of Corynebacterium glutamicum with 7 g/L oAB after 2 h (before supplementation with oAB), 5 h (after supplementation with 7 g/L oAB), 25 h, and 49.5 h. Plus oAB standard.

Figure 24 shows the optical density at 600 nm (ODeoo) followed over 58 h of fermentation cultures at pH 7 of Corynebacterium glutamicum supplemented after 3, 4, 5, 6, and 7 h to a final concentration of 0 g/L supplement (squares), 15 g/L oAB (triangles), 35 g/L oAB (crosses), and 80 g/L oAB (circles).

Figure 25 shows the optical density at 600 nm (ODeoo) followed over 58 h of fermentation cultures at pH 7 of Corynebacterium glutamicum supplemented after 3, 4, 5, 6, and 7 h to a final concentration of 35 g/L oAB. Figure 26 shows the optical density at 600 nm (ODeoo) followed over 58 h of fermentation cultures at pH 7 of Corynebacterium glutamicum supplemented after 3, 4, 5, 6, and 7 h to a final concentration of

80 g/L oAB. Figure 27 shows a scheme of the knockout procedure based on vector system pEMG.

Figure 28 shows the genetic map of the P. putida KT2440 trpEGDC cluster (above) and genetic map of the P. putida KT2440 trpDC deletion (below) (gph: phosphoglycolate phosphatase (partial); trpE: anthranilate synthase component I; GDSL: (J DSL family lipase (unrelated to L-tryptophane biosynthesis); hyp: hypothetical protein; trpG: anthranilate synthase component II; erg: cAMP- regulatory protein (partial); TS 1 : deletion flank 1 ; TS2: deletion flank 2).

Figure 29 shows the genetic map of the P. putida T2440 serCpheAtyrA cluster (above) and genetic map of the indented P. putida KT2440 pheA deletion (below) (gyrA: ON A gyrase sub unit A(partial); serC: phosphoserine aminotransferase; pheA: chorismate mutase / prephenate dehydratase; tyrA: prephenate dehydrogenase / 3 -phosphoshikimate 1 - carboxyvinyltransferase; cmk: cytidylate kinase; TS 1 : deletion flank 1 ; TS2: deletion flank 2).

Figure 30 shows the integration of a cross-flow ultrafiltration module (cut of value of 750 kDa) for continuous fermentations with cell retention in a 1 L lab scale bioreactor.

Figure 31 shows the integration of a centrifuge (Centritech Lab I I I ) for continuous fermentations with cell retention in a 1 L lab scale bioreactor. Figure 32 shows the fermentation of C. glutamicum AtrpD for the production of oAB during batch- and continuous mode using a culture volume of 1 L, a dilution rate 0.05 L/h, cultivation temperature 30 °C at pH 7 controlled with 1 M NH4OH, air flow rate of 0.2 L/min for aeration, pCh controlled at 30 % air saturation by adjusting the stirrer speed between 200 and 1200 rpm. Cell retention achieved by cross-flow ultrafiltration.

Figure 33 shows the fermentation of C. glutamicum AtrpD::trpD5Agpi for the production of oAB during batch- and continuous mode using a culture volume of 1 L, a dilution rate 0.01 L h, cultivation temperature 30 °C at pH 7 controlled with 1 M NH..OH and 1 M HCI, air flow rate of 0.2 L/min for aeration, p(¾ controlled at 30 % air saturation by adjusting the stirrer speed between 200 and 1400 rpm. Cell retention achieved by centrifugation.

Figure 34 shows the fermentation of C. glutamicum AirpD for the continuous production of oAB using a culture volume of 1 L, a dilution rate of 0.025 L h or 0.05 L/h as stated in the diagram, cultivation temperature 30 °C. pH 7 controlled with 1 M NHiOH, air flow rate of 0.2 L/min for aeration, pCh controlled at 30 % air saturation by adjusting the stirrer speed between 200 and 1200 rpm. Cell retention achieved by cross-flow ultrafiltration. Figure 35 shows the decomposition of 2-aminobenzoic acid (anthranilic acid, AA) in aniline with a catalyst. The kinetics of dec arboxy lation of anthranilic acid (AA) dissolved 40 wt% in aniline in the presence of zeolite Y at 160 °C, 180 °C and 200°C is shown.

Figure 36 shows the decomposition of 2-aminobenzoic acid (anthranilic acid, AA) in aniline with a catalyst. The kinetics of decarboxylation of anthranilic acid (AA) dissolved 40 wt% in aniline and in 10% Water-90% Aniline in the presence of zeolite Y at 200°C is shown.

Figure 37 shows the decomposition of 2-aminobenzoic acid (anthranilic acid, AA) in aniline with a catalyst. The kinetics of dec arboxy lation of anthranilic acid (AA) dissolved 40 wt% in aniline in the presence of zeolite Y at 180°C is shown.

Figure 38 shows the decomposition of 2-aminobenzoic acid (anthranilic acid, AA) in aniline with a catalyst. The kinetics of dec arboxy lation of anthranilic acid (AA) dissolved 40 wt% in aniline in the presence of zeolite Y at 180°C is shown.

Figure 39 shows the I R Spectrum of a freshly prepared H-Y catalyst and one after reaction as above using AA isolated from hydrolyzed corn starch. The typical bands for aniline are marked. Absorption of this higly basic species takes place. The OH region is also highlighted. A reduction of H signal is compatible with some partial ion exchange that could take place with trace elements present, however the OH groups seam not be active in the catalysis, but is probably the A l-O-Si bridge the responsible. List of tables

Table 1 shows vectors and plasmids used and/or generated in this study.

Table 2 shows bacterial strains used and/or generated in this study.

Table 3 shows primers used in this study.

Table 4 shows characteristics towards oAB production of bacterial strains used and/or generated in this study (CDW: cell dry weight; Y: yield; μ„^'. maximal growth rate; STY: space time yield).

Table 5 shows biochemical characteristics of C. glutamicum AtrpD raw extracts producing different variants of TrpEG towards TrpEG activity.

Table 6 shows the adsorption of AA 0.5% in water by HAP. zeolite Y (G0257 and G055) and ZSM5 (ZSM5-27 and ZSM5-55), as described in Example 6.

Table 7 shows the A A adsorption capacities of metal-exchanged zeolite G0257 in g AA/ kg adsorbent in distilled water and buffered aqueous solution after 10 and 60 minutes, as described in Example 6.

Examples

Example 1 - Production of o-aminobenzoate with E. coli

The strain E. coli W31 10 trpD9923 (Escherichia coli::trpD9923; Table 2) was purchased from the E. coli Genetic Resource Center at Yale University. The strain had been created by random mutagenesis and contained a mutated trpD gene called trpD9923. An inactivation of the enzyme was achieved by a random point mutation (( i->T) in the related gene, resulting in a nonsense mutation in the transferase activity encoding region. As the result only seven amino acids of the original transferase domain are translated (Ikeda M, Appl Microbiol Biotechnol, 2006, 69:615). The related truncated enzyme of the trpD9923 gene looses its ability to catalyse the reaction of anthranilate phosphoribosyl transferase, but maintains its anthranilate synthase activity. The strain can therefore synthesize anthranilate, but cannot metabolize it further to L -tryptophan and is thus L -tryptophan auxotroph. This leads to an accumulation of anthranilate.

The strain was grown in 50 mL shake flasks with a 10 mL culture volume at 28 °C at 140 rpm. The medium used was the mineral medium M9 (1 g/L (NELt Cl, 0.5 g/L NaCl, 0.05 g/L thiamin, 1 g/L ΚΉ2ΡΟ4, 1 g/L K2HPO4, 0.247 g/L MgSO H-O (Merck, Darmstadt), 0.015 g/L CaCl ₂ , and 10 g/L glucose. The pH was adjusted to 7 with 5 mol L sodium hydroxide solution.) with 20 mg/L L- tryptophan. The strain produced 60 mg/L anthranilate after 25.5 h as measured by HPLC-DAD (254 nm) (Table 4 and Example 3). The strain was further optimized by inactivating the phosphotransferase system (PTS) using knockout deletion. Accordingly, a the resulting PTS-deficient strain Escherichia coli: :trpD9923Ahpr (Table 2) was generated and adapted to growth on glucose and tested for anthranilate production using a 25 mL shake flask fermentation at 37 °C at 150 rpm with a culture volume of 10 mL. The same medium as for the pts positive strain was used. It produced 69 mg/L of anthranilate after 25 hours as measured by HPLC-DAD (254 nm) (Table 4 and Example 3).

Example 2 - Strain selection for developing a microbial strain that produces o-aminobenzoate Several strains commonly used industrial biotechnology were investigated regarding their natural tolerance towards oAB and to organic solvents in question for later product extraction (e.g. 1 - dodecanol or octanol). Furthermore, the influence of the used pH value was investigated. It was known from previous experiments that at least for E. coli the toxicity of aromatic compounds depends strongly on the pH value, probably because protonated acids can much more efficiently penetrate the cell membrane and accordingly their toxicity rises severely. The extraction of oAB in an organic phase is supposed to work best at low pH values, close to the isoelectric point of oAB. The most relevant host strains considered were: Escherichia coli K12, Corynebacterium glutamicum DSM 20300 (= ATCC 13032), Pseudomonas putida KT2440, Bacillus subtilis subsp. 168, Saccharomyces cerevisiae DSM 70449, and Pichia pastoris X33 (DSM 70382).

To investigate the chosen strains towards the mentioned characteristics shake flasks and small scale fermentation experiments were performed and the organisms were supplemented with the toxic substances (aminobenzoates, solvents) at differed concentrations, different pi I values, and the behaviour of the strains was studied.

The organisms (commercially obtained from ordered from the DSMZ, German collection of Microorganisms and cell cultures, Braunschweig) were firstly cultivated in shake flasks and in each of them in suiting, published standard minimal medium in order to characterize their growth under standard cultivation conditions (C. glutamicum: Preculture: BHI medium (37 g/L; Brain-Heart- Infusion; Becton Dickenson and Company, Heidelberg); Main culture pH 7: CGXII-MOPS medium (42 g/L MOPS buffer, 20 g/L (NH ₄ ) ₂ S0 ₄ , 5 g/L urea (Fisher Scientific, Schwerte), 1 g/L KH ₂ P0 ₄ , 1 g/L K2HPO4, 0.25 g/L MgS0 ₄ -7H ₂ 0 (Merck, Darmstadt), 0.01 g/L CaCl ₂ , and 10 g/L glucose (autoclaved separately). The pH was adjusted to 7 with 5 mol/L sodium hydroxide solution. The following components were added after sterile filtration: 2 mg/'L biotin, 0.01 g/L MnS0 ₄ -H ₂ 0 (Merck, Darmstadt), 0.01 g/L FeS0 ₄ -7H ₂ 0 (Merck, Darmstadt), 1 mg/L ZnS0 ₄ -7H ₂ 0, 0.2 mg/L CuS0 ₄ -5H ₂ 0 (Merck, Darmstadt), 0.02 mg/L NiCl ₂ -6H ₂ 0 (Merck, Darmstadt), and 0.03 g/L 3.4 -dihydroxybenzoic acid (Acros Organics, Nidderau); Main culture pH 5: CGXII-MES medium (39 g L MES buffer, 20 g/L (NH ₄ ) ₂ S0 ₄ , 5 g/L urea (Fisher Scientific, Schwerte), 1 g/L KH ₂ P0 ₄ , 1 g/L K ₂ HP0 ₄ , 0.25 g/L MgS0 ₄ -7H ₂ 0 (Merck, Darmstadt), 0.01 g/L CaCl _¾ and 10 g/L glucose (autoclaved separately). The pH was adjusted to 5 with 5 mol/L sodium hydroxide solution. The following components were added after sterile filtration: 2 mg/L biotin, 0.01 g/L MnS0 ₄ -H ₂ 0 (Merck, Darmstadt), 0.01 g/L FeS0 ₄ -7H ₂ 0 (Merck, Darmstadt), 1 mg/L ZnS0 -7H ₂ 0, 0.2 mg/L CuS0 -5H ₂ 0 (Merck, Darmstadt), 0.02 mg/L NiCl ₂ -6H ₂ 0 (Merck, Darmstadt), and 0.03 g/L 3.4-dihydroxybenzoic acid (Acros Organics, Nidderau); E. coli: Preculture: LB medium (Luria-Bertani; Roth, Karlsruhe); Main culture pH 7: M9 medium (1 g/L (NH,) ₂ C1, 0.5 g/L NaCl, 0.05 g/L Thiamin chloride, 1 g/L KH -PO.,. 1 g/L K ₂ HP0 ₄ , 0.247 g/L MgS0 ₄ -7H ₂ 0 (Merck, Darmstadt), 0.015 g/L CaCl ₂ , and 10 g/L glucose. The pH was adjusted to 7 with 5 mol/L sodium hydroxide solution.); Main culture pH 5 : M9-MES medium (19.5 g/L MES buffer, 1 g/L (Nft^Cl, 0.5 g/L NaCl, 0.05 g/L Thiamin chloride, 1 g/L K.H. PO4. 1 g/L K.H O4. 0.247 g/L MgS0 ₄ -7H ₂ 0 (Merck, Darmstadt), 0.015 g/L CaCl ₂ , and 10 g/L glucose. The pH was adjusted to 7 with 5 mol/L sodium hydroxide solution.); B. subtilis: Preculture: LB medium; Main culture pH 7: M9- SL medium (1 g/L (Ν¾)£1, 0.5 g/L NaCl, 0.05 g/L thiamin, 1 g/L KH ₂ P0 ₄ , 1 g/L K ₂ HP0 ₄ , 0.247 g/L MgS0 ₄ -7H ₂ 0, 0.015 g/L CaCl ₂ , 1.0 mg/L MnC! H -O, 1.35 mg/L FeCl-6H ₂ 0, 1.7 mg/L ZnS0 ₄ -7H ₂ 0, 0.04 mg/L CuCl-2H ₂ 0, 0.06 mg/L CoCl ₂ -6H ₂ 0, 0.06 mg/L Na ₂ Mo0 ₄ -2H ₂ 0, and 10 g/L glucose. The pH was adjusted to 7 with 5 mol L sodium hydroxide solution.); Main culture pH 5 : M9-SL-MES medium (19.5 g/L MES buffer, 1 g/L (NH ₄ ) ₂ C1, 0.5 g/L NaCl, 0.05 g/L thiamin, 1 g/L KH ₂ P0 ₄ , 1 g/L K. HPO4. 0.247 g/L MgS0 ₄ -7H ₂ 0, 0.015 g/L CaCl _¾ 1.0 mg/L MnCl H ₂ 0, 1.35 mg/L FeCl-6H ₂ 0, 1.7 mg/L ZnS0 ₄ -7H ₂ 0, 0.04 mg/L CuC1 2H ₂ 0, 0.06 mg/L CoCl ₂ -6H ₂ 0, 0.06 mg/L Na ₂ Mo0 ₄ -2H ₂ 0, and 10 g/L glucose. The pH was adjusted to 5 with 5 mol/L sodium hydroxide solution.); P. putida: Preculture: LB medium; Main culture pH 7: Brunner medium (0.5 g/L (NFLt) ₂ S0 ₄ , 1 g/L KH.-PO4. 1 g/L K. HPO4. 0.2 g/L MgS0 ₄ -7H ₂ 0, 0.05 g/L CaCl ₂ , 5.0 mg/L EDTA, 2.0 mg/L FeS0 ₄ -7H ₂ 0, 0.03 mg ^/ L MnCl H ₂ 0, 0.1 mg/L ZnS0 ₄ -7H ₂ 0, 0.01 mg/L CuC1 2H ₂ 0, 0.2 mg/L CoCl ₂ -6H ₂ 0, 0.03 mg/L Na ₂ Mo0 ₄ -2H ₂ 0, 0.3 mg/L H3BO3, 0.02 mg/L NiCl ₂ -6H ₂ 0, and 10 g/L glucose. The pH was adjusted to 7 with 5 mol/L sodium hydroxide solution.); Main culture pH 5: Brunner-MES medium (19.5 g/L MES buffer, 0.5 g/L (NH4) ₂ S0 ₄ , 1 g/L H ₂ P0 ₄ , 1 g/L K ₂ HP0 ₄ , 0.2 g/L MgS0 ₄ -7H ₂ 0, 0.05 g/L CaCl ₂ , 5.0 mg/L EDTA, 2.0 mg/L FeS0 ₄ -7H ₂ 0, 0.03 mg/L MnC! H O, 0.1 mg/L ZnS0 ₄ -7H ₂ 0, 0.01 mg/L CuCl-2H ₂ 0, 0.2 mg/L CoCl ₂ -6H ₂ 0, 0.03 mg/L Na ₂ Mo0 ₄ -2H ₂ 0, 0.3 mg/L H3BO3, 0.02 mg/L NiCl ₂ -6H ₂ 0, and 10 g L glucose. The pH was adjusted to 5 with 5 mol/L sodium hydroxide solution.); S. cerevisiae: Preculture: YPD medium (10 g/L yeast extract, 20 g/L peptone, 20 g/L glucose); Main culture pH 5 : SCM medium (5 g/L (NH4) ₂ S0 ₄ , 3 g/L KH ₂ P0 ₄ , 2.5 g/L MgS0 ₄ -7H ₂ 0, 0.5 g/L NaCl, 20 mg/L inositol, 5 mg/L thiamin, 1.32 mg/L riboflavin, 17.8 mg/L calcium pantothenat, 18.4 mg/L nicotinic acid, 5.16 mg/L pyridoxine-HCl, 0.18 mg/L biotin, 0.12 mg/L folic acid, 40.6 mg/L FeCl ₃ -6H ₂ 0, 6.0 mg/L ZnCl ₂ -4H ₂ 0, 3.0 mg/L CaCl ₂ -2H ₂ 0, 5.76 mg/L CuS0 ₄ -5H ₂ 0, 6.0 mg/L CoCl ₂ -6H ₂ 0, 6.0 mg/L Na ₂ Mo0 ₄ -2H ₂ 0, 1.56 mg/L H3BO3, and 5 g/L glucose. The pH was adjusted to 5 with 5 mol L sodium hydroxide solution.) (Sambrock J; Fritsch, EF, and Maniatis, T: Molecular Cloning - a laboratory manual. Cold Spring Larbor Laboratory Press; 1989; New York; Harwood R. Cutting SM: Molecular biological methods for Bacillus. Chichester, England: John Wiley & Sons Ltd; 1990; Keilhauer, C, Eggeling, L, Sahm, H: .1 Bacterid, 1993, 175(17): 5595). These experiments were performed with a medium pH of 7 and pH 5 to evaluate the difference in toxicity of aminobenzoate at low pH values. The yeast strains were only investigated at pH values below 7 (5.5 and 3.5). The different pH values were stabilized by addition of different buffer systems (MOPS, MES).

100-mL-shake flask cultures of the strains (pH 5 (with MES buffer) were incubated for 33 h and the growth was followed by measurement of the optical density at 600 nm (ODeoo) over the time (triple determination). C. giutamicum was additionally cultivated at pH 7 (with MOPS buffer). Figure 11 shows the results.

The collected data showed that C. giutamicum exhibits the best growth at pH 5 of all investigated organisms. E. coli, P. putida, and S. cerevisiae did grow extremely weak at pH 5. Nevertheless, in some of the cultures, due to metabolic activity the pH dropped additionally below 4. To exclude growth effects due to metabolic activities the experiments were performed in fermenters with automatic pH regulation. Already visible was, as expected, that cultivation at low pH decreases the growth rate of all tested organisms. However, to compare the growth of all considered organisms under standard cultivation conditions at pH 7 the experiment was repeated with all minimal media supplemented with MOPS buffer. Additionally, the analogously cultures were supplemented with oAB (final concentration of 3 g/L, added in the beginning of the exponential growth phase in three portions (after 2, 3, 4 h incubation time)). 100-mL-shake flask cultures of the strains at pH 7 with or without supplementation with 3 g/L oAB were incubated for 25 h and the growth was followed by measurement of the optical density at 600 nm (ΟΟβοο) over the time (double determination) (Figure 12 and Figure 13). C. glutamicum reached the highest ODeoo values under the chosen cultivation conditions and furthermore its growth was not inhibited by the addition of 3 g/L oAB, whereas the growth of other organisms, like for example B. subtilis was at 3 g/L oAB already strongly decreased. From these results C. glutamicum was identified as the preferred candidate for an oAB production at pH 7 (Figure 14 and Figure 15).

In order to investigate the resistance of the considered strains against para -aminobenzoate (pAB) and solvents (octanol and 1 -dodecanol), the strains were cultivated at pH 7 (S. cerevisiae at pH 5.5) as described before and supplemented with 3 g/L pAB (Figure 16). The growth of C. glutamicum, E. coli, P. putida, and P. pastoris was not inhibited by pAB under these conditions, whereas the growth of S. cerevisiae and B. subtilis was severely reduced. The toxicity of the extraction solvent octanol was investigated. The solubility of octanol in water is about 0.1 g/L, therefore cultures were supplemented with such an octanol amount and in addition with 3 g/L. It was shown that already at 0.1 g/L but even clearer at 3 g/L octanol all organisms were unable to grow. Only P. putida has shown minor growth activity after octanol addition (Figure 17 and Figure 18).

As in all performed experiments C. glutamicum had shown the highest cell densities and best resistance capabilities, this strain was analogously cultivated with 5 g/L oAB and 5 g/L pAB at pH 5 and pH 7, respectively. No growth inhibition was detected under these conditions (Figure 19 and Figure 20 ). At pH 5, the cell density is halved, resulting in the conclusion that the fermentation should more preferably be performed at pH 7.

Dodecanol was investigated as an alternative extraction solvent. Dodecanol exhibits solubility in water of about 40 mg/L. C. glutamicum was cultivated as described above at pH 5 and 7 under supplementation with 40 mg/L 1 -dodecanol. Dodecanol did not influence the growth of the strain under these conditions (see Figure 19 and Figure 20).

The results obtained in the shake flask experiments with C. glutamicum were reproduced in small-scale fermenters (Figure 21 ). To further characterize the resistance of C. glutamicum against aminobenzoates and 1 -dodecanol stress at pH 7 small-scale fermentations of the organism were performed using a 1 L-4fo Id- fermentation unit (HiTecZang) with regulated pH, agitation, oxygen supply, temperature, and under glucose feeding. The glucose-fed allowed the strain to grow to higher cell densities.

A fermentation experiment was performed with C. glutamicum in minimal medium pH 7 (as described above but without addition of MOPS, biotin, protocathecuat and urea) and supplementation of one fermenter with nothing (positive control), one fermenter with 7 g/L oAB, one fermenter with 7 g/L pAB and one fermenter with 40 mg/L 1 -dodecanol. The resulting growth curves show that the growth of C. glutamicum is not influenced by the added 1 -dodecanol. This confirmed the results obtained in shake flasks (see Figure 22). Supplementation with oAB and pAB, respectively, resulted in a prolonged lag-phase but finally same cell densities were reached (see Figure 22 ). It was concluded that 1 -dodecanol is a suitable oAB extraction solvent.

The concentration of aminobenzoate was further increased to find the border at which it significantly inhibits the growth of C. glutamicum. It had to be excluded that the aminobenzoates were degraded by the organism and therefor the growth is restored after the prolonged lag-phase. To investigate the aminobenzoate stability in the fermenters culture supernatant samples were collected during the fermentation and analyzed by HPLC-DAD (254 nm) (see Figure 23 and Example 3). The aminobenzoate was not significantly degraded during cultivation.

In a further experiment the concentration of oAB was increased to investigate the effect on the C. glutamicum strain. The four fermenters were hereto supplemented with 0 g/L, 15 g/L, 35 g/L and 80 g/L of oAB. To achieve solubility of the acid in such high concentrations at pH 7 the acid was titrated with ammonium hydroxide to form the corresponding ammonium salt of oAB in stock solutions later added to the fermenters in five portions over 5 h (after cultivation of 3, 4, 5, 6, and 7 h). The rest of the cultivation conditions was kept as described above. The addition of 80 g/L oAB led to a dying (lysis) of the cells over 54 hours and 35 g/L oAB resulted in a growth inhibition (see Figure 24. Figure 25, and Figure 26). Cultures supplemented with 15 g/L oAB have shown an extensive lag-phase over 34- 48 h before growth was recovered and cells entered the exponential phase (see Figure 24 ). An oAB degradation during fermentation was again excluded by HPLC measurements of culture supernatants (data not shown).

Taken together it was concluded that the production of oAB should most preferably be done with a Corynebacterium glutamicum ATCC 13032 strain. Furthermore it was concluded that the process fermentation should preferably be performed at pH 5-7, and more preferably at pH 7 and the extraction (if needed) should preferably be conducted with 1 -dodecanol as extraction solvent. Octanol was not a suitable extraction solvent. Example 3 - Production of o-aminobenzoate with C. glutamicum

A bacterial strain that produces anthranilate in a g/L scale was developed. This work was based on genetic manipulations of tryptophan producers (e.g. in E. coli and C. glutamicum) and included the development and implementation of a strategy to genetically manipulate central metabolism, common aromatic biosynthesis and the L-tryptophan branch pathway in order to gain an anthranilate (oAB) producer. Based on the above described results, the metabolic engineering was preferably based on Corynebacterium glutamicum ATCC 13032. General cultivation of Escherichia coli DHSa based strains

If not denoted differently all chemicals were acquired from Sigma-Aldrich (Sternheim) and all applied enzymes from New England Biolabs (Schwalbach). E. coli strains were cultivated under sterile conditions in LB medium (Luria-Bertani; Roth, Karlsruhe) or on LB-agar (Abcr GmbH & Co. Kg, Karlsruhe) plates at 37 °C. Selective cultivation was achieved by adding antibiotics (100 mg/L ampicillin sodium salt; 50/25 mg L kanamycin sulfate; 100 mg/L spectinomycin sulfate). For plasmid preparation the stains have been cultivated in 15 mL-tubes for 14 h-16 h with constant shaking at 200 rpm in 3 mL LB medium and a selecting antibiotic at 37 °C (Kuhner Shaker ISF-4-W; Adolf Kuhner AG, Basel (Switzerland)). General molecular biology methods

PGR reactions were generally performed using Platinum ^® Taq DNA Polymerase (5 U/μΕ; Life technologies, Darmstadt) with ca. 50 ng PGR template and 10 pmol of the responding primer pair (see Table 3). PGR conditions: 98 °C 5 min, 98 °C 30 sec, 56 °C 30 sec, 72 °C I min/kb, 3 Ox, 16 °C hold (Mastercycler® nexus - Cycler; Eppendorf, Hamburg). Plasmid DNA purification was done using the NucleoSpin ^® Plasmid Pure kit (Macherey & Nagel, Duren) following the manufacturer's instructions. Plasmid DNA digestion was done in a 20 μΕ scales with 5 U of the related restriction enzyme(s) and ca. 1 μg plasmid DNA in buffer, recommend by the enzyme supplier, and incubated for about 2 h at 37 °C and after that analyzed by agarose gel electrophoresis. Agarose gels were made with 1 % agarose (Abcr GmbH & Co. Kg, Karlsruhe) dissolved in lxTAE electrophoresis buffer (Promega, Fitchburg USA). The supplied electric field was 90 V-120 V (EPS300; Pharmacia Biotech; VWR International GmbH, Darmstadt) and the DNA was detected via ethidium bromide (0.3-0.5 mg/L) or Midori Green (Midori Green Advance DNA Strain; N IPPON Genetics EUROPE GmbH, Duren). Gel documentation: Gene Genius®; VWR; Darmstadt. For the agarose gel DNA extraction the NucleoSpin ^® Gel and PGR Clean-up (Macherey & Nagel, Duren) kit was used as specified by the manufacturer. For DNA ligation a T4-DNA ligase (20 U) was used in the buffer supplied by the manufacturer. The used insert to plasmid mass ratio was 4: 1. The mixture was incubated for ca. 15 h at 16 °C and afterwards incubated at 65 °C for 10 min. Plasmids and ligation reactions were introduced into electro-competent Escherichia coli DH5a cells (ElectroMAX™ DH5a-E™ Competent Cells; Lfe Technologies, Darmstadt). After electro -trans formation (conditions: -20 ms exponentially decaying pulse, 2.5 kV/cm, 25 F, 200 Ω) in 0.2 cm gap electroporation cuvettes (BioRad, Hercules, CA) using a Gene Pulser Xcell System (BioRad, Hercules, CA), 800 xL LB recovery medium was immediately added and the suspension was transferred into 1.5 rriL microcentrifuge tube. After 1 h at 37 °C, a cell suspension were spread onto LB agar plates, supplemented with appropriate antibiotics, and incubated at 37 °C over night. Clones were collected and correct transformation confirmed via restriction analysis, colony PCR and/or sequencing. For colony PCR analysis of colonies a colony was picked with a sterile toothpick, transferred to a separate plate (master plate), dissolved in 1 \il DM SO and boiled for 10 min at 98 °C, before being added into a standard PCR mixture. The correct cloning of all plasmids and generation of the related mutants was proven by PCR and/or ON A sequencing. General cultivation of Coryncbactcrium glutamicum A TCC 13032 based strains

C. glutamicum strains were cultivated under sterile conditions in BH I medium (37 g/L; Brain-Heart- Infusion; Bee ton Dickenson and Company, Heidelberg) tubes, shake flasks or on agar plates at 30 °C. Selective cultivation was achieved by adding antibiotics (15 mg/L kanamycin sulfate; 100 mg/L spectinomycin sulfate).

For strain characterizations the first pre- culture was started from an isolated colony and cultivated for 10 h with constant shaking at 400 rpm in 4 mL BHI medium and, depending on the strain, supplemented with a suiting antibiotic or aromatic amino acids. The second pre-culture was grown in 50 mL CGXIi-MOPS medium (42 g/L MOPS buffer, 20 g/L (NH ₄ ) ₂ S0 ₄ , 5 g/L urea (Fisher Scientific, Schwerte), 3.7 g/L Brain-Heart-Infusion, 1 g/L KILPO,. 1 g/L K. I IPO.,. 0.25 g/L MgS0 ₄ -7H ₂ 0 (Merck, Darmstadt), 0.01 g/L CaCh, and 10 g/L glucose (autoclaved separately). The pH was adjusted to 7 with 5 mol/L sodium hydroxide solution. The following components were added after sterile filtration: 2 mg/L biotin, 0.01 g/L MnSO,, H.O (Merck, Darmstadt), 0.01 g/L FeSO H -O (Merck, Darmstadt), 1 mg/L ZnSO H -O. 0.2 mg L CuS0 ₄ -5H ₂ 0 (Merck, Darmstadt), 0.02 mg/L NiCl ₂ -6H ₂ 0 (Merck, Darmstadt), and 0.03 g L 3.4-dihydroxybenzoic acid (Acros Organics, Nidderau)) in a 250 mL shake flask after inoculation with 1 mL of the first pre-culture and cultivated 16 h-18 h with constant shaking at 400 rpm. The fermentation culture was grown in 100 mL CGXII medium (20 g L (NH4) ₂ S0 ₄ , 5 g/L urea, 1 g/L KH ₂ P0 ₄ , 1 g/L K ₂ HP0 ₄ , 0.25 g/L MgS0 ₄ -7H ₂ 0, 0.01 g/L CaCl ₂ , 100 iWL polypropylenglycol (autoclaved separately), and 18 g/L glucose (autoclaved separately). The pH was adjusted to 7 with 5 mol/L sodium hydroxide solution. The following components were added after sterile filtration: 2 mg/L biotin, 0.01 g/L MnS0 -H ₂ 0, 0.01 g/L FeS0 - 7H ₂ 0, 1 mg/L ZnS0 ₄ -7H ₂ 0, 0.2 mg/L CuS0 ₄ -5H ₂ 0, 0.02 mg/L NiQ ₂ -6H ₂ 0, and 0.03 g/L 3.4-dihydroxybenzoic acid). Each bioreactor (DasBox, Eppendorf, Hamburg) was inoculated with the second pre-culture to a final ΟΟβοο of 0.5. The initial stirring speed was set to 100 rpm and air was supplied at 2.2 L/h (0.37 V/(V*min). Dissolved oxygen was continuously monitored (OxyFermFDA 120; Hamilton, Bonaduz (Switzerland)) and maintained at 30 % air saturation by automatic adjustment of the stirring speed. The pH was measured (EasyFermPlus K8 120; Hamilton, Bonaduz (Switzerland)) and maintained at 7 by automatic addition of 1 M NaOH and the temperature was kept at 30 °C. To prevent foam formation the addition of 10 % polypropylene glycol solution was automatically added by a conductivity sensor measuring the conductivity above the fermentation broth. The fermentation was performed as a repeated batch process. Three times an additional feeding with 18 g L glucose (autoclaved separately) was performed. During the fermentation cell dry weight (BDW), glucose concentration, and the anthranilate concentrations were measured offline from 1.5 mL fermenter samples. A 1 mL aliquot of each sample was centrifuged for 5 min at 16000 x g (5415R; Eppendorf, Hamburg) in a weighted reaction tube. To determine the cell dry weight the pellet was dried for at least 72 h at 60 °C (Hybridisation oven; Appligene Oncor Lifescreen, Watford (UK)). Afterwards an analytical balance (La 230 S; Satorius AG, Gottingen) was used for determining the exact pellet weight. The supernatant was analyzed via HPLC-DAD (1100; Agilent Technologies, Santa Clara (USA)) and YSI (YSI-Select 2700; Kreienbaum Neoscience GmBH, Langenfeld). The collected data is shown in Table 4. General manipulation of Corynebacterium glutamicum A TCCI3 32

Plasmid DNA was transferred into Corynebacterium glutamicum ATCC 13032 by electroporation. For transformation cells were harvested from a 200 mL culture grown in BHI medium and in the exponential growth phase (OD ₆₀ o = 1.75-2.0) by centrifugation (4000 g, 10 min, 4 °C (581 OR; Eppendorf, Hamburg)) and washed three times with 20 mL ice-cold TG buffer (1 mM Tris-HCl, 10 % glycerin, pH 7.0). The pellet was resuspended in 2 mL 10 % glycerin solution and directly used for transformation or stored until use at -80 °C. For transformation plasmid DNA (ca. 1 μg) was first pipetted in an ice-cold 0.2 cm gap electroporation cuvette followed by 150 μΕ cell suspension. After incubation on ice for 5 min electroporation was performed (conditions: ~20 ms exponentially decaying pulse, 2.5 kV/cm, 25 F, 200 Ω). The suspension was transferred into a reaction tubes with 500 xL preheated (46 °C) BHI-medium and incubated at 30 °C for 1 h with constant shaking at 400 rpm. The cells were spread on BHI medium agar plates containing the corresponding antibiotics.

For the expression of designated target genes the E. coli-C. glutamicum shuttle vector pEKEx2 was employed (Eikmanns BJ. Kleinertz E, Liebl W, Sahm H, Gene, 1991, 102:93). The 8.16 kb vector derives from vector pU18 and encodes a kanamycin-resistance-cassette, a pBLi point of origin V (oriV) for proliferation in C. glutamicum, a pU18 oriV for proliferation in E. coli, a strong tac- promotor (isopropyl β-D-l -thiogalactopyranoside (IPTG)-inducible), and a lacIQ gene. Alternatively, the E. coli-C. glutamicum shuttle vector pCRB210 was employed (Yukawa H. Inui M, US20130302860(A1), 2013), which is compatible for co-transformation with pEKEx2. The 4.96 kb vector encodes a spectinomycin-resistance-cassette, a p ^' ASE I point of origin V (oriV) for proliferation in C. glutamicum, an oriV for proliferation in E. coli, and a gap A -promotor for constitutive gene expression. Gene- fragments selected for overexpression in C. glutamicum or mutants thereof were synthesized by MWG Operon GmbH under removal of all naturally occurring Sbfl, BamHl, Bglll, Ncol, and Ndel restriction sites by single nucleotide exchanges (silent mutations) and insertion of a ribosomal binding site upstream of the corresponding ORF (apart from the gene sequences of aroG, aroK, glnA, and aroL, which were amplified from their natural loci by PGR. see primers (Table 3)). in case of the trpEG and aroG genes single nucleotide exchanges were included into gene to access feedback-resistant version of the related enzymes. The resulting sequences were cloned into the multiple cloning site of the pEKEx2 vector via Sbfl and BamHI restriction and re- ligation. Apart from glnA gene sequence, this was cloned into the pCRB210 vector via Ncol restriction and re-ligation yielding pCKB-glnA (Table 1). The synthesized gene fragments basically have the structure: "Sbfl-Bglll-RBS -gene-BamHF . This allows the iterative ligation of target genes (as Bglll- 5am_ffl- fragments) into the vector via singular BamHI restriction sites. The plasmids were sequenced to confirm correct cloning and sequence. C. glutamicum (and mutants thereof) transformants were screened for antibiotics resistance and the correct generation of the related mutants was proven by PGR.

For the deletion of target genes, as well as for the insertion of genes or other sequences fragments into the genome of C. glutamicum the E. coli-C. glutamicum shuttle vector pKl9mobsacB was employed (S chafer A, Tauch A, Jager W, Kalinowski J, Thierbach G, Puhler A, Gene, 1994, 145:69). The 5.72 kb vector derives from vector pU18 and encodes a kanamycin-resistance-cassette, no point of origin for proliferation in C. glutamicum, a pU18 oriV for proliferation in E. coli, a lacZa gene, and a sacB gene. The target sequences were synthesized by MWG Operon GmbH (apart from trp D and csm, which were PGR amplified from the C. glutamicum ATCC 13032 genome with specific primers (Table 3) with removal of all naturally occurring Hindlll and EcoRI restriction sites by single nucleotide exchanges (silent mutations) and cloned into the multiple cloning site of the vector via Hindlll and EcoRI restriction and re-ligation. For gene deletions (or integrations) the cloned DNA sequences consist of 300-500 bp of 5 '-flanking region of the target gene (including the first 6 codons of the gene), 21 bp of foreign missense-sequence (or a sequence to be integrated into the genome), and 300-500 bp of 3 '- flanking region of the target gene (including the last 6 codons of the gene of interest). C. glutamicum transformants were screened for single cross-over events integrating the plasmids into the genome by kanamycin-selection, as the resulting plasmids cannot proliferate in C. glutamicum by themself. The BHI medium agar plates (containing 15 mg L kanamycin) carrying trans formants were cultivated at 30 °C for 1 -7 days. Afterwards isolated clones were screened for double-cross-over events, resulting into the in frame deletion of the target genes, leaving behind only the 21 bp of foreign missense sequence (or instead sequences to be integrated into the genome), by selection with sucrose (Colonies were cultivated for 4 h- 6 h in 5 mL Bi l l medium and then plated in two different dilutions (1 : 100 and 1 : 10000) on BHI medium agar plates containing 10 % (w/v) sucrose. The sacB gene on the vector encodes a levansucrase, which transforms sucrose into a lethal product and, as a result, prevents the growth of all clones still carrying the plasmid sequences in their genome. All colonies that grew on the sucrose containing plates were analyzed by co-transfer on BHI medium agar plates supplemented with 15 mg/L kanamycin as well as BH I medium agar plates and cultivated at 30 °C for 1 -7 days. Colonies that grew only on the plates without kanamycin supplementation were analyzed by colony PGR to confirm the correct recombination. The correct cloning of all plasmids and generation of the related mutants was proven by PCR and/or DNA-sequencing. High performance liquid chromatography

An Agilent 1100 series HPLC-DAD system (with diode array detector; Agilent Technologies, Santa Clara (USA)) was used to quantify oAB concentrations in culture supernatants. As stationary phase a CI 8 column Luna® HPLC-column (4.6 x 250 mm; 3 μιη; Phenomenex) was used at 20 °C with a binary solvent system consisting of methanol (solvent B) and water containing 0.1 % formic acid (solvent A) was used. 10 \\L of diluted culture supernatants were injected. The following gradient with a flow rate of 0.5 mL/min was applied: 0-1 min, 2 % B; 1 -2 min, 2-10 % B; 2-12 min, 10-70 % B; 12- 23 min, 70-90 % B; 23-25 min, 90 -98 % B, 25-27 min, 98 % B; 27-27.5 min, 98 %-2 % B; 27.5-30 min, 2 % B. The oAB concentration was determined from the signal integration at 254 nm (retention time: 18.9 min) using an external calibration curve.

Engineering of the trpD gene in C. glutamicum strains

The strain Corynebacterium glutamicum. AtrpD (see Table 2) was created by in- frame deletion of the trpD gene (encoding anthranilate phosphoribosyl transferase; Cgl3032; SEQ ID NO: 1) using the E. coli-C. glutamicum shuttle vector p 19 mobs a cB . The 5 '-flanking region of the trpD open reading frame (including the first 7 codons of the gene) and the 3 '-flanking region of the target gene (including the last 8 codons of trpO open reading frame) were amplified from genomic DNA, isolated from Corynebacterium glutamicum ATCC 13032, by PCR using the primer pairs Del-trpD-1 and Del-trpD- 2, and Del-trpD-3 and Del-trpD-4, respectively (Table 3, SEQ I NO: 66-69). The resulting two fragments were combined by crossover PCR using the primer pair Del-trpD- 1 and Del-trpD-4. The resulting fragment was cloned into the multiple cloning site of the pK19mobsacB vector via Smal restriction and re-ligation. The resulting vector was applied for in- frame deletion of trpD from the genome of Corynebacterium glutamicum ATCC 13032 and introduced in- frame a 24 bp foreign missense-sequence into the genome, which was employed for the crossover PCR (resulting sequences in trpD locus: SEQ ID No.2). Correct mutants were isolated as described above and correct gene deletion proven by PCR. This yielded an L-tryptophan auxotroph strain with oAB accumulation. The strain was characterized towards its properties as an oAB producer as described above with addition of 0.1 mM L-tryptophan or 0.1 mM indole (see Table 4 ).

Six versions of the trpD gene were cloned into vector pEKEx2 via Sbfl and BamHI restriction and re- ligation as described above. The six different versions of the trpO gene were generated with different start codons, and spacer lengths between ribosomal binding site and start codon of trpO (SEQ ID No. 3-8) by PCR using C. glutamicum genomic DNA as a template. The applied primers for the six fragments generation were forward primers (Ex-trpD-1 to -6, Table 2, SEQ ID No. 70-75), which included the changed start codon and spacer between ribosomal binding site and start codon, together with the for each version unchanged reverse primer (Ex-trpD-rev, Table 2. SEQ ID No. 76). This method was chosen in order to achieve a reduction of TrpD protein translation in the produced strains. The strain Corynebacterium glutamicum AtrpD ( Table 2) was transformed with the resulting plasmids (Table 1) and the resulting strains Corynebacterium glutamicum AtrpD/pEKEx2-trpDl -6 (Table 2) grew without the addition of L-tryptophan and were, accordingly, characterized towards their properties as an oAB producer as described above with supplementation of 25 mg L kanamycin and 1 μΜ IPTG (isopropyl- -D-thiogalactopyranosid) to the medium (see Table 4). Also the strains Corynebacterium glutamicum AtrpD/pEKEx2 - trpD5- 6 produced enough L-tryptophan to enable biomass formation; they accumulated significant amounts of oAB (Table 4). The versions of the trpD gene trpDl-3, trpD5, and trpD6 were furthermore each integrated into the genome of the strain Corynebacterium glutamicum AtrpD (Table 2) by homologous recombination using the pKl9mobsacB vector as described above. The 5 '-flanking region of the trpO gene versions and the 3 '-flanking region of the target gene were amplified from genomic DNA, isolated from Corynebacterium glutamicum ATCC 13032, by PCR using the primer pairs Del-trpD-1 and Ko-ti pD- 1 . as well as Del-trpD-3 and Del-trpD-4, respectively (Table 3, SEQ ID No. 66, 77, 68, and 69). The resulting two fragments were combined with the different trpO variants by crossover PCR using the primer pairs Ex-trpD-1 -3, -5 or - 6 and Ko-trpD-2 (Table 3, SEQ ID No. 70-72, 74, 75, and 78). The resulting five fragments were cloned into the multiple cloning site of the pKl9mobsacB vector via Smal restriction and re-ligation. The resulting plasmids (Table 1) were applied for introduction of the six trpD versions into the genome of strain Corynebacterium glutamicum AtrpD. Additionally, 10 foreign nucleotides (GCCCTGCAGG, SEQ I D NO: 92) were integrated into the genomes, as a result of the cloning procedure, upstream of the ribosomal binding site of the /rpD gene, that were employed for the crossover PCR. Correct mutants were isolated as described above and the integrations were verified by sequencing of the trpD region. The resulting strains Corynebacterium glutamicum AtrpD trpDl-3, trpD5, or trpD6 (Table 2) grow without the addition of L-tryptophan or kanamycin and were accordingly characterized towards their properties as an oAB producer as described above without further medium supplements (see Table 4 ).

The described method for gene expression reduction can be applied for other genes instead of performing gene deletions. Engineering of the csm gene in C. glutamicum strains

The strains Corynebacterium glutamicum Acsm, Corynebacterium glutamicum AtrpDAcsm, and Corynebacterium glutamicum AtrpD: : trpD5Acsm (Table 2) were created by in- frame deletion of the csm gene (encoding chorismate mutase; Cgl0853) using the E. coli-C. glutamicum shuttle vector pKl9tnobsacB . The 5 '-flanking region of the csm open reading frame (including the first 6 codons of the gene) and the 3 '-flanking region of the target gene (including the last 6 codons of csm open reading frame) were amplified from genomic ON A, isolated from Corynebacterium glutamicum ATCC 13032, by PCR using the primer pairs Del-csm-1 and Del-csm-2, and Del-csm-3 and Del-csm-4, respectively (Table 3, SEQ I D NO: 79-82). The resulting two fragments were combined by crossover PCR using the primer pair Del-csm-1 and Del-csm-4. The resulting fragment was cloned into the multiple cloning site of the pK\9mobsacB vector via Smal restriction and re-ligation, as described above. The resulting vector was applied for in- frame deletion of csm from the genome of Corynebacterium glutamicum ATCC 13032, Cglutamicum AtrpD: :trpD5, and from the strain Corynebacterium glutamicum AtrpD, respectively, and introduced in- frame a 24 bp foreign missense-sequence into the genome, which was employed for the crossover PCR (resulting sequence in csm locus: SEQ I D No.10). Correct mutants were isolated as described above and correct gene deletion proven by PCR. This yielded the L-tyrosin and L-phenylalanin auxotroph strains C. glutamicum Acsm and C .glutamicum AtrpD: :trpD 5 Acsm, as well as the L-tyrosin, L-phenylalanin, and L-tryptophan auxotroph strain C. glutamicum AtrpDAcsm. Strains were characterized towards their properties as an oAB producer strains as described above with supplementation of the medium with 0.1 mM L-tyrosin and 0.1 mM L-phenylalanin for the C. glutamicum Acsm strain and with supplementation of the medium with 0.1 mM L-tyrosin, 0.1 mM L- phenylalanin, and 0.1 mM L-tryptophan for the C. glutamicum AtrpDAcsm strain (Table 2). The generated strains had an impaired growth rate, compared to the wild type strain as well as strain C. glutamicum AtrpD, and under standard cultivation conditions they did not accumulated significant amounts of oAB (Table 4 ). Six versions of the csm gene were cloned into vector pEKEx2 via Sbfl and BamHI restriction and re- ligation as described above. The six different versions of the csm gene were generated with different start codons, and spacer lengths between ribosomal binding site and start codon of csm (SEQ ID No. 1 1 -16) by PGR using C. glutamicum genomic O A as a template. The applied primers for the six fragments generation were forward primers (Ex-csm-1 to -6), which included the changed start codon and spacer between ribosomal binding site and start codon, together with the for each version unchanged reverse primer (Ex-csm-rev, Table 3, SEQ I D No. 89). This method was chosen in order to achieve a reduction of Csm protein translation in the produced strains. The strain C. glutamicum Acsm (Table 2) was transformed with the resulting plasmids (Table I ) and the resulting strains C. glutamicum A sm/pEKEx2-csml-6 (Table 2) grew without the addition of L-tyrosin, L-phenylalanin, and L-tryptophan and were, accordingly, characterized towards their properties as an oAB producer as described above with supplementation of 25 mg/L kanamycin and 1 μΜ IPTG to the medium (see Table 4). The strain C. glutamicum AtrpD: :trpD5 Acsm (Table 2) was transformed with the empty vector control pEKEx2 (Table I ) and the resulting strain C. glutamicum AtrpD: :trpD5 AcsmlpEKEx2 (Table 2) grew without the addition of L-tryptophan and was, accordingly, characterized towards its properties as an oAB producer as described above with supplementation of 25 mg/L kanamycin and 0.1 mM IPTG to the medium (Table 4).The generated strains produced enough aromatic amino acids to enable biomass formation, but under standard cultivation conditions they did not accumulated significant amounts of oAB (Table 4).

Engineering of the common aromatic pathway and the L-tryptophan branch of C. glutamicum strains

Engineering of the trpEG gene in C. glutamicum strains

The aromatic biosynthesis pathway of C. glutamicum has been elucidated and the genes related to the encoding the enzymes for the biosynthesis of oAB are known (Ikeda M, Appl Microbiol Biotechnol, 2006, 69:615). oAB is an intermediate of the L-tryptophan biosynthesis pathway and is derived from chorismate (CHO) by an anthranilate synthase, consisting of an amidotransferase (TrpE; donor: L- glutamine) and an anthranilate synthase unit (TrpEG), which releases a pyruvate molecule under formation of oAB and L-glutamate. The increased expression of TrpEG-encoding genes has been shown to lead to an accumulation of L-tryptophan in Escherichis coli strains (Ikeda M, Appl Microbiol Biotechnol, 2006, 69:615). Nevertheless, the enzyme has been reported to be strongly feedback- inhibited by L-tryptophan (Ikeda M, Appl Microbiol Biotechnol, 2006, 69:615), resulting in the application of feedback-resistant versions of the TrpEG proteins. Feedback-resistant versions of TrpEG protein have been reported based on trpEG sequences from Brevibacterium lactofermentum (C. glutamicum), E. coli and Salmonella typhimurium (Calguri et al., J Bio Chem, 1991, 8328-8335; Kwak et al, J Biochem Mol Bio, 1999, 20-24, Ikeda M, Appl Microbiol Biotechnol, 2006, 69:615 and Matsui et al., J Bac, 1987, 5330-5332). The most favorable trpEG versions based on the E. coli genes were expressed in C. glutamicum strains, as well as selected amino acid exchanges on the trpEG sequence of the natural TrpEG protein of C. glutamicum ATCC13032 were implemented and expressed in C. glutamicum strains. Taken together four feedback-resistant TrpEG proteins (based on E. coli TrpEG: TrpEGS40R and TrpEGS40F (SEQ ID No. 50-51); and based on C. glutamicum TrpEG: TrpEGS38R and TrpEGS38F) (SEQ ID No. 52-53) were characterized, in comparison to the natural TrpEG protein of C. glutamicum. The five versions of the trpEG genes were synthesized by Eurofins MWG Operon GmbH and cloned into vector pEKEx2 via Sb/I and BamHl restriction and re- ligation as described above and correct cloning was verified by sequencing. The strains C. glutamicum ATCC13032 and C. glutamicum AtrpD::trpD5 (Table 2) were transformed with the resulting plasmids (Table 3) (C. glutamicum AtrpD::trpD5 only with plasmid pEKEx2- trpEG ⁵⁴⁰⁷ ) and the resulting strains C. glutamicum! pEKEx2-trpEG, C. glutamicum/pEKEx-trpEG ^S3SF , C. glutamicum/ pEKEx2 - trpEG ^S3 * ^R , C. glutamicumlpEKEx2-trpEG^ , C. glutamicum /pEKEx2-trpEG ^S40R , C. glutamicum AtrpD::trpD5/pEKEx2-trpEG ^SAOF (Table 2) were characterized towards their properties as an oAB producer, as described above, with supplementation of 25 mg/L kanamycin and 0.1 mM IPTG to the medium (see Table 4 ). Additionally the cell lysates of the five C. glutamicum AtrpD-based strains carrying the pEK ExJ-trpEG variants were analyzed towards the anthranilate synthase activity, according to Caliguri and Bauerle (1991). Cell lysates were produced from cell pellets generated from 30 niL cultures of C. glutamicum AtrpD-based strains carrying the pEKExl-trpFG variants in 300 niL- shake flask with 30 mL BH I medium supplemented with 25 mg/L kanamycin and 1 mM IPTG and grown at 30 °C with shaking at 150 rpm (final O _wll 4). The cell pellets were lysed using 0.1 mm zirconia beads (Zymo Research Coperation, Irvine, CA) and resuspended in assay buffer (50 mM KH2PO4/K2HPO4, 20 mM L-glutamine, 10 mM MgCl?, and 25 μΜ chorimate) according to Caliguri and Bauerle (1991) at 25 °C. The emission was measured at 390 nm using a sp ectrofluorometer (excitation: 390 nm). All tested variants showed a higher specific activity for anthranilate synthase activity compared to the cell lysate with the natural C. glutamicum TrpEG protein. Three biological replicates were used for measurements. Both E. coli TrpEG protein variants had a higher activity and a lower KM compared the C. glutamicum TrpEG protein variants (determination of KM and v _mtK values was performed with Graphpad Prism (La Jolla, CA). The cell lysate of strain C. glutamicum AtrpD/pEKEx2-trpEG ^SA<X! has shown the best performance (see Table 5).

Table 5 Biochemical characteristics of C. glutamicum AtrpD raw extracts producing different variants of TrpEG towards TrpEG activity. Strain Κ _Μ (μηιοΙ) V _MI (mU/mg)

C. glutamicum AtrpD/pEKEx2-trpEG 1 1 .4 12.7

C. glutamicum AtrpD/pEKEx2-trpEG ^S3SF 16.2 68.8

C. glutamicum AtrpD/pEKEx2-trpEG ^S3m 21.6 231.2

C. glutamicum AtrpD/pEKEx2-trpEG ^S40F 1.9 427.5

C. glutamicum AtrpD/pEKEx2-trpEG ^S40R 1.9 317.9

Engineering of the aroG gene in C. glutamicum strains

I n both C. glutamicum and E. coli it has been reported, that carbon flux through the common aromatic pathway up to chorismate is primarily controlled at the first reaction of 3 -deoxy-D-arabino- heptulosonate 7- phosphate synthase (Ikeda M, Appl Microbiol Biotechnol, 2006, 69:615). In C. glutamicum, two types of DSs with different subunit sizes exist. One is an L-tyrosine-sensitive DS with a predicted molecular mass of 39 kDa (type I-DS; the aro product; Cgl0950) and an L-pheny lalanine - and L-tyrosine-sensitive DS with a predicted molecular mass of 51 kDa (type l l-DS; the aroll product; Cgl2098). The type l l-DS forms a polypeptide complex with chorismate mutase, which converts chorismate to prephenate. The type l l-DS exhibits its activity by itself, while CM activity requires the presence of the type ll-DS protein (Ikeda M, Appl Microbiol Biotechnol, 2006, 69:615). A feedback- resistant Arol DS has been described (AroIS187C) (Ikeda M, Appl Microbiol Biotechnol, 2006, 69:615). The feedback-resistant equivalent DS of Aroll, AroG ^ftr (AroGD146N; SEQ ID No. 55) from E. coli, which is not inhibited by the aromatic amino acids, was expressed in C. glutamicum strains to enhance the carbon flux through the common aromatic pathway towards CHO and oAB. The aroG ^D146N gene was amplified by PGR using the primers Ex-aroG-1 and Ex-aroG-2 (Table 3, SEQ ID No. 109- 1 10), which included the restrictions sites and ribosomal binding site, as described above. The resulting DNA fragment was and cloned into vector pEKEx2 via Sbfl and BamHl restriction and re-ligation, as described above, and correct cloning was verified by sequencing. The strain C. glutamicum AtrpD::trpD5 (Table 2) was transformed with the resulting plasmid (Table I ) and the resulting strain C. glutamicum AtrpD::trpD5/pEKEx2-aroG ^D146N (Table 2) was characterized towards its properties as an oAB producer, as described above, with supplementation of 25 mg/L kanamycin and 0.1 mM IPTG to the medium (see Table 4).

Engineering of the trpEG and aroG gene in C. glutamicum strains To allow investigation of the influence of combined expression of AroGD146N and TrpEGS40F in C. glutamicum strains the DNA sequences encoding these proteins were fused onto the pEKEx2 vector. The aroG ^D146N sequence was fused to pEKEx2-irp£'G ^S40F to gain pEKExl-trpEG^ -aroG ^D146N (Table 1) and the gene trpEG ^&40F was fused to the plasmid pEKEx2 -a roG ^D146N to gain pEKEx2-aroG ^D146N - trpEG^ ⁴⁰ ^ (Table 1), both via Bglll and BamHI restriction and re-ligation, as described above, and correct cloning was verified by sequencing. The strain C. glutamicum AtrpD::trpD5 (Table 2) was transformed with each of the resulting plasmids and the resulting strains C. glutamicum. AtrpD::trpD5/pEKEx2-aroG ^D146N -trpEG ^S40F and C. glutamicum AtrpD: : trpD5/pEKEx2 -trpEG ^S4ae - aroG ^D146N (Table 2) were characterized towards their properties as oAB producers, as described above, with supplementation of 25 mg/L kanamycin and 0.1 mM IPTG to the medium (see Table 4). Additionally, the strain C. glutamicum AtrpD: trpD5Acsm (Table 2) was transformed with plasmid pEKEx2 -aro G ^D146N -irpEG ^SAm ' and the resulting strain C. glutamicum AtrpD::trpD5Acsm IpEKExl- aroG ^D146N -trpEG ^s,40F (Table 2) was characterized towards its properties as an oAB producer, as described above, with supplementation of 25 mg L kanamycin, 0.1 mM L-tyrosin, 0.1 mM L- phenylalanin, and 0.1 mM IPTG to the medium (see Table 4).

Engineering of the aroK and aroL gene in C. glutamicum strains

To prevent an accumulation of intermediates in the common aromatic amino acid biosynthesis pathway of C. glutamicum strains the gene aroK (encodes shikimate kinase; Cgll622; SEQ ID No. 94) from C. glutamicum and the gene aroL (encodes shikimate kinase; CP000948; SEQ ID No. 93) from E. coli were investigated. The aroK and the aroL genes were separately amplified by PGR using the primers Ex-aroK-1 and Ex-aroK-2 (Table 3, SEQ ID No. 101 -102) and Ex-aroL-1 and Ex-aroL-2 (Table 3, SEQ ID No. 99-100), respectively. The primers included the restrictions sites and ribosomal binding site, as described above. The resulting DNA fragments were separately integrated into vector pEKEx2 via Sbfl and BamHI restriction and re-ligation, as described above, to yield the plasmids pEKEx2-aroK and pEKEx2-</m/, and correct cloning was verified by sequencing. The strain C. glutamicum AtrpD::trpD5 (Table 2) was transformed with the resulting plasmids (Table 3) and the resulting strains C. glutamicum AtrpD::trpD5/pEKEx2-aroL and C. glutamicum AtrpD : : trpD5l pEKEx2 - roK (Table 2) were characterized towards their properties as oAB producers, as described above, with supplementation of 25 mg/L kanamycin and 0.1 mM IPTG to the medium (see Table 4). In strain C. glutamicum AtrpD::trpD5/pEKEx2-aroL no significant accumulation of shikimate and 3- dehydroshikimat was detectable by HPLC-DAD analyses of culture supernatants.

Engineering of the glnA gene in C. glutamicum strains To analyze the impact of the L-glutamine synthetase (GlnA; Cgl2214, SEQ ID No. 95) activity on the production of oAB by C. glutamicum strains the L-glutamine synthetase gene, gin A, was expressed in strain C. glutamicum strains. The glnA gene was amplified by PGR using the primers Ex -gin A- 1 and Ex-glnA-2 (Table 3, SEQ ID No. 97-98), which included the restrictions sites and ribosomal binding site, as described above. The resulting DNA fragment was and cloned into vector pCRB210 via Ncol restriction and re-ligation, as described above, and correct cloning was verified by sequencing. The strain C. glutamicum AtrpD::trpD5 was transformed with the resulting plasmid (Table 3) and the resulting strain C. glutamicum AtrpD::trpD5/pCRB-glnA (Table 2) was characterized towards its properties as an oAB producer, as described above, with supplementation of 25 mg/L kanamycin and 0.1 mM IPTG to the medium (see Table 4). As a control the strain C. glutamicum AtrpD :trpD was transformed with the empty vector control pCRB210 (Table 1) and the resulting strain C. glutamicum AtrpD trpD5/pCRB210 (Table 2) was characterized towards its properties as an oAB producer, as described above, with supplementation of 25 mg/L kanamycin and 0.1 mM IPTG to the medium (see Table 4). Although the growth of the strain was reduced a beneficial effect on the anthranilate accumulation of C. glutamicum AtrpD::trpD5 could not be observed. It cannot be ruled out that the increased activity of GlnA will increase the accumulation of anthranilate in a more advanced oAB producer strain as it has been observed in E. coli (Sun et al, Appl Environ Microbiol, 2013, 4024- 4030). Engineering of the central metabolism of C. glutamicum strains

Engineering of the pyk gene in C. glutamicum strains

A pyruvate kinase (Pyk) consumes phosphenolpyruvate by production of pyruvate and ATP. The effect of an in- frame deletion of the pyk gene (SEQ ID NO: 23) in C. glutamicum strains was investigated. The strain C. glutamicum AtrpD : : trpD5 Apyk (Table 2) was created by in-frame deletion of the pyk gene (encoding pyruvate kinase; Cgl2089) using the E. coli-C. glutamicum shuttle vector pKl9mobsacB , as described above. The resulting plasmid pSB082 (Table 1) was applied for in-frame deletion of pyk gene from the genome of C. glutamicum AtrpD::trpD5, and introduced in- frame a 24 bp foreign missense-sequence into the genome (resulting sequence in pyk locus: SEQ I D No. 24). Correct mutants were isolated as described above and correct gene deletion proven by PGR. This yielded strain C. glutamicum AtrpD: : trpD5 Apyk, which was characterized towards its properties as an oAB producer strain, as described above (Table 4).

Engineering of the gpi gene in C. glutamicum strains

In order to direct the carbon flux towards the pentose phosphate pathway the effect of an in-frame deletion of the glucose-6-phosphate isomerase (Gpi) encoding gene (SEQ ID NO: 27) in C. glutamicum strains was investigated. The strain C. glutamicum AtrpD: : trpD5 Agpi (Table 2) was created by in-frame deletion of the gpi gene (encoding glucose-6-phosphate isomerase; Cgl0851) using the E. coli-C. glutamicum shuttle vector pKl9mobsacB, as described above. The resulting plasmid pSB064 (Table 1) was applied for in- frame deletion of gpi gene from the genome of C. glutamicum AtrpD::trpD5, and introduced in- frame a 24 bp foreign missense-sequence into the genome (resulting sequence in gpi locus: SEQ ID No .28). Correct mutants were isolated as described above and correct gene deletion proven by PCR. This yielded strain C. glutamicum AtrpD::trpD5Agpi, which was characterized towards its properties as an oAB producer strain, as described above (Table 4). Engineering of the pepco gene in C. glutamicum strains

The phosphoenolpyruvate carboxylase (Pepco) consumes PEP by transforming it into oxaloacetate. In order to access an increased PEP pool an in-frame deletion of the phosphoenolpyruvate carboxylase encoding gene (SEQ I D NO: 21) in C. glutamicum strains was investigated. The strain C. glutamicum AtrpD::trpD5 Apepco (Table 2) was created by in- frame deletion of the pepco gene (encoding phosphoenolpyruvate carboxylase; Cgll585) using the E. coli-C. glutamicum shuttle vector pKl9mobsacB , as described above. The resulting plasmid pSB061 (Table 1) was applied for in-frame deletion of pepco gene from the genome of C. glutamicum AtrpD::trpD5, and introduced in- frame a 24 bp foreign missense-sequence into the genome (resulting sequence in pepco locus: SEQ ID No.22). Correct mutants were isolated as described above (with additional supplementation of 1 % acetate to BHI medium agar plates) and correct gene deletion proven by PCR. This yielded strain C. glutamicum AtrpD::trpD5 Apepco , which was characterized towards its properties as an oAB producer strain, as described above (Table 4).

Engineering of the ptsG and the hpr gene in C. glutamicum strains

Glucose and fructose uptake of C. glutamicum is mainly utilized by the phosphotransferase system (PTS). This multi- enzyme complex performs a translocation of glucose via a pho sphory 1 ation cascade mediated by phosphoenolpyruvate with concomitant glucose phosphorylation to glucose-6-phosphate across the cytoplasmic membrane (Siebold et al. 2001). The system consists of two soluble components enzyme I and histidine rich protein (Hpr) and a membrane bound/ associated enzyme complex (Tanaka et al. 2008). The strains C. glutamicum AtrpD::trpD5Ahpr and C. glutamicum AtrpD::trpD5 AptsG (Table 2) were created by in- frame deletion of the hpr gene (encoding histidine rich protein; Cgll937, SEQ ID NO: 17) and the ptsG gene (encoding phosphotransferase system unit G; Cgll360, SEQ ID NO: 19), respectively, using the E. coli-C. glutamicum shuttle vector pKl9mobsacB, as described above. The resulting plasmids pSB060 and pSB079 (Table 1) were applied for in-frame deletion of hpr gene and ptsG gene, respectively, from the genome of C. glutamicum AtrpD::trpD5 , and introduced in-frame a 24 bp foreign missense-sequence into the genome (resulting sequence in hpr/ptsG locus: SEQ ID No.18; ptsG locus: 20). Correct mutants were isolated as described above (with additional supplementation of 5 g/L ribose to BHI medium agar plates) and correct gene deletion proven by PCR. This yielded strain C. glutamicum AtrpD: : trpD5 Ahpr and C. glutamicum AtrpD::trpD5AptsG, which were characterized towards its properties as an oAB producer strain, as described above (Table 4). Both strains did not show significant growth under standard fermentation conditions, as described above, and as a result did not accumulate oAB.

Engineering of the ppk gene in C. glutamicum strains

The phosphoenolpyruvate carboxykinase (Ppk) can, as a part of the glyoxylate pathway, recycle oxaloacetate into PEP. In order to access an increased PEP pool a plasmid-based expression of the phosphoenolpyruvate carboxykinase encoding gene (Cgl2862; SEQ ID NO: 36) in C. glutamicum strains was investigated. The ppk gene was synthesized by Eurofms MWG Operon GmbH and cloned into vector pEKEx2 via Sbfl and BamHI restriction and re- ligation, as described above, resulting in the plasmid pSi 51)72 (Table 1), and correct cloning was verified by sequencing. The strain C. glutamicum AtrpD :trpD5/pSB072 (Table 2) was created by transformation of strain C. glutamicum AtrpD::trpD5 with plasmid pSB072, as described above. Correct mutants were isolated as described above and successful transformation with the target gene proven by colony PCR. This yielded strain C. glutamicum AtrpD:: trpD51 p SB 072, which was characterized towards its properties as an oAB producer strain, as described above (Table 4 ).

Engineering of the pps gene in C. glutamicum strains

The phosphoenolpyruvate synthase (Pps) catalysis, as part of the gluconeogenesis, the formation of PEP by pyruvate recycling and consumption of ATP to AMP. In order to access an increased PEP pool a plasmid-based expression of the phosphoenolpyruvate synthase encoding gene together with the PEP synthase binding domain encoding gene (Cgi0551 and Cgl0552; SEQ I NO: 33-35) in C. glutamicum strains was investigated. The pps genes were together synthesized by Eurofins MWG Operon GmbH and cloned into vector pEKEx2 via Sbfl and BamHI restriction and re-ligation, as described above, resulting in the plasmid pSB073, and correct cloning was verified by sequencing. The strain C. glutamicum AtrpD::trpD5lpSB073 (Table 2) was created by transformation of strain C. glutamicum AtrpD::trpD5 with plasmid pSB073, as described above. Correct mutants were isolated as described above and successful transformation with the target gene proven by colony PCR. This yielded strain C. glutamicum AtrpD:: trpD51 p SB 073, which was characterized towards its properties as an oAB producer strain, as described above (Table 4). Engineering of the zwfl together with the opcA gene in C. glutamicum strains

I n Zymomonas mohilis and E. coli the enhanced production of the enzyme catalyzing the first reaction of the pentose phosphate pathway (PPP) (formation of ribulose-5 -phosphate from glucoce-6-phosphate by glucose-6-phosphat dehydrogenase (Zwfl and OpcA)) was reported to result in an increased carbon flux into the PPP and as a result an increase of the erythrose-4-phosphate (E4P) pool. In order to access an increased E4P pool a plasmid-based expression of the glucose-6-phosphat dehydrogenase encoding genes {zwfl (Cgll 576) and opcA (Cgll577; S EQ ID NO: 37-39) in C. glutamicum strains was investigated. The genes were together synthesized by Euro fins MWG Operon GmbH and cloned into vector pEKEx2 via Sb/I and BamHI restriction and re-ligation, as described above, resulting in the plasmid pSB078 (Table 1), and correct cloning was verified by sequencing. The strain C. glutamicum AtrpD::trpD5lpSB078 (Table 2) was created by transformation of strain C. glutamicum AtrpD::trpD5 with plasmid pSB078, as described above. Correct mutants were isolated as described above and successful transformation with the target gene proven by colony PGR. This yielded strain C. glutamicum AtrpD :trpD5/pSB078, which was characterized towards its properties as an oAB producer strain, as described abovc(Table 4).

Engineering of tal and tkt genes in C. glutamicum strains

An enhanced carbon flux through the PPP and an increased E4P pool has been described by the enhanced expression of transketolase and transaldolase in E. coli. Transketolase (Tkt) and transaldose (Tal) encoding genes oi E. coli (tktEC (ECDH10B_3110) and talEC ( ECDH ! OB 2629 ); SEQ ID NO: 41 and 43) as well as the equivalent genes of C. glutamicum (tktCG (Cgll 574) and talCG (Cgll575); SEQ I NO: 40 and 42) were plasmid-based expressed in C. glutamicum strains. The four sequences were synthesized by Eurofins MWG Operon GmbH and separately integrated into vector pE Ex2 via Sbfl and BamHI restriction and re-ligation, as described above, resulting in the plasmids pSB074- pSB077 (Table 1), and correct cloning was verified by sequencing. The strains C. glutamicum AtrpD trpD5/pSB074 - pSB077 (Table 2) were created by transformation of strain C. glutamicum AtrpD::trpD5 with each of the plasmids, as described above. Correct mutants were isolated as described above and successful transformation with the target gene proven by colony PGR. This yielded strains C. glutamicum AtrpD:: trp∑>51 pSB 074 - p SB 077, which were characterized towards their properties as an oAB producer strains, as described above (Table 4).

To allow investigation of the effect of combined expression of transketolase and transaldolase in C. glutamicum strains the ON A. sequences encoding these proteins were fused onto the pEKEx2 vector. The talCG sequence was fused to pSB075 to gain pSB085 (Ta le 1) and the gene talEC was fused to the plasmid pSB077 to gain pSB086 (Table 1), both via Bglll and BamHI restriction and re-ligation, as described above, and correct cloning was verified by sequencing. The strain C. glutamicum AtrpD::trpD5 was transformed with each of the resulting plasmids and the resulting strains C. glutamicum AtrpD: : trpD5lpSB085 and C. glutamicum AtrpD trpD5/pSB086 (Table 2) were characterized towards their properties as oAB producers, as described above, with supplementation of 25 mg L kanamycin and 0.1 mM BPTG to the medium (Table 4).

Engineering of the galP, the iolT2 and the ppgk gene in C. glutamicum strains

I n order to restore efficient glucose uptake in Pts-deficient C. glutamicum strains a plasmid -based expression of galactopermease encoding gene (galP, Cgl2409; SEQ I D NO: 30) in combination with a polyphosphoglucokinase (a glucose phosphorylating enzyme) encoding gene (ppgk (Cgl910; SEQ ID NO: 32) in C. glutamicum strains was investigated. Furthermore, the inositolpermease T2 unit (of the inositolpermease) encoding gene (iolT2, Cgl3058; SEQ ID NO: 31) in combination with a polyphosphoglucokinase encoding gene (ppgk (Cgl910; SEQ I NO: 32) in C. glutamicum strains was investigated. The four genes were synthesized by Eurofins MWG Operon GmbH and each integrated into vector pEKEx2 via Sbfl and BamHI restriction and re- ligation, as described above, resulting in the plasmid pSB068, pSB070, and pSB071 (Table 1) and correct integration was verified by sequencing. The ppgk sequence was fused to galP sequence and iolT2, respectively, onto the pEKEx2 vector. The galP sequence was fused to pSB071 to gain pSB083 (Table 1) and the gene iolT2 was fused to the plasmid pSB071 to gain pSB084 (Table 1), both via Bglll and BamHI restriction and re- ligation, as described above, and correct cloning was verified by sequencing. The strains C. glutamicum AtrpDv.trpD5h.ptsG and C. glutamicum AtrpD::trpD5Ahpr (Table 2) were each transformed with the plasmids pSB083 and pSB084, respectively, and the resulting strains C. glutamicum AtrpD::trpD5AptsG/pSB083 , C. glutamicum AtrpD trpD5AptsGlpSB084, C. glutamicum AtrpD::trpD5AhprlpSB083, and C. glutamicum AtrpD: :trpD5AhprlpSB084 (Table 2) were characterized towards their properties as oAB producers, as described above, with supplementation of 25 mg/L kanamycin and 0.1 mM IPTG to the medium (Table 4). The strain AtrpD::trpD5Ahpr/pSB083 did not show significant growth under standard fermentation conditions, as described above, and as a result did not accumulate oAB. Engineering of the o. ili export from C. glutamicum strains

With increasing production rates of oAB the export of the product from producing cells can easily become a production limiting step. Furthermore, intracellular accumulation of oAB very likely can have a significant toxic effect on cells. No oAB exporter has been described so far. Recently, a shikimate permease (QsuA), belonging to the family of major facilator systems, was described by Kubota et ah, which facilitates the import of shikimate and quinate (Kubota et ah, Molecular Microbiol, 2014, 92(2), 356) in C. glutamicum. In order to test the effect of the shikimate permease on intra- and extracellular concentrations of oAB in C. glutamicum strains, the related gene (qsuA, Cgl0492; SEQ ID NO: 96) was synthesized by Eurofms MWG Operon GmbH and integrated into vector pEKEx2 via Sbfl and BamHI restriction and re-ligation, as described above, resulting in the plasmid pSB096 (Table 1), and correct cloning was verified by sequencing. The strain C. glutamicum AtrpD :trpD5/pSB096 (Table 2) was created by transformation of strain C. glutamicum AtrpD::trpD5 with plasmid pSB096, as described above. Correct mutants were isolated as described above and successful transformation with the target gene proven by colony PGR. This yielded strain C. glutamicum AtrpD trpD5/pSB096, which was characterized towards its properties as an oAB producer strain, as described above (Table 4 ).

Example 4 - Production of o-amiiiobeiizoate with P. putida

Strains derived from Pseudomonas putida KT2440 were developed, that produce oAB. In order to achieve an oAB accumulation by P. putida strains, the trpOC genes (PP 0421 and PP 0422; SEQ ID NO: 63), encoding anthranilate phosphoribosyltransferase and indole-3 -glycerol phosphate synthase, were chosen to be disrupted in Pseudomonas putida KT2440. The construction of the deletion mutant P. putida AtrpDC, based on a knockout plasmid pEMG-Del-trpDC (Table 1), was approached by the method describe in Martinez-Garcia et al (Martinez -Garcia et al, Environ Microbiol, 201 1, 13(10), 2702) using the vector pEMG. This approach results in the removal of the gene sequences (gen of interest, GO I) without residual cloning scars or markers. The concept of the method is shown in the Figure 27. In Step A - B the designed knockout vector pEMG, bearing the disruption flanks TS 1 and TS2, a kanamycin marker and the l-Scel sites, is integrated into the genome of P. putida. In step C the double strand break is induced by the transformation of a second plasmid (pSW-2), bearing the meganuclease l-Scel genes. The break is repaired in vivo, by homologous recombination of either the TS 1 or the TS2 region, resulting in either P. putida wild type or P. putida AtrpDC (step D). Step E shows the PGR screening which is required to distinguish and isolate the different strains. Deletion flanks were chosen such that the open reading frame of the genes is removed while keeping neighbouring genes intact (Figure 28). This strategy can lead to greater strain stability in long-term continuous cultivation. The 800 bp disruption flanks TS 1 and TS2 were amplified using a Phusion® High-Fidelity DNA Polymerase (New England Biolabs). The Phusion PGR was performed according to the manufacturer's manual (25 cycles, 61,8°C (TS1), 70,2°C (TS2) 1 :30 minutes). The PGR fragment TS 1 was digested with BamHI and Xhol and the PCR fragment TS2 was digested with Xhol and Sbfl. The plasmid pEMG and the fused TS I -TS2 PCR product were digested with BamHI and Sb/I. All restriction mixtures were purified using a High Pure PCR Product Purification Kit (Roche). The ligation of pEMG-Oel-trpDC , using the individual flanks TS l and TS2, was performed with a T4 DNA ligase (Thermo Fisher Scientific) according to the manufacturer's instructions. The ligation mixture was transformed to the chemically competent E. coli DH5a λρϊτ strain. Positive plasmids from the three point ligation mixture consisting of the two individual flanks and the digested vector pEMG were identified by restriction analysis and verified by sequencing, which confirmed the successful construction of pEMG-Del-irpDC (Table 1). The knockout vector pEMG-Del-irpDC was integrated into the genome of P. putida (Figure 27. step A-B) by tri-parental mating using the acceptor strain P. putida, the donor strain E. coli DH5a )jpir/pEMG-Del-trpDC and a helper strain E. coli HB101 pRK2013 (Martinez-Garcia et al, Environ Microbiol, 2011, 13(10), 2702). P. putida /pEMG-Del- trpDC was isolated from the cell mixture by using cetrimide agar plates with 50 mg/ ^' L kanamycin and single colonies were re-streaked on LB -kanamycin plates. Genome integration of the knockout vector was confirmed in single colonies via colony PCR. To introduce the double strand breakage in the genome (Figure 27. step C), a second plasmid (pSW-2) bearing the l-Scel meganuclease gene, was transformed into the newly constructed strain. Therefore electro-competent P. putida /pEMG-Del- trpDC cells were obtained according to Choi et al (Choi et al, J Microbiol Methods, 2005, 64(3), 391). The electroporation was performed using a Biorad Gene Pulser Xcell Electroporator (2.5 kV, 200 ohm, 25 μΤ) and the cell suspension was plated out on LB-gentamycin (30 mg/'L) and LB-kanamycin- gentamycin plates (30 mg L and 50 mg/ ^' L, respectively). Following the protocol of the knockout procedure described in Martinez-Garcia et al an induction of the l-Scel meganuclease is necessary. However, this step was omitted due to practical experience implying a leaky expression of the l-Scel meganuclease. To distinguish the desired P. putida AtrpDC knockout strain from the P. putida and the P. putida/pEMG-Oel-trpDC strain, single colonies were streaked out on LB and LB- kanamycin plates. Kanamycin sensitive colonies were checked via colony PCR as described above, in addition kanamycin sensitive colonies were checked for L-phenylalanine auxotrophy on minimal medium plates with 20 mM glucose and with and without 0.1 mM L -tryptophane. Gene deletions were verified by PCR analysis and sequencing. The resulting strain P. putida AtrpDC (Table 2) was not able to grow on minimal medium without L -tryptophane supplementation.

In addition, the pheA gene ( PP I 769. SEQ ID NO: 64), encoding chorismate mutase, was disrupted in the Pseudomonas putida AtrpDC strain. The pheA gene encodes the first step of L-phenylalanine and L- tyrosine biosynthesis, which likely competes with anthranilate production. The construction of the deletion mutant P. putida AtrpDC A pheA, based on a knockout plasmid pEMG-Dcl-/;//c.-i. was approached by the method describe in Martinez -Garcia et al (Martinez-Garcia et al, Environ Microbiol, 2011, 13(10), 2702) using the vector pEMG, as described above. Deletion flanks were chosen such that the open reading frame of the gene is removed while keeping neighbouring genes intact (Figure 29). This strategy can lead to greater strain stability in long-term continuous cultivation. In case of pheA this meant that the eight 5' nucleotides of the gene remained due to an overlap with the serC gene. To construct the 800 bp disruption flanks TS I and TS2 the following primers were used JK038f, JK039r, JK040f, and JK041r (Table 3, SEQ ID NO: 103-106). The PGR fragment TS 1 was digested with BamHI and Xhol and the PGR fragment TS2 was digested with Xhol and Sbfl. The plasmid pEMG and the fused TS I -TS2 PGR product were digested with BamHI and Sbfl. All restriction mixtures were purified using a High Pure PGR Product Purification Kit (Roche). The ligation of pEMG-Del-p/ze^, using the individual flanks TS1 and TS2, was performed with a T4 DNA ligase (Thermo Fisher Scientific) according to the manufacturer's instructions. The ligation mixture was transformed to the chemically competent E. coli DH5a λρϊτ strain. Positive plasmids from the three point ligation mixture consisting of the two individual flanks and the digested vector pEMG were identified by restriction analysis and verified by sequencing, which confirmed the successful construction of pEMG-Del-p/ze i (Table 1). The knockout vector p¥MG-Dc\-pheA was integrated into the genome of P. putida AtrpDC (Figure 27. step A-B) by tri-parental mating using the acceptor strain P. putida AtrpDC, the donor strain E. coli DH5a pir/pEMG-Del-pAe^ and a helper strain E. coli HB101 pRK2013 (Martinez- Garcia et al, Environ Microbiol, 2011, 13(10), 2702). P. putida AtrpDC/pEMG-Oel-pheA was isolated from the cell mixture by using cetrimide agar plates with 50 mg/L kanamycin and single colonies were re- streaked on LB -kanamycin plates. Genome integration of the knockout vector was confirmed in single colonies via colony PGR. To introduce the double strand breakage in the genome (Figure 27. step C), a second plasmid (pSW-2) bearing the l-Scel meganuclease gene, was transformed into the newly constructed strain. Therefore electro-competent P. putida AtrpDC ρΕ\Κ ΐ-Γ\·1-/>Λ<·. / cells were obtained according to Choi et al (Choi et al, J Microbiol Methods, 2005, 64(3), 391). The electroporation was performed using a Biorad Gene Pulser Xcell Electroporator (2.5 kV, 200 ohm, 25 μ¥) and the cell suspension was plated out on LB-gentamycin (30 mg L) and LB-kanamycin- gentamycin plates (30 mg/L and 50 mg/L, respectively). Following the protocol of the knockout procedure described in Martinez-Garcia et al an induction of the l-Scel meganuclease is necessary. However this step was omitted due to practical experience implying a leaky expression of the l-Scel meganuclease. To distinguish the desired P. putida AtrpDCApheA (Table 2) knockout strain from the P. putida AtrpDC and the P. putida AtrpDC/pEMG-Oel-pheA strain, single colonies were streaked out on LB and LB- kanamycin plates. Kanamycin sensitive colonies were checked via colony PC R as described above. In addition kanamycin sensitive colonies were checked for L -phenylalanine auxotrophy on minimal medium plates with 20 mM glucose, 0.1 mM L-tryptophane, and with and without 1 niM L -phenylalanine . Gene deletions were verified by PGR analysis and sequencing. Since P. putida KT2440 possesses a L -phenylalanine -4 -hydroxylase (PheA) which converts L -phenylalanine to L-tyrosine, L-tyrosine-auxotrophies were rescued by the addition of L-phenylalanine. Since both L- tryptophane, and L-phenylalanine auxotrophs were obtained by single gene disruptions, it was assumed that no redundant genes encoding alternative enzymes were present in P. putida KT2440.

The first dedicated reaction of the aromatic amino acid biosynthesis pathway is catalysed by DAM synthase isoenzymes. It has been shown in several organisms that over expression of a feedback- insensitive mutant of this enzyme leads to an increased flux through the aromatic s biosynthesis pathway (Ikeda M, Appl Microbiol Biotechnol, 2006, 69:615). Therefore, in order to increase anthranilate production, a feedback -ins ensitive 3 -deoxy-D-arabino-heptulosonate 7 -phosphate ( DAMP) synthase, encoded by the aroG ^m46N gene (SEQ ID NO: 55), was expressed in the strains P. putida KT2440, P. putida AtrpDC and P. putida AtrpDCApheA via vector pSEVA234 (under control of the lacP-Vtrc system; encoding a kanamycin resistance and a pBBRl origin of replication) (Silva-Rocha et al, Nucleic Acids Research, 2013, D666-75). The aroG ^Di46N gene was restricted from a donor plasmid (pCAS-2JF- aroG ^{m m} ) using EcoRI and BamHl restriction enzymes and standard conditions, as described above (Example 3), resulting in a 1 1 17 bp fragment. The vector pSEVA234 was restricted with the same enzymes. The fragments were purified using a High Pure PGR Product Purification Kit (Roche). The ligation of the fragment to gain pSEVA234-aroG ^z)/' ' ^&Y (Table 1) was performed with a T4 DNA ligase (Thermo Fisher Scientific) according to the manufacturer's instructions. The ligation mixture was introduced into electro- competent Escherichia coli DH5a cells (ElectroMAX™ DH5a- E™ Competent Cells; Life Technologies, Darmstadt). After electro-transformation (conditions: -20 ms exponentially decaying pulse, 2.5 kV/cm, 25 F, 200 Ω) in 0.2 cm gap electroporation cuvettes (BioRad, Hercules, CA) using a Gene Pulser Xcell System (BioRad, Hercules, CA), 800 μΕ LB recovery medium (Luria-Bertani; Roth, Karlsruhe) was immediately added and the suspension was transferred into 1.5 mL microcentrifuge tube. After 1 h at 37 °C, a cell suspension were spread onto LB agar plates (Abcr GmbH & Co. Kg, Karlsruhe), supplemented with 50 mg/L kanamycin, and incubated at 37 °C over night. Clones were collected and correct transformation confirmed via restriction analysis and sequencing. For plasmid preparation the stains were cultivated in 15 mL-tubes for 14 h-16 h with constant shaking at 200 rpm in 3 mL LB medium and 50 mg/L kanamycin at 37 °C (Kuhner Shaker ISF-4-W; Adolf Kuhner AG, Basel (Switzerland)). Plasmid DNA purification was done using the NucleoSpin ^® Plasmid Pure kit (Macherey & Nagel, Diiren) following the manufacturer's instructions. Electro-competent P. putida KT2440, P. putida AtrpDC, and P. putida AtrpDCApheA cells were obtained according to Choi et al (Choi et al, .1 Microbiol Methods, 2005, 64(3):391). The electroporation was performed using a Biorad Gene Pulser Xcell Electroporator (2.5 kV, 200 ohm, 25 μΡ) and the cell suspension was plated out on LB-agar plates supplemented with 50 m/L kanamycin. This resulted in the generation of the strains P. putida KT2440/ pSEVA234-aroG ^D46N , P. putida AtrpDC/ pSE VA234-aroG ^D46N , and P. putida AtrpDCApheA/pSEVA234-aroG ^D4 (Table 2). Additionally, the empty vector pSEVA234 was transformed into to the strains P. putida AtrpDC and P. putida AtrpDCApheA to gain the strains P. putida AtrpDCIpSEVA234 , and P. putida AtrpDCApheA/pSEVA234 (Table 2).

oAB is an intermediate of the L-tryptophan biosynthesis pathway and is derived from chorismate by an anthranilate synthase, consisting of an amidotransferase (TrpE; donor: L-glutamine) and an anthranilate synthase unit (TrpEG), which releases a pyruvate molecule under formation of oAB and L- glutamate. The increased expression of Trp EG - encoding genes has been shown to lead to an accumulation of L-tryptophan in Escherichia coli strains (Ikeda M, Appl Microbiol Biotechnol, 2006, 69:615). Nevertheless, the enzyme has been reported to be strongly feedback-inhibited by L- tryptophan, resulting in the application of feedback-resistant versions of the TrpEG proteins. Feedback- resistant versions of TrpEG protein have been reported based on trpEG sequences from Salmonella typhimurium (Ikeda M, Appl Microbiol Biotechnol, 2006, 69:615). The most favorable trpEG version, based on the S. typhimurium genes, irpEG ⁸⁴ * (SEQ ID No. 53) was expressed in P. putida strains. The feedback-inhibition-resistant version of the trpEG gene was cloned into P. putida vector pSEVA234 to enable plasmid-based expression in the strains P. putida T2440, P. putida AtrpDC, and P. putida AtrpDCApheA (under control of the lacF-Ptrc system; encoding a kanamycin resistance and a pBBRl origin of replication) (Silva-Rocha et al, Nucleic Acids Research, 2013, D666-75). The trpEG^ ⁰⁷ gene was restricted from a donor plasmid using Bglll and BamHI restriction enzymes and standard conditions, as described above (Example 3), resulting in a 2200 bp fragment. The vector pSEVA234 was restricted with BamHI, resulting in a 4550 bp fragment. The fragments were purified using a High Pure PGR Product Purification Kit (Roche). The ligation of the fragments to gain pSEVA234- trpEG^ ^1' (Table 1) was performed with a T4 DNA ligase (Thermo Fisher Scientific) according to the manufacturer's instructions. The ligation mixture was introduced into electro- competent Escherichia coli DH5a cells (ElectroMAX™ DH5a-E™ Competent Cells; Life Technologies, Darmstadt). After electro-transformation (conditions: -20 ms exponentially decaying pulse, 2.5 kV/cm, 25 F, 200 Ω) in 0.2 cm gap electroporation cuvettes (BioRad, Hercules, CA) using a Gene Pulser Xcell System (BioRad, Hercules, CA), 800 \xL LB recovery medium (Luria-Bertani; Roth, Karlsruhe) was immediately added and the suspension was transferred into 1.5 mL microcentrifuge tube. After 1 h at 37 °C, a cell suspension were spread onto LB agar plates (Abcr GmbH & Co. Kg, Karlsruhe), supplemented with 50 mg/L kanamycin, and incubated at 37 °C over night. Clones were collected and correct transformation confirmed via restriction analysis and sequencing. For plasmid preparation the stains were cultivated in 15 mL-tubes for 14 h-16 h with constant shaking at 200 rpm in 3 mL LB medium and 50 mg L kanamycin at 37 °C (Kuhner Shaker ISF-4-W; Adolf Kuhner AG, Basel (Switzerland)). Plasmid ON A purification was done using the NucleoSpin ^® Plasmid Pure kit (Macherey & Nagel, Diiren) following the manufacturer's instructions. Electro- competent P. putida KT2440, P. putida AtrpDC, and P. putida AtrpDCApheA cells were obtained according to Choi et al (Choi et al, J Microbiol Methods, 2005, 64(3), 391). The electroporation was performed using a Biorad Gene Pulser Xcell Electroporator (2.5 kV, 200 ohm, 25 μ¥) and the cell suspension was plated out on LB-agar plates supplemented with 50 m/L kanamycin. This resulted in the generation of the strains P. putida KT2440/ pSEVA234-trpEG ^S40F , P. putida AtrpDC/pSEVA234-trpEG ^S40F , and P. putida A trpDCAph eA IpSE VA234- trpEG^^ (Table 2).

All resulting P. putida strains (P. putida KT2440, P. putida AtrpDC/pSEVA234, P. putida AtrpDC pheA/pSEVA234, P. putida/pSEVA234-aroG ^D1A ™, P. putida/pSEVA234-trpEG ^S40F , P. putida AtrpDC/pSEVA234-aroG ^D146 , P. putida AtrpDC/pSEVA234-trpEG ^SA0F , P. putida AtrpDC ApheA/pSEVA234-aro G ^D146N , and P. putida AtrpDCApheA/pSEVA234-trpEG ^{SA F} (Table 2) were characterized towards their properties as oAB producers in shake-flask cultures. Strains were grown in minimal medium according to Wierckx et al. (Wierckx et al., A EM, 2005, 71 :8221) at 30 °C and shaking at 250 rpm. The composition of the minimal medium was: 2 g/L (NH ₄ ) ₂ S0 ₄ , 1.63 g/L NaH ₂ P0 ₄ , 3.88 g/L K ₂ HP0 ₄ , 0.1 g/L MgS0 ₄ -7H ₂ 0, 1.0 mg/L CaCl ₂ , 10 mg/L EDTA, 5.0 mg/L FeSO.r7H.-0. 1.0 mg/L MnCl H ₂ 0, 2.0 mg/L ZnSO. 7H.-0, 0.2 mg/L CuS0 ₄ -5H ₂ 0, 0.4 mg/L CoCl ₂ -6H ₂ 0, 0.2 mg/L Na ₂ Mo0 ₄ -2H ₂ 0, and 5 g/L glucose. The pH was adjusted to 7 with 5 mol/L sodium hydroxide solution. The strains were pre-cultured in the minimal medium and inoculated into the main cultures of 50 mL in 500 mL-shake flasks with ODeoo of approximately 0.1. Cultures of plasmid-carrying strains were supplemented with 50 mg L kanamycin. Cultures of strains with a disruption of the trpDC gene were additionally supplemented with 0.1 mM L-tryptophane and cultures of strains with a disruption of the pheA gene were additionally supplemented with 1 mM L- phenylalanine. The lacF system was induced at the beginning of the midexponential phase by addition of 1 mM IPTG to enable oAB production. The characteristics of the strains as oAB producers were measured as described in example 3. During the cultivation ODeoo, glucose concentration, and the anthranilate concentrations were measured offline from 1.5 mL culture samples. A 1 mL aliquot of each sample was centrifuged for 5 min at 16000 x g (5415 ; Eppendorf, Hamburg) in reaction tube, after ODeoo measurement using a photometer. The supernatant was analyzed via HPLC-DAD (1 100; Agilent Technologies, Santa Clara (USA)) and YSI (YSI-Select 2700; Kreienbaum Neoscience GmBH, Lang enf eld). The collected data of Example 3 and Example 4 are shown in Table 4.

Table I Vectors and plasmids used and/or generated in the invention

JJejsigjiiation Description and relevant genotype Source/reference

expression of trpEG^^

Table 2 Bacterial strains used and/or generated in the invention

Designation Description and relevant Source/reference

genotype

Escherichia coli Strain DH5a; supE44, OlacU169 (Ϊ80 Hanahan and lacZDMIS), hsdR17 (rk-mk+), recAl, Meselson, 1983 endAl, thil, gyrA, re!A

Escherichia coli: :trpD9923 Strain W3110 with trpD9923 E. coli Genetic mutation Resource Center (Yale

University)

Escherichia coli: :trpD9923 ishpr In frame deletion of hpr in trpD9923 Balderas-Hernandez et strain al, 2009

Corynebacterium glutamicum Strain ATCC 13032 DSMZ

Corynebacterium Strain ATCC 13032 with empty This study glutamicum/pEKEx2 expression vector pEKEx2

Corynebacterium glutamicum In frame deletion of trpD (Cgl3032) This study AtrpD

Corynebacterium glutamicum In frame deletion of csm (Cgl0853) This study

Acsm

Corynebacterium glutamicum In frame deletion of trpD and csm This study AtrpDAcsm

Corynebacterium glutamicum Integration of trpDl into C. This study AtrpDv.tr pD\ glutamicum AtrpD

Corynebacterium glutamicum Integration of trpDl into C. This study AtrpDv.trpDl glutamicum AtrpD

Corynebacterium glutamicum Integration of trpD3 into C. This study AtrpDv.trpU glutamicum AtrpD

Corynebacterium glutamicum Integration of trpDS into C. This study AtrpD::trpD5 glutamicum AtrpD

Corynebacterium glutamicum Integration of trp/Xt into C. This study AtrpD::trpD6 glutamicum AtrpD

Corynebacterium glutamicum. AtrpD::trpD5 strain with empty This study AtrpD::trpD5lpEKEx2 expression vector pEKEx2

Corynebacterium glutamicum. AtrpD::trpD5Acsm strain with empty This study A trpD : : trpD5 Acs m. expression vector pEKEx2

Corynebacterium glutamicum In frame deletion of ptsG (Cgll360) This study AtrpD ::trpD5AptsG in AtrpD::trpD5 strain

Corynebacterium glutamicum In frame deletion of hpr (Cgll937) in This study AtrpD v.trpDS Ahpr AtrpD::trpD5 strain

Corynebacterium glutamicum In frame deletion of pepco (Cgll585) This study AtrpD ::trpD5 Apepco in AtrpD::trpD5 strain

Corynebacterium glutamicum In frame deletion of gpi (Cgl0851) in This study

AtrpD ::trpD5Agpi AtrpDv.trpDS strain

Corynebacterium. glutamicum In frame deletion of pyk (Cgl2089) in This study AtrpD ::trpD5 Apyk AtrpD::trpD5 strain

Corynebacterium glutamicum Expression of trpDl in AtrpD strain This study AtrpD lpEKEx2-trpDl

Corynebacterium glutamicum Expression of trpD2 in AtrpD strain This study AtrpD lpEKEx2-trpD2

Corynebacterium glutamicum Expression of trpD3 in AtrpD strain This study AtrpD lpEKEx2-trpD3

Corynebacterium. glutamicum Expression of trpD4 in AtrpD strain This study AtrpD lpEKEx2-trpD4

Corynebacterium glutamicum Expression of trpDS in AtrpD strain This study AtrpD IpEKEx2-trpD5

Corynebacterium glutamicum Expression of trpD6 in AtrpD strain This study AtrpD lpEKEx2-trpD6

Corynebacterium glutamicum Expression of csml in Acsm strain This study Acsm/pEKEx2 -csml

Corynebacterium glutamicum Expression of csm2 in Acsm strain This study Acsm/pEKEx2- csm2

Corynebacterium glutamicum. Expression of csm3 in Acsm strain This study AcsmlpEKEx2- csm3

Corynebacterium glutamicum Expression of csm4 in Acsm strain This study AcsmlpEKEx2- csm4

Corynebacterium glutamicum Expression of csm 5 in Acsm strain This study AcsmlpEKEx2-csm5

trpEG^

Tab!e 3 Primers used in t e invention Del-csm-1 GTC TCC CCA ATC AAA TCA TCA 79

CCC GGG ATC CAC TAA ACT TAA ACA GTC ACC TGC ATT AGT 80

Del-csm-2 CAT

TGT TTA AGT TTA GTG GAT CCC GGG CGC GGA AAA CTC GGA 81

Del-csm-3 T AA

Del-csm-4 TGA TGA TGT GCC CGT CCA CAG 82

Ko-csm-1 Crr GCC GAC GAG CGT AGA AAT 107

Ko-csm-2 GTT GAG GAC TAG GTT GAC TTG 108

GCC CTG CAG GAC GTG GCA GAA TAG TGT GCA TGA CTA ATC ΐ 83

Ex-csml

CAG GTG AC

GCC CTG CAG GAC GTG GCA GAA TAG TGT GC ^' G TGA CTA ATC ΐ 84

Ex-csm2

CAG GTG AC

GCC CTG CAG GAC GTG GCA GAA TAG TGT GCT TGA CTA ATC ΐ 85

Ex-csm3

CAG GTG AC

GCC CTG CAG GAC GTG GCA GAA TAG ATG ACT AAT GCA GGT 86

Ex-csm4

GAC

GCC CTG CAG GAC GTG GC A GAA T AG GTG AC T AAT GCA GGT 87

Ex-csm5

GAC

GCC CTG CAG GAC GTG GCA GAA TAG TTG ACT AAT GCA GGT 88

Ex-csm6

GAC

Ex-csm-rev CCC GGG ATC CTT ATC CGA GTT TTC CGC GTC C 89

GCC CTG CAG GTA AAA AAA GGA TTT GAT TCA TGA C TT CTC 70

Ex-lrpD 1

CAG CAA CAC TG

GCC CTG CAG GTA AAA AAA GGA TTT GAT TCG TGA CTT CTC 71

Ex- trpD2

CAG CAA CAC TG

GCC C TG CAG GTA AAA AAA GGA TTT GAT TCT TGA CTT' CTC 72

Ex- trpD3

CAG CAA CAC TG

GCC CTG CAG GTA AAA AAA GGA TTA TGA CTT CTC CAG CAA 73

Ex- trpD4

CAC TG

GCC C TG CAG GTA AAA AAA GGA TTG TGA CTT CTC CAG CAA 74

Ex- trpD5

CAC TG

GCC CTG CAG GT A AAA AAA GGA TTT TGA CTT CTC CAG CAA 75

Ex- trpD6

CAC TG

Ex-trpD-rev CCC GGG ATC CCT AGT CAT TGG AAG ACT CCT T 76

GC CCT GCA GGA GAT C TG AAA GGA GGC CCT TCA GAT GAA 109

Ex-aroG-1 TTA TCA GAA C GA CGA T

Ex-aroG-2 CCC GGG ATC CTT ACC CGC GAC GCG CTT TTA C 110

Ex-glnA - 1 TTA GAG GAG ACA CCA TAT GGC GTT TGA AAC CCC GGA AGA 97

GAA CCA TGG GC T AGC CTC GAG TTA GCA GTC GAA GTA CAA 98

Ex-glnA -2

TTC

GGC CCT GCA GGG AAA GGA GGC CCT TCA GAT GAC ACA ACC 99

Ex-aroL -1

TCT TTT TCT GA

Ex-aroL -2 CCC GGG ATC CTC AAC AAT TGA TCG TCT GTG C 100

GGC CCT GCA GGG AAA GGA GGC CCT TC A GAT GAA TGA TCA 101

Ex-aroK - 1

AAT TCA CTT AG

Ex-aroK -2 CCC GGG ATC CTT AAT CGA TTT CTA GAT GAT GC 102

ATT CGA GCT CGG TAG CCG GGG ATC CAC TAG ATC GAA ACC 103

JK038f GGC ATC JK039r CTG AAC TCG AGT CAG CCA TGC TCC TTC TC 104

JK040f GCA TGG CTG ACT CGA GTT CAG GGG CCT TGG GGC T 105

TAG AAG CTT GC A TGC CTG CAG GCA GTG AGT CGA CCA GGC 106

JK041r CAA AG

Table 4 Characteristics towards oAB production of bacterial strains used and/or generated in the invention (CDW: cell drv weight; Y: yield; μ„„ _ν : maximal growth rate; STY: space time yield).

AtrpDCApheA/

_P SEVA234-trpEG ^S40F

Example 5 - Cultivation of C. glutamicum strains in continuous fermenters with ceil retention Continuous fermentations were used for the production of anthranilate to achieve an optimized space- time yield. A cell retention system was applied during continuous fermentation to increase the biomass concentration without increasing the glucose concentration of the feed solution. Two different systems were used for cell retention experiments in lab scale: a hollow fibre filtration module from JM Water Separations (WaterSep Technology Corporation, Marlborough, MA, USA) with a cut of value of 750 k Da and a disposable centrifuge (Centritech Lab III) from Pneumatic Scale Angelus (Allen Road, Stow, OH. USA). Both systems are connected to 1 L lab-scale bioreactors (OmniFerm, HiTec Zang, Herzogenrath, Germany). The integration of the filtration module is shown in Figure 30 and the integration of the centrifuge is shown in Figure 31. Continuous fermentation of strain C. glutamicum AtrpD

A preculture of C. glutamicum AtrpO (see Table 2) was prepared using a sterile 250 ml. Erlenmeyer flask filled with 50 mL sterile BHIS medium containing 37 g/L Brain-Heart-Infusion (BHI) and 91 g/L sorbitol. Incubation for 24 h at 28 °C with a shaking frequency of 140 rpm resulted in an ODeoo of 6.0. A culture volume of 9 mL was transferred into 91 mL fresh CGXII-MOPS derived medium containing 42 g/L MOPS buffer, 20 g/L (NH ₂ SO ₄ , 5 g/L urea, 3.7 g/L Brain-Heart- Infusion, 9.1 g/L sorbitol, 1 g/L KH. PO.;. 1 g/L K2HPO4, 0.25 g/L MgS0 ₄ -7H ₂ 0, 0.01 g/L CaCk and 10 g/L glucose (autoclaved separately). The following components were added after sterile filtration: 2 mg L biotin, 0.1 mM L- tryptophan, 0.01 g/L MnS0 ₄ H ₂ 0, 0.01 g/L FeS0 ₄ -7H ₂ 0, 1 mg/L ZnS0 ₄ -7H ₂ 0, 0.2 mg/L CuSO.r5 H.-0, 0.02 mg/L NiCl ₂ -6H ₂ 0 and 0.03 g/L 3.4-dihydroxybenzoic acid. Incubation of the 100 mL culture volume in a 500 mL Erlenmeyer flask for 24 h at 28 °C with a shaking frequency of 140 rpm resulted in an ODeoo of 18.7. A culture volume of 27 mL was centrifuged, the supernatant was discarded and the pellet resuspended in 25 mL sterile isotonic PBS buffer. The suspension was injected into a fermenter with 1 I. sterile CGXII derived medium containing 20 g L (NH ₄ ) ₂ S0 ₄ , 1 g/L KH ₂ P0 ₄ , 1 g/L K. HPO,,. 0.25 g/L MgS0 ₄ -7H ₂ 0, 0.01 g/L CaCl ₂ , 100 μΕ/L polypropylenglycol and 10 g/L glucose (autoclaved separately). The following components were added after sterile filtration: 2 mg/L biotin, 0.1 mM L-tryptophan, 0.01 g/L MnS0 ₄ ¾0, 0.01 g/L FeS0 ₄ -7H ₂ 0, 1 mg/L ZnS0 ₄ -7H ₂ 0, 0.2 mg/L CuS0 ₄ -5H ₂ 0, 0.02 mg/L NiCl ₂ -6H ₂ 0 and 0.03 g/L 3.4-dihydroxybenzoic acid. The same medium was used for feeding during continuous operation. The dissolved oxygen concentration was controlled at 30 % by adjusting the stirrer speed between 200 and 1200 rpm. A constant gas flow rate of 0.2 1/h air was adjusted for aeration and the pH was controlled at pH 7 with 1 M NHiOH. The fermentation results are shown in Figure 32. During the batch phase within the first 24 h glucose was consumed and biomass and oAB where produced. The fermenter was switched to continuous operation after a cultivation time of 24 h by feeding 50 mL/h sterile medium and purging 50 mL h fermentation broth including biomass and oAB. A constant production of oAB was achieved during continuous cultivation. After a cultivation time of 100 h the cell retention was started using a cross-flow ultrafiltration as specified in Figure 30. An increase of dry cell weight was achieved by cell-retention as shown in Figure 32.

Continuous fermentation of strain C. glutamicum AtrpD::trpD5Agpi

A preculture of C. glutamicum Δ trpD: : trpD5Agpi (Table 2) was prepared using a sterile 250 mL Erlenmeyer flask filled with 50 mL sterile BHIS medium containing 37 g/L Brain-Heart-Infusion (BHI) and 91 g/L sorbitol. Incubation for 24 h at 28 °C with a shaking frequency of 140 rpm resulted in an ODeoo of 8.2. A culture volume of 10 mL was transferred into 90 mL CGXII-MOPS derived medium containing 42 g/L MOPS buffer, 20 g/L (NH ₄ ) ₂ S0 ₄ , 5 g/L urea, 3.7 g/L Brain-Heart-Infusion, 9.1 g/L sorbitol, 1 g/L KH.-PO.,. 1 g L K. HPO.,. 0.25 g/L MgS0 ₄ -7H ₂ 0, 0.01 g/L CaCl ₂ and 10 g/L glucose (autoclaved separately). The following components were added after sterile filtration: 2 mg/L biotin, 0.01 g/L MnS0 ₄ -H ₂ 0, 0.01 g/L FeS0 ₄ -7H ₂ 0, 1 mg/L ZnS0 ₄ -7H ₂ 0, 0.2 mg/L CuS0 ₄ -5H ₂ 0, 0.02 mg/L NiCl ₂ -6H ₂ 0 and 0.03 g/L 3.4-dihydroxybenzoic acid. Incubation of the 100 mL culture volume in two 250 mL Erlenmeyer flask for 24 h at 30 °C with a shaking frequency of 300 rpm resulted in an OD«x) of 20.5. A volume of 55 mL culture broth was injected into a fermenter with 1 I, CGXII derived medium containing 20 g/L (N¾) ₂ S0 ₄ , I g/L KH ₂ P0 ₄ , 1 g/L K. I I O,,. 0.25 g/L MgS0 ₄ -7H ₂ 0, 0.01 g/L CaCl ₂ , 100 μΙΛ. polypropylenglycol and 10 g/L glucose (autoclaved separately). The following components were added after sterile filtration: 2 mg/L biotin, 0.01 g/L MnS0 ₄ -H ₂ 0, 0.01 g/L FeS0 ₄ -7H ₂ 0, 1 mg/L ZnS0 ₄ -7H ₂ 0, 0.2 mg/L CuS0 ₄ -5H ₂ 0, 0.02 mg/L NiClr6H ₂ 0 and 0.03 g/L 3.4-dihydroxybenzoic acid. A high concentrated medium was used for feeding containing 20 g/L (ΝΗ,) ₂ 80 ₄ , 10 g/L KH ₂ P0 ₄ , 10 g/L K ₂ HP0 ₄ , 2.5 g/L MgS0 ₄ -7H ₂ 0, 0.1 g/L CaCl ₂ , 2 mL/L polypropylenglycol and 100 g/L glucose (autoclaved separately). The following components were added after sterile filtration: 20 mg/L biotin, 0.1 g/L MnS0 ₄ -H ₂ 0, 0.1 g/L FeS0 ₄ -7H ₂ 0, 10 mg/L ZnS0 ₄ -7H ₂ 0, 2 mg/L CuS0 ₄ -5H ₂ 0, 0.2 mg/L NiCl ₂ -6H ₂ 0 and 0.3 g/L 3.4- dihydroxybenzoic acid. The dissolved oxygen concentration was controlled at 30 % by adjusting the stirrer speed between 200 and 1400 rpm. A constant gas flow rate of 0.2 1/h air was adjusted for aeration and the pH was controlled at pH 7 with 1 M NH ₄ OH and 1 M HC1. The fermentation results are shown in Figure 33. During the batch phase within the first 40 h glucose was consumed and biomass and oAB where produced. The fermeter was switched to continuous operation after a cultivation time of 40 h by feeding 10 mL/h medium and purging 10 mL/h fermentation broth including biomass and oAB. At the beginning of the continuous operation the biomass concentration was not high enough to completely consume the added glucose in the feed solution resulting in an accumulation of glucose in the reactor as shown in Figure 33. After 136 h the cell concentration was high enough to completely consume the added glucose. Cell retention was started after 184 h using a centrifuge, as shown in Figure 31. An increase of dry cell weight was achieved by cell recycling as pictured in Figure 33. A continuous production of oAB was achieved during 400 h continuous fermentation.

Continuous fermentation of strain C. glutamicum AtrpD with complex carbon sources

A preculture of C. glutamicum AtrpD (Table 2) was prepared using a sterile 250 mL Erlenmeyer flask filled with 50 mL sterile BHIS medium containing 37 g/L Brain-Heart- In fusion and 91 g/L sorbitol. Incubation for 24 h at 28 °C with a shaking frequency of 140 rpm resulted in an ΟΟβοο of 7.53. A culture volume of 7 mL was transferred into 93 mL fresh CGXII-MOPS derived medium containing 42 g/L MOPS buffer, 20 g/L (NH ₄ ) ₂ S0 ₄ , 5 g/L urea, 3.7 g/L Brain-Heart-Infusion, 9.1 g/L sorbitol, 1 g/L KH2PO4, I g/L K2HPO4, 0.25 g/L MgSO,-7H;0, 0.01 g/L CaCl ₂ and 10 g/L glucose (autoclaved separately). The following components were added after sterile filtration: 2 mg/L biotin, 0.1 mM L- tryptophan, 0.01 g/L MnSfV lLO. 0.01 g/L FeS0 ₄ -7H ₂ 0, 1 mg/L ZnSO H.O. 0.2 mg/L CuS0 ₄ -5H ₂ 0, 0.02 mg/L NiCl ₂ -6H ₂ 0 and 0.03 g/L 3.4-dihydroxybenzoic acid. Incubation of the 100 mL culture volume in a 500 mL Erlenmeyer flask for 24 h at 28 °C with a shaking frequency of 140 rpm resulted in an ΟΟβοο of 21.3. A culture volume of 25 mL was centrifuged, the supernatant was discarded and the pellet resuspended in 25 mL sterile isotonic PBS buffer. The suspension was injected into a fermenter with 1 L sterile CGXII derived medium containing 20 g/L (NH ) ₂ S0 ₄ , 1 g L KH ₂ P0 ₄ , 1 g/L K ₂ HP0 ₄ , 0.25 g/L MgS0 ₄ -7H ₂ 0, 0.01 g/L CaCl ₂ , 100 μΕ/L polypropylenglycol 2000 and 10 g/L glucose (autoclaved separately). The following components were added after sterile filtration: 2 mg L biotin, 0.1 mM L-tryptophan, 0.01 g/L MnSOWLO. 0.01 g/L FeS0 ₄ -7H ₂ 0, 1 mg/L ZnS0 -7H ₂ 0, 0.2 mg/L CuS0 ₄ -5H ₂ 0, 0.02 mg/L NiCl ₂ -6H ₂ 0 and 0.03 g/L 3.4-dihydroxybenzoic acid. The same medium was used for feeding during the first 772 hours of continuous operation (feed medium 1). After 772 hours, the feed composition was changed to feed medium 2 containing 40 g/L (N¾) ₂ S0 ₄ , 0,5 g/L KH. PO.:, 5 mL/L corn steep liquor (Sigma Aldrich, C4648, lot number: MKBP5720V), 1 mL/L polypropylenglycol, 267 g/L glucose syrup (Cargill C*Sweet D02767, lot number: 03285882), autoclaved separately, to achieve a final glucose concentration in the feed solution of 200 g/L and 0.1 mM L-tryptophan (sterile filtered and added after autoclaving). The dissolved oxygen concentration was controlled at 30 % by adjusting the stirrer speed between 200 and 1200 rpm. A constant gas flow rate of 0.2 1/h air was adjusted for aeration and the pH was controlled at pH 7 with 1 M NH4OH. The results of the continuous fermentation with feed media 2 are shown in Figure 34. The preceded 772 h of continuous fermentation with feed medium 1 are not shown. Continuous cell retention was achieved by cross-flow ultrafiltration as specified in Figure 30 and a continuous production of anthranilate was achieved during the cultivation as shown in Figure 34. The continuous operation was temporarily interrupted as indicated by a stop phase in Figure 34 due to an accumulation of glucose. Feed and harvest were restarted after the glucose level decreased.

Example 6 -Adsorption/Desorption of aiithranilic Acid (A A)

In order to get the anthranilic acid (AA), a solution with higher concentration, A A was transferred from the aqueous solution into an organic solvent by performing an adsorption-desorption operation. To do so, the adsorption capacity of AA on different types of adsorbents was tested.

Zeolite Y (Zeolyst International, catalog no. CBV600) and ZSM5 (Siid-Chemie/Clariant catalog no. H-MFI-27) were selected as zeolites, which function as molecu lar sieves for different molecules. Hydroxyapatite (Caio(P04)e(OH)2) (Sigma-Aldrich catalog no. 289396) was tested due to its ability in the adsorption of A A and some other similar compounds in different solvents. Adsorption test: Adsorbents were already calcined at 300 °C for 3 h to release any remained moisture. A solution of A A (0.5 wt%) in water was prepared. 20 mL of this solution was transferred to a 50 niL flask containing 0.2 g adsorbent. After a certain period of time under stirring, the concentration of A A in water was analysed by HPLC. The decrease of AA concentration in water was considered as the adsorbed A A.

Synthesis of metal- exchanged zeolite: given the improved adsorption capacity of Ca-incorporated zeolite (following W. H. Goodman, US 4910336, 1990), Ca-exchanged zeolites were prepared by ion exchange to be tested in the adsorption of A A (S. M. Seo, S. Y. Coi, .1. M. Suh, K. J. Jung, N. H. Heo and W. T. Lim, Bull. Korean Chem. Soc. 2009, 30: 1703). 3 g of zeolite as powder was added to a solution of Ca(N03)2.4H20 (0.5 M). The slurry was stirred for 4h and then the solution was replaced by a fresh one and this procedure was repeated two more times. Finally, the solids were separated by centrifuge and dried at 80 °C and calcined at 300 °C for 3 h. Four other metal- exchanged zeolite Y samples using K. Na, Mg and Fe were prepared with same method as explained above. The samples were then labelled K-Y. Na-Y, Ca-Y, Mg-Y and Fe-Y. The pore size of zeolite H-ZSM5 can be in the range of 0.4-0.6 nm, preferably 0.5 nm. The pore size of zeolite H-Y can be in the range of 0.6 nm -0.9 nm, preferably 0.7-0.9 nm (e.g. as obtained from Zeolyst International, catalog number CBV600).

The adsorption test of A A was performed using the adsorbents mentioned above. The results of these tests are shown in Table 6.

Table 6: Adsorption of A A 0.5%

Zeolite Y showed the highest ability in adsorption of A A from aqueous solution. Ca exchanged Zeolite Y could improve the adsorption of A A to nearly 50%. Fe exchanged Zeolite Y could improve the adsorption of A A to nearly double. The tests were performed with the AA solution in distilled water and also in buffer solution containing (NH ₄ ) ₂ S0 ₄ (20 g/L), Na ₂ HP0 ₄ (1 g/L), KH. PO4 (1 g/L) and MgS0 ₄ (0.25 g/L). Later, MgS0 ₄ was eliminated from the buffer solution, because it caused the formation of insoluble magnesium anthranilate salt. Among the above metal-exchanged zeolites, Fe-Y showed highest A A adsorption capacity (higher than 50 g AA/ kg Fe-Y). Fortunately, the buffer solution contents did not influence the adsorption of AA and the adsorption capacities were similar to those in distilled water. The IR spectra of the Fe-Y sample after adsorption test showed some characteristic peaks of A A which can be an evidence for the existence of A A on the surface of Fe-Y most probably chelating with Fe.

Table 7 : A A adsorption capacities ^* of metal-exchanged zeolite Y in distilled water and buffered aqueous solution after 10 and 60 minutes of contacting.

Adsorbent Na-Y K-Y Mg-Y Ca-Y Fe-Y distilled water ( 10 24.2 29.6 29.2 36.2 50.6 min)

distilled water (60 22.6 31.6 34 42.2 54.2 r solution (10 25 27.4 27.6 36.8 51.2

buffer solution (60 25.9 22.3 22.8 25.6 45.3 min)

a g AA/ kg adsorbent

The Fe-Y after adsorption test and the fresh Fe-Y sample were analysed by ICP-MS. This analysis showed that the Fe/Al ratio in both samples were 1.25 and 1.3 respectively which are very close to each other. Therefore, the amount of Fe leaching was negligible. In case of C ^' a-Y. 64 % Ca leaching was detected after the adsorption test.

Decomposition of A A desorbed on Fe-Y was studied at the temperatures up to 550 °C using thermal gravimetric analyser equipped with a mass analyser (TGA-MS). As can be seen in Figure 7, no ANL was formed at the temperatures less than 400 °C. At higher than 400 °C, A A was decomposed directly to polyaniline which was observed as dark materials on the sample holder. A small amount of ANL was detected at higher than 400 °C. Therefore, direct thermal conversion of AA adsorbed on the adsorbent to ANL was not successful. Desorption of AA from adsorbent into liquid phase

The desorption test of A A from Fe-Y into water showed that the adsorption of A A by metal- exchanged zeolite in aqueous solution is reversible. Subsequently, the desorption of AA into an organic solvent was tested. 1 -dodecanol was selected as organic solvent due to the high solubility of A A in it and also its very low miscibility in water (0.004 g/L). Besides, its boiling point (259 °C) is much higher than aniline (183 °C) and water so that its separation from the final mixture was more convenient. Desorption of AA from Fe-Y into 1 -dodecanol was performed by suspending 0.2 g Fe-Y containing 10.8 mg A A in 2 rriL 1 -dodecanol. The slurry was stirred for 0.5 h at the temperature range of 25-120 °C. The desorption results are shown in Figure 8 in which maximum 27.8% A A could be desorbed into 1 -dodecanol phase at 120 °C. Examplc 7 - reaction kinetics of the decarboxylation of anthranilic acid to aniline by thermal decarboxylation in an organic solvent

Decarboxylation of AA dissolved in 1-dodecanol

The decarboxylation of AA 3 wt% dissolved in 1 -dodecanol at 160 °C was tested. Catalysts with different characters were screened. The results can be seen in Figure 9. The blank test without any catalyst showed only 1.4% conversion, while in the presence of zeolite Y (CBV600, "G0257") resulted in more than 70% conversion to ANL in lh.

High AA conversion over zeolite Y (CBV600, "G0257") can be due to its highly acidic character and also the pore size (0.7-0.8 nm). ZSM5 (MFI-27) despite of possessing acidic character, has smaller pore size (0.5 nm) so that A A molecules cannot penetrate into them and consequently do not have access to active sites. Hydrotalcite (HTC, Mg6 Al2(C03 ) (OH) 16. 4¾0) has basic character and the low conversion of A A indicates that basic sites are not as active as acidic sites in the decarboxylation of A A. The addition of Na to the zeolite H-Y ("G0257") decreases its acidic character and thus a low conversion of A A was achieved. The ammonia form (CBV500, "G055") also has a lower number of acidic sites than H-Y (CBV600, "G0257").

As can be seen in Figure 10, the decarboxylation of AA profile in 1 -dodecanol at 160 °C and 180 °C showed that the A A conversion and the ANL formation follow zero order kinetics. The simulated models for the reaction at 160 °C and 180 °C could calculate rate coefficients (k) as 0.235 mol L .h ^_1 and 0.522 mol.L/'.tf ¹ respectively. The selectivity of these reactions to ANL was 100% and a mass balance of 95-110% (with an average of 100.4%) was observed.

Decarboxylation of AA dissolved in Aniline

Aniline is a much better solvent for A A than dodecanol and, being also the product of the decarboxylation, its use presented great advantages compared to 1-dodecanol. We found that up to 30% AA could be dissolved in aniline at room temperature (20°C), up to 40%o A A could be dissolved in aniline at 50°C and that up to 50% A A could be dissolved at 90°C.

The decarboxylation of A A 40 wt%> dissolved in aniline at 160, 180 and 200°C was tested. The blank test without any catalyst showed, as before, negligible conversion, while in the presence of zeolite Y (CBV600, "G0257") resulted in more than 99% conversion to ANL in 2h. The results are shown in Figure 35. We were also able to fit the data according to a simple kinetic model: Where "r" is the reaction rate in g aniline per g catalyst per hour, "k°" is the kinetic constant also in g aniline per g catalyst per hour, "E _a " is the reaction activation energy in kJ/mol, "R" is the ideal gas costant of 8.31 J / K mol, "T" is the temperature in K, "AA" is the concentration of AA and "n" is the reaction order. Figure 35 also shows the fittings of this reaction model at the different temperatures. The results of the fittings can be summarized as follows:

k°=55492 g AN / g CAT h

E _a =37.4 kJ/mol

n= 0.16

Fi ure 35 shows the decomposition of 2-aminobenzoic acid (anthranilic acid, AA) in aniline with a catalyst. The kinetics of dec arboxy lation of anthranilic acid (AA) dissolved 40 wt% in aniline in the presence of zeolite Y at 160 °C, 180 °C and 200°C is shown.

Anthranilic Acid (40 wt%) was dissolved in aniline. At this concentration the acid was perfectly soluble in the organic solvent once heated to about 50°C. 80 mL of the above solutions are then transferred into an autoclave of 160 mL, 1.5 g of zeolite catalyst are added and heated to 160 °C, 180 °C or 200°C. Samples were taken at different time intervals and analysed by HPLC methods to determine the rate of aniline formation.

Decarboxylation of AA dissolved in Aniline in presence of water

As the antranilic acid separated from the crystallization process could present some residual moisture (approximately 10%), we decided to look at the kinetic of the reaction in presence of water. Experiments analoge to the previous one were conducted, except that 10 wt% water was added on purpose to the reaction mixture. We found that the overall kinetic and reaction profile remains unchanged. Results for the reaction temperature 200°C are presented in Figure 36. Figure 36 shows the decomposition of 2-aminobenzoic acid (anthranilic acid, AA) in aniline with a catalyst. The kinetics of decarboxylation of anthranilic acid (AA) dissolved 40 wt% in aniline and in 10% Water-90% Aniline in the presence of zeolite Y at 200°C is shown.

Anthranilic Acid (40 wt%) was dissolved in aniline. At this concentration the acid was perfectly soluble in the organic solvent once heated to about 50°C. 80 mL of the above solutions are then transferred into an autoclave of 160 mL, then 10% (8 g) of water was added, 1.5 g of zeolite catalyst are added and heated to 160 °C, 180 °C or 200°C. Samples were taken at different time intervals and analysed by 1 1 PLC methods to determine the rate of aniline formation. Decarboxylation of AA dissolved in Aniline from different organic sources

According to the present invention, anthranilic acid could be produced biologically by fermentation of different sugar mediums. It is expactable that the differences in trace elements present in the different media (i.e. hydrolized corn stach, sugar cane juice, or glucose) could affect the anthranilic acid that should be further decarboxylated to aniline. To verify if these differences could interfere with the catalyst and the decarboxylation reaction, different anthranilic acid, separated by crystallization from different growth media were tested. The comparison was always done against the chemically pure anthranilic acid purchased by the laboratory supplier Sigma Aldrich.

The different A A was tested in the same conditions as the previous tests starting from a 40 % solution of A A in Aniline at 180°C. Results are summarized in figure 47. All the A A sources were fully converted to aniline in less than 90 minutes. The fastest conversion is shown by the pure chemical material (Aldrich). This is comparable with the material produce by fermentation of chemically pure glucose (VN35). Slightly slower is the amterial isolated from hydrolized corn starch (VN32 and 33) followed by the material isolated from sugar cane juice (VN 34). This sequence follows the degree of refining of the media used. The purest sugar source resulted in faster decomposition of the A A isolated from it. This could be explained by the presence of trace elements, like ions or minerals, which could interfere with the zeolite catalyst. However the effect is so limited that would not interfere with the decarboxylation reaction, which proceed very quickly to completion independently from the source of the A A.

Figure 37 shows the decomposition of 2-aminobenzoic acid (anthranilic acid, AA) in aniline with a catalyst. The kinetics of dec arboxy lation of anthranilic acid (AA) dissolved 40 wt% in aniline in the presence of zeolite Y at 180°C is shown.

Anthranilic Acid was obtained by crystallization from different sugar media like chemically pure glucose, hydrolysed corn stach and sugar cane juice. The resulting isolated Anthranilic Acid (40 wt%) was dissolved in aniline. At this concentration the acid was perfectly soluble in the organic solvent once heated to about 50°C. 80 mL of the above solutions are then transferred into an autoclave of 160 mL, 1.5 g of zeolite catalyst are added and heated to 180 °C. Samples were taken at different time intervals and analysed by H I X ^' methods to determine the rate of aniline formation.

Catalyst Stability for the Decarboxylation of AA dissolved in Aniline from Hydrolized corn starch

Considering the effect of the trace elements observed before, we tested the catalyst stability over several runs. We performed this by recycling the catalyst and by using it in the following reaction without any wasching step, following the reaction procedure described before. The results are presented in Figure 38. The difference in activity between the freshly prepared catalyst and the reused one is minimal. The reused seems to be even faster or better than the fresh system. This showed that there is no poisoning or inhibition effect of the trace elements on the catalyst, which remains active in the decarboxylation reaction. To further confirm that we analysed post mortem the catalyst and compared it with the freshly prepared one. The 1 R spectrum of the 2 catalyst is shown in Figure 39. Beside some absorbed aniline species, only a small decrease in OH signal is present. This OH could be quenched after exchange with other ions (trace elements), however they are probably not the active site for the catalysis, as the strong acidity of the zeolites comes from the A l-O-Si bridge.

Figure 38 shows the decomposition of 2-aminobenzoic acid (anthranilic acid, AA) in aniline with a catalyst. The kinetics of dec arboxy lation of anthranilic acid (AA) dissolved 40 wt% in aniline in the presence of zeolite Y at 180°C is shown.

Anthranilic Acid was obtained by crystallization from hydrolysed corn stach. The resulting isolated Anthranilic Acid (40 wt%) was dissolved in aniline. At this concentration the acid was perfectly soluble in the organic solvent once heated to about 50°C. 80 mL of the above solutions are then transferred into an autoclave of 160 mL, 1.5 g of zeolite catalyst are added and heated to 180 °C. The catalyst was isolated and recycled for a subsequent reaction without further treatment or washing. Samples were taken at different time intervals and analysed by HPLC methods to determine the rate of aniline formation.

Figure 39 shows the IR Spectrum of a freshly prepared H-Y catalyst and one after reaction as above using A A isolated from hydrolyzed corn starch. The typical bands for aniline are marked. Absorption of this higly basic specie take place indeed. The OH region is also highlighted. A reduction of OH signal is compatible with some partial ion exchange that could take place with trace elements present, however the OH groups seams not be active in the catalysis, but is probably the Al-O-Si bridge the responsible.

Example 8 dissolution and crystallization of oAB from fermentation broth An industrial sugar source was used for fermentation of Corynehacterium glutamicum. 100 ml of the broth with a content of 2 1 ,5 g/L glucose, 3,7 g/L lactate was mixed with 1 Og anthrani l ic acid in a 250 ml laboratory reactor. The slurry had a pH-value of 4.7. The solubltiy f oAB was about 4.5 g/L at room temperature at pH 4.5. A fterwards 9.12 g NaOH (32%) were added in 4 portions until a solution was present. The solution had a pH-value of 8.3. F llowing 7,. g HC1 (37%) w ere added within 1 hour. By dosing of HC1 first nucleation took place as from a pH- value of 5.8. The number of crystals increased during dosing. At the end of dosing the suspension had a pH-value of 3.6. Afterwards the slurry was stirred for 1 hour. Finally the slurry was filtered.

Fi ltration:

The fi ltration was performed accord ing to the VDI guideline "VDI 2762" at 0.2 bar and a filter area of 1 2.6 cm ² . 125 g suspension was filtered and washed with 25 g water. 91 g mother l iquor and 23.1 g wet solid were measured. The resistance of the fi lter cake being determined according to "VDI 2762" accounted to 5 x 10 ⁱ⁰ 1/m ² . The filter cake was dried in a drying furnace. A fter drying 9.1 g of dry solid were measured, corresponding to a yield of 91%.

The fraction of anthrani l ic acid being in the mother l iquor was 0.72%. The ash content of the solid was determined as an indication for the salt content of anthranilic acid. The ash content was determined to be at 0.55%.

Example 9 - Solubltiy of oAB in I -dodecanol

The solubility of oAB in dodecanol was tested with increasing temperature.

SLE = solid liquid equi l ibrium

SLE oAB / 1 -dodecanol input 1 -doc ecanol

fuli= 270,74 g

empty= 170,67 g

clean= 100,07 g

Input oAB:

No. target ful l empty c lean sum concentration temperature

g g g g g wt % °C

1 5 25,49 20,36 5,13 5, 13 4,88 40

2 10 30,36 20.4 1 9,95 15,08 13, 10 68

3 1 5 35,22 20,32 14,90 29,98 23,05 93

4 25 53,81 29,10 24,7 1 54,69 35,34 1 1 1

5 25 53,51 28,87 24,64 79,33 44,22 1 19 6 25 53,72 29,19 24,53 103,86 50,93 123

The solubility of A B in 1 -dodecanol was increasing with rising temperature.