WO2013023999A1 - Regioselective carboxylation of nonnatural substrate compounds using decarboxylases - Google Patents

Abstract

The present invention relates to the regioselective carboxylation of nonnatural substrate compounds using a decarboxylase.

Description

REGIOSELEC IVE CARBOXYLATION OF NONNATURAL SUBSTRATE COMPOUNDS USING

DECARBOXYLASES

FIELD OF THE INVENTION

The present invention relates to the regioselective carboxylation of nonnatural substrate compounds using a decarboxylase.

BACKGROUND OF THE INVENTION

Due to the increasing abundance of carbon dioxide as the major greenhouse gas on one side and dwindling petroleum resources on the other side, the

development of CO₂-fixation reactions for the production of organic molecules is one of the major challenges in synthetic organic chemistry. Based on the electron- accepting properties of CO₂, an increasing number of chemical CO₂-fixation reactions have been reported (1 ). However, only few made it to the industrial level, most notably the Kolbe-Schmitt-reaction and the synthesis of urea. More recently, the synthesis of carbonate esters (via carboxylation of epoxides) (2) and the production of lactones (via carboxylation of butadiene) (3) have been developed. Despite these isolated success stories, the use of CO₂ as a raw material for organic synthesis is still heavily underdeveloped.

Biological C0₂-Fixation in Biosynthesis

Natural processes for the fixation of carbon dioxide proceed on large scale. Autotrophic organisms synthesize complex organic molecules (e.g. carbohydrates, fatty acids) by assimilation of CO₂ either through chemosynthesis (at the expense of an external reducing agent, such as H₂, H₂S, S°, Fe²⁺, NH₃, NO₂ ^") or via

photosynthesis (catalysed by ribulose-1 ,5-bisphosphate carboxylase/oxygenase, RuBisCO) (4).

To date, four major pathways of biological CO₂-fixation are known (5):

i) Calvin-Benson-Bassham-cycle (6),

ii) reductive TCA (Arnon-Buchanan) cycle (7),

iii) reductive Acetyl-CoA (Wood-Ljungdahl) pathway (8),

iv) Acyl-CoA carboxylase pathways: 3-hydroxypropionate/malonyl-CoA cycle (8), 3-hydroxypropionate/4-hydroxybutyrate cycle (10), dicarboxylate/4-hydroxy- butyrate pathway (1 1 ), and the ethylmalonyl-CoA pathway (12).

In chemosynthesis, carboxylation reactions are mediated by the cofactor biotin, whereas photosynthesis depends on Mg²⁺-dependent RuBisCO. These reactions are high-energy demanding (13, 14)· and require ATP for the fixation of CO₂ onto biotin, which serves as activated C0₂-derivative for the carboxylation of Acyl- CoA (15). Alternatively, reductive carboxylation (e.g. pyruvate + C0₂ yielding malate) depends on NADPH.

Due to the high complexity of these specialised biosynthetic pathways, their application for the production of well-defined organic materials requires sophisticated metabolic engineering.

CO -fixation in biodegradation-detoxification pathways

In contrast to highly directed biosynthetic carboxylation pathways, several unspecific C0₂-fixation reactions occur in catabolic (biodegradation) pathways. The prime goal of biodegradation is making toxic compounds more hydrophilic, thereby enhancing their solubility in water in order to reduce their affinity for sensitive lipophilic biological components, such as membranes, proteins, etc. In order to make these processes more efficient, catabolic enzymes usually possess relaxed substrate specificities and are able to act on a large variety of substrates. Carboxylases involved in biodegradation have been exploited to an astonishingly limited extent for preparative biotransformations (16, 17, 18). Although the exact role of these enzymes and their 'true' substrates in nature are currently unknown, it is assumed that they serve as detoxification mechanism under oxygen-limited conditions for lipophilic electron-rich (hetero)aromatics, such as phenols and pyrrols, to yield the corresponding water-soluble carboxylic acids (19). Due to the lack of a strong thermodynamic driving force, equilibria close to unity are common. In view of preparative-scale applications, equilibria may be driven towards the

synthesis/carboxylation direction by the use of excess of C0₂— applied either under i) atmospheric pressure,

ii) concentrated carbonate/bicarbonate solution or

iii) in supercritical form (20).

Matsui T. et al. (Appl Microbiol Biotechnol. 2006 Nov;73(1 ):95-102) describe the regioselective carboxylation of the natural substrate 1 ,3-dihydroxybenzene by 2,6-dihydroxybenzoate decarboxylase of Pandoraea sp. 12B-2. The biocatalyst was employed as wild-type whole cells of Pandoraea sp. 12B-2 induced with 2,6- dihydroxybenzoic acid. In addition to the natural substrate (1 ,3-dihydroxybenzene, conv. = 48%) only one more substrate (1 ,2-dihydroxybenzene, conv. = 22%) was carboxylated at reasonable rates. The conversion of phenol to the p-carboxylated product 4-hydroxybenzoate (conv. = 1 .5%) is within the analytical error margin. In addition sixteen further nonnatural substrates (benzene, phenol, 1 ,2- dihydroxybenzene, 1 ,4-dihydroxybenzene, 1 ,2,3-trihydroxybenzene, 1 ,3,5- trihydroxybenzene, o- and m-chlorophenol, o- and m-cresol, o- and m- methoxyphenol, onitrophenol, o, m- and p-aminophenol) were tested, however, none of them were carboxylated at all by whole cells of Pandoraea sp. 12B-2.

DETAILED DESCRIPTION OF THE INVENTION

Carboxylation reactions are important biosynthetic processes, the enzymes involved are usually highly substrate-specific and consequently it was rather surprising, that nonnatural (man-made) substrates have been carboxylated at reasonable rates.

The present invention relates to a process for the regioselective carboxylation of nonnatural substrate compounds using a decarboxylase.

One aspect of the invention relates to such a process, wherein the

carboxylation is an o- or β-carboxylation.

A further aspect of the invention relates to such a process, wherein at least

25% of the nonnatural substrate are converted, preferably at least 30%.

A further aspect of the invention relates to such a process, wherein the decarboxylase is selected from the group consisting of dihydroxybenzoic acid decarboxylase (DHBD), dihydroxyphthaiic acid decarboxylase (DHPD), salicylic acid decarboxylase (SAD), pyruvate decarboxylase (PDC), phenolic acid decarboxylase (PAD), and ferulic acid decarboxylase (FDC).

A further aspect of the invention relates to such a process, wherein the nonnatural substrate is compound of formula I,

wherein

A is a mono- or bicyclic aromatic or heteroaromatic ring; and

each R1 is independently from one another selected from the group consisting of halogen, carboxy, -(CH₂)_nOH, -NH₂, Ci.₃alkoxy, Ci.₃alkyl, d.₃alkenyl, and

n is 0, 1 , 2, 3 or 4.

A further aspect of the invention relates to such a process, wherein the nonnatural substrate is hydroxystyrene. A further aspect of the invention relates to such a process, wherein the side chain R = -CH=CH₂ of said hydroxystyrene is carboxylated.

A further aspect of the invention relates to a process for the regioselective carboxylation of natural and nonnatural substrate compounds, wherein an organic cosolvent is added in order to enhance the solubility of the substrate and/or product(s) and to increase the conversion.

A further aspect of the invention relates to such process, wherein the organic cosolvent is selected from the group consisting of D SO, DMF, EtOH, acetone, acetonitrile, A/-methyl-2-pyrrolidone, glycerin, formamide, and 1 ,4-dioxane.

A further aspect of the invention relates to a process for manufacturing a carboxylated compound, comprising the steps of:

i) resuspending host cells which overexpress the decarboxylase in a buffer, ii) adding the non-natural substrate compound;

iii) adding a carbon source; and

iv) obtaining the carboxylated compound.

A further aspect of the invention relates to such process, wherein said carbon source is C0₂ or a carbonate buffer.

A further aspect of the invention relates to a compound obtained in such processes for use in pharmaceutical, cosmetic and/or food industry.

Specificity is a key feature of enzymatic reactions. Enzymes allow many chemical reactions to occur within the homeostasis constraints of a living system.

Enzymes function as organic catalysts. In the context of this invention the non-natural substrate relates to a substrate which is not known to be natural to the respective enzyme.

In the context of this invention mono- or bicyclic aromatic or heteroaromatic ring refers to a mono- or bicyclic aromatic or heteroaromatic ring system comprising 5 to 10 ring atoms selected from C, N, O and S. Examples of such aromatic or heteroaromatic mono- or bicyclic ring systems include but are not limited to phenyl, pyridyl, pyrrolyl, pyrrazolyl, napthyl, quinolinyl, isoquinolinyl, tetrahydroisoquinyl, indolyl, indazolyl, benzimidazolyl, benzthiadiazolyl and imidazopyridinyl.

In the context of this invention halogen refers to F, CI, Br and I.

DESCRIPTION OF THE FIGURES

Fig. 1 : Carboxylation reaction of resorcinol using 2,6-DHBD Rsp applying various organic cosolvents categorized in four groups Fig. 2: Carboxylation of catechol applying SAD Tm in presence of selected water-miscible organic cosolvents at concentrations of 0% - 50% (v/v)

Fig. 3: Carboxylation reaction of resorcinol applying 2,6-DHBD Rsp using various bicarbonate salts; IL = ionic liquid (1 -butyl-3-methylimidazolium hydrogen carbonate solution -50% in methanol/water 2/3; 200μΙ_)

Fig. 4: Substrate concentration study of the carboxylation of resorcinol employing 2,6-DHBD Rsp as biocatalyst

Fig. 5: Carboxylation reaction of resorcinol applying 2,6-DHBD Rsp at different temperatures

Fig. 6. Carboxylation reaction of resorcinol applying 2,6-DHBD Rsp at 10°C and different pH values.

Fig 7: Cosolvent study for the carboxylation of resorcinol and catechol applying 2,6-DHBD Rsp and SAD Tm as biocatalyst

EXAMPLES

Example 1 : Search for enzymes (decarboxylases)

A search for appropriate decarboxylases in common databases (Brenda, NCBI) based on available amino acid sequences was conducted, with the main focus of enzymes catalysing the (de)-carboxylation of aromatic (phenolic) and hetero- aromatic substrates. A phylogenetic tree was calculated showing that there are three main groups of decarboxylases for which amino acid sequences are available:

i) 2,3-dihydroxybenzoic acid decarboxylases (2.3-DHBD)

ii) 4-hydroxybenzoic acid decarboxylases (4-HBD)

iii) 4,5-dihydroxyphthalic acid decarboxylases (4,5-DHPD)

Out of the phylogenetic tree four enzymes were chosen in which for only one enzyme (2,3-DHBD from Aspergillus oryzae) evidence on protein level was described regarding to the protein knowledge database (UniProtKB/Swiss-Prot). For all other enzymes the amino acid sequence was only predicted and the three other enzymes (4,5-DHPD) were more or less randomly selected from the group of 4,5- dihydroxyphthalic acid decarboxylases.

Enzymes:

(i) 2,3-dihydroxybenzoic acid decarboxylase from Aspergillus oryzae (Swiss- Port: evidence on protein level)

(ii) 4,5- dihydroxyphthalic acid decarboxlase from Bordetella pertussis (Swiss- Port: predicted) (iii) 4,5-dihydroxyphthalic acid decarboxlase from Comamonas testosteroni (Swiss-Port: predicted)

(iv) 4,5-dihydroxyphthalic acid decarboxlase from Verminephrobacter eiseniae (Swiss-Port: predicted) Example 2: Overexpression of the decarboxylases

The genes were synthesized at geneart (Germany, Regensburg) and subcloned in a common pET vector (pET 21 a). The obtained plasmid was

transformed in a standard Escherichia coli host [BL21 (DE3)] for overexpression using IPTG for induction.

Successful overexpression was obtained for the decarboxylases of the

Aspergillus, Comamonas and Bordetella strain whereas no overexpression was observed in case of the Verminephrobacter strain.

In order to verify if the overexpressed proteins are soluble the cells were disrupted using ultrasonication and the separated supernatant and remaining pellet were analyzed by SDS-page. The decarboxylases of the Aspergillus and Bordetella strain were nicely soluble whereas in case of the Comamonas strain the protein was mostly insoluble and remained in the pellet and in case of the protein of the

Verminephrobacter strain no overexpression at all was observed.

In order to enhance the solubility and overexpression of the two enzymes from the Comamonas and Verminephrobacter organism, respectively, two strategies were applied. For the solubility problem on the one hand a different expression host - BL21 (DE3)pLysS - was used, whereas for a better protein expression a

coexpression with a chaperone was carried out. The use of the BL21 (DE3)pLysS expression host was not successful since no overexpression at all for both enzymes was observed. In case of the coexpression using a chaperone the overexpression of the 4,5-DHPD of Verminephrobacter was successful and a protein band could be detected by SDS-page after induction with IPTG.

Enzymes for the o-carboxylation

2,3-Dihydroxybenzoic acid decarboxylase from Aspergillus oryzae (2,3- DHBD Ao), salicylic acid decarboxylase from Trichosporon moniliiforme (SAD Tm) and 2,6-dihydroxybenzoic acid decarboxylase from Rhizobium sp. (2,6-DHBD Rsp) were synthesized and ligated into a pET 21 a (+) vector at Geneart AG (Germany, Regensburg). The obtained plasmids were transformed in a standard E. coli host [BL21 (DE3)] using IPTG-induction for overexpression. Enzymes for the β-carboxylation

Phenolic acid decarboxylases from Lactobacillus plantarum (PAD Lp) and from Bacillus amyloliquefaciens (PAD Ba) subcloned in a pET 28a (+) vector were kindly provided by Byung-Gee Kim (School of Chemical and Biological Engineering, Institute of Bioengineering, Seoul National University). The plasmids were

transformed in a standard E. coli host [BL21 (DE3)] using IPTG-induction for overexpression.

Phenolic acid decarboxylases from Mycobacterium colombiense (PAD Mc), Methylobacterium sp. (PAD Msp), Pantoea sp. (PAD Psp), Lactoccocus lactis (PAD LI), and ferulic acid decarboxylase from Enterobacter sp. (FDC Esp) were synthesized at geneart AG (Germany, Regensburg) and subcloned in a common pET vector [pET 21 a (+)]. The obtained plasmids were transformed in a standard E. coli host [BL21 (DE3)] using IPTG-induction for overexpression.

The heterologous overexpression of all enzymes was performed as follows: For preculturing 500 ml_ LB medium [Trypton (10 g/L), yeast extract (5 g/L), NaCI (5 g/L)] supplemented with the appropriate antibiotics [ampicillin (100 μg/mL) for 2,3- DHBD Ao, SAD Tm, 2,6-DHBD Rsp, PAD Mc, PAD Msp, PAD Psp. PAD LI and FDC Esp; kanamycin (50 μg/mL) for PAD Lp and PAD Ba] were inoculated with 3 mL ONC (starter culture) and incubated at 37°C and 120 rpm until an OD₆₀o of 0.6- 1 .0 was reached. Then IPTG [450 μg/mL, 2 mM (for 2,3-DHBD Ao) or 175 μg/mL, 0.5 mM (for all other decarboxylases)] was added for induction and the cells left overnight at 20 °C and 120 rpm. The next day the cells were harvested by

centrifugation (20 min, 8000 rpm, 4°C), washed with phosphate buffer (5 mL, 50 mM, pH 7.5) and centrifuged under the same conditions. The cell pellet thus obtained was resuspended in phosphate buffer (5 mL, 50 mM, pH 7.5), shock frozen in liquid nitrogen followed by lyophilization. Lyophilized cells were stored at +4°C.

Table 1 : Summary of all overexpressed enzymes and wild-type organisms applied for the o- and -carboxylation.

ocarboxylation entry name enzyme organism

2,3-DHBD Ao 2,3-Dihydroxybenzoic acid decarboxylase Aspergillus oryzae

2,6-DHBD_Rsp 2,6-Dihydroxybenzoic acid decarboxylase Rhizobium. sp.

Trichosporon

SAD Tm Salicylic acid decarboxylase

moniliiforme

jS-carboxylaiioo

entry name enzyme organism

PAD Lp Phenolic acid decarboxylase Lactobacillus plantarum

Bacillus

PAD Ba Phenolic acid decarboxylase

amyloliquefaciens

Mycobacterium

PAD Mc Phenolic acid decarboxylase

colombiense

PADJV!sp Phenolic acid decarboxylase Methylobacterium sp.

PAD Psp Phenolic acid decarboxylase Pantoea sp.

PADJJ Phenolic acid decarboxylase Lactoccocus lactis0 FDC Esp Ferulic acid decarboxylase Enterobacter sp.

Mycobacterium

1 NCIMB 10420 wild-type microbial cells

paraffinicum

2 FCC073³ wild-type microbial cells Fusarium solani3 DSM10 wild-type microbial cells Bacillus subtilis 168

FCC = Fab-Crew-Collection, in-house strain collection;

According to SDS-PAGE analysis, successful overexpression was obtained for all enzymes mentioned above, which were employed as lyophilized whole-cell biocatalyst to a range of substrates. Separate blank experiments using empty E. coli host cells ensured that the host did not display any enzyme activity for the

carboxylation (see below).

In order to examine the potential of both the ocarboxylation as well as the β- carboxylation further substrates were investigated. A summary of all substrates which were carboxylated at the aromatic ring system in o/ )o-position to the OH group applying 2,3-DHBD Ao, 2.6-DHBD Rsp and SAD Tm is shown in Table 2. The obtained results for the β-carboxylation applying PAD Lp, PAD Ba, PAD-Mc, PAD Msp, PAD Psp, PAD LI, FDC Esp and the wild-type microbial cells (NCIMB 10420, FCC073 and DSM10) are summarized in Table 3. Table 2: Substrate scope of the ocarboxylation applying benzoic acid decarboxylases

enzymes overexpressed in E. coii 2,3-DHBD_Ao 2,6-DHBD_Rs SAD Tm substrates

Arrows indicate observed carboxylation site; ^a2,4- as well as 2,6-DHBA were formed during the carboxylation process; ^bonly 2,6-DHBA was obtained as carboxylated product.

Table 3: Substrate scope of the S-carboxylation applying phenolic acid decarboxylases as well as wild-type microbial cells

enzymes conv. [%] conv. [%] conv. [%] conv. [%] conv. [%] conv. [%] conv. [%]

PAD_Lp^a 2° 30 3 22^d 3° 17° 22

PAD_Ba^a 18° 26° 5 34^' 10 ^: 6° 20

PAD_Mc^a 18^d 28 1 1 1 28 30 26

PAD_ sp^a 19^c 25° 4 24° 10^c 26° 15

5 _{1 8}cd

PAD_Psp^a 17° 20° 35^c 13° 14

PAD_Ll^a 4 8 4 35^c 2 26° 10

FDC_Esp^a 23° 21 ° 5 31 ° 20^c 21 ° ^d 18

NCIMB 10420^b 33 12 1 5 42 8 7

FCC073¹ 1 1 13 <1 <1 28 3 1

DS 10^b 20 18 <1 3 37 5 4

Arrows indicate observed carboxylation site. ^aE. co/ whole cells containing the overexpressed enzyme, ^bwhole wild-type microbial cells; Conversion obtained applying 20% v/v acetonitrile; ^d minor side product < 5% was detected;

Example 3: Screening - decarboxylation

In a first screening in order to check whether the enzymes are active at all the activity towards the decarboxylation was investigated since all four enzymes are decarboxylases applying the corresponding natural substrates. For the 2,3-DHBD from A. oryzae the natural substrate is 2,3-dihydroxybenzoic acid which is supposed to be decarboxylated yielding the corresponding diol (catechol). For the 4,5-DHPDs from B. pertussis, C. testosteroni and V. eiseniae the natural substrate is 4,5-dihydroxyphthalic acid obtaining 3,4-dihydroxybenzoic acid as expected product.

2,3-Dihydroxybenzoic _ . , . 4,5-Dihydroxyphthalic 3,4-Dihydroxybenzoic acid Catechol acid acid

Scheme 1 : Decarboxylation of 2,3-dihydroxybenzoic acid and 4,5- dihydroxyphthalic acid

General screening procedure of the decarboxylation

For the activity screening either lyophilized whole cells (30mg) resuspended in phosphate buffer (1 ml_, pH 7.0, 100 mM) or a cell suspension obtained after cell disruption of lyophilized or freshly harvested cells using ultrasonication (1 ml_, supernatant and pellet) was used. The substrate was added as a stock solution (10 - 20 μΙ_) to a final concentration of 10 mM. The mixture was shaken at 30 °C and

120 rpm. After 24h the reaction mixture was acidified with HCI (3M, 100 μΙ_) and the products were extracted with EtOAc (500 μΙ_). The organic layer was dried over Na₂S0₄ and the solvent was evaporated. The remaining residue was dissolved in HPLC eluent and analyzed on HPLC to determine the conversion.

Both the 2,3-DHBD from A. oryzae and the 4,5-DHPD from B. pertussis showed excellent decarboxylation activity (conv. >99%, Table 4) employing the corresponding natural substrates. In the case of the 4,5-DHPD from C. testosteroni neither the enzyme overexpressed in E. coli BL21 (DE3) nor the enzyme which was over- expressed together with a chaperone showed any significant activity. For the 4,5- DHPD from V. eiseniae the coexpression together with a chaperone could dramatically enhance the decarboxylation activity (conv. >99%, Table 4). Table 4: Results of the decarboxylation using the natural substrate enzymes

conversion - overexpressed organism host substrate

decarboxylation in E. coli

2,3-DHBD Aspergillus oryzae BL21 (DE3) 2,3-DHBA >99%

4,5-DHPD Bordetella pertussis BL21 (DE3) 4,5-DHPA >99%

Comamonas

4,5-DHPD BL21 (DE3) 4,5-DHPA <2%

testosteroni

Comamonas BL21 (DE3) +

4,5-DHPD 4,5-DHPA <2%

testosteroni chaperone

Verminephrobacter

4,5-DHPD BL21 (DE3) 4,5-DHPA >1 %

eiseniae

Verminephrobacter BL21 (DE3) +

4,5-DHPD 4,5-DHPA >99%

eiseniae chaperone

Beside the natural substrate a range of various nonnatural aromatic and hetero- aromatic substrates were applied for the decarboxylation reaction in order to gain more insight towards the substrate acceptance of the decarboxylases.

Phthalic Phthalic acid Phthalic acid

4-HBA 2.3- DHBA 3.4-DHBA Pyrrol acid

acid mono-ester di-ester

Regarding the decarboxylation activity applying a nonnatural substrate 2,3- DHBD from A. oryzae also accepted 4,5-DHPA (4,5-dihydroxyphthalic acid) as substrate and vice versa, 4,5-DHPD from B. pertussis (disrupted cell suspension) accepted 2,3-DHBA (2,3-dihydroxybenzoic acid) as substrate obtaining an excellent conversion (both >99%).

Table 5: Results of the decarboxylation using a nonnatural substrate

enzymes

conversion - overexpressed organism host substrate

decarboxylation in E. coli

2,3-DHBD Aspergillus oryzae BL21 (DE3) 4,5-DHPA >99%

4,5-DHPD Bordetella pertussis BL21 (DE3) 2,3-DHBA 7%

4,5-DHPD^'3 Bordetella pertussis BL21 (DE3) 2,3-DHBA >99% ^a cell suspension of disrupted cell was used

Example 4: Screening - carboxylation

Like for the decarboxylation screening first the natural substrate was used for the screening of the carboxylation activity. For the 2,3-DHBD from A. oryzae the natural substrate is catechol which is supposed to be carboxylated regioselective in o/ !0-position to the OH-group yielding the corresponding 2,3-dihydroxybenzoic acid. For the 4,5-DHPDs from B. pertussis, C. testosteroni and V, eiseniae the natural substrate is 3,4-dihydroxybenzoic acid obtaining 4,5-dihydroxyphthalic acid as expected product. In order to push the equilibrium towards carboxylation bicarbonate was added to the biotransformation which generated a certain C0₂ pressure in the tightly closed reaction vessel.

3,4-Dihydroxybenzoic 4,5-Dihydroxyphthaiic acid acid

Scheme 2: Carboxylation of catechol and 3,4-dihydroxybenzoic acid

General screening procedure of the carboxylation

All carboxylation reactions were performed in glass vials capped with septums.

Lyophilized whole cells (30 mg E. coli host cells containing the corresponding overexpressed enzyme or whole wild-type microbial cells) were resuspended in phosphate buffer (1 ml_ or 900 μΙ_, pH 5.5, 100 mM) and rehydrated for 30min. The substrate was added either directly or as a stock solution (1 0 - 1 00 μΙ_) to yield a final concentration of 1 0 mM, followed by addition of KHC0₃ (0.1 - 3M, 1 0 - 300 mg). In case of the jS-carboxylation all screenings were performed with and without the addition of organic cosolvent (20% v/v acetonitrile). For the experiments using 20% acetonitrile resuspension was performed with 800μΙ_ phosphate buffer to obtain a final volume of 1 ml of each biotransformation. After the addition of potassium bicarbonate the vials were immediately tightly closed and the mixture was shaken at 30 °C and 1 20 rpm. After 24h the reaction mixture was acidified with HCI (3M, 700 μΙ_) and the products were extracted with EtOAc (1 ml_). The organic layer was dried over Na₂S0₄ and the solvent was evaporated. The remaining residue was dissolved in HPLC eluent and analyzed on HPLC to determine the conversion. Alternative the reaction mixture was centrifuged (1 3000 rpm, 15 min), an aliquot of 1 00 iL of each sample was diluted with 1 ml_ of an H₂0/acetonitrile mixture (v/v = 50%) supplemented with trifluoroacetic acid (3% v/v, 30 μΙ_). After incubation at room temperature for 5min, the samples were again centrifuged (13000 rpm, 15 min) and analyzed on reverse-phase HPLC to determine the conversion. Products were identified by comparison with authentic reference material.

For the ocarboxylation applying olivetol (8b) and naphthol (10) the substrate stock solution was prepared in acetonitrile to overcome solubility problems, leading to a final acetonitrile concentration of 10% v/v in the reaction mixture. For the β- carboxylation applying substrate 14-17 and 19 the substrate stock solution was prepared in acetonitrile to overcome solubility problems, leading to a final acetonitrile concentration of 1 % v/v in the reaction mixture. The work up was performed as described above.

Control experiments

The absence of competing carboxylation activity of empty host cells (without piasmid encoding for the corresponding decarboxylase) was verified in separate blank experiments.

In order to determine the spontaneous (non-enzymatic) background reactions blank experiments were performed without the addition of the biocatalyst retaining all other reaction conditions. All conversion data (Table 2 and 3) are corrected for this value to accurately describe the enzyme-catalyzed activity.

Results

The 2,3-DHBD from A. oryzae showed a moderate carboxylation activity (conv. = 25%, Table 6) employing the natural substrate catechol. For the 4,5-DHPD from B. pertussis a carboxylation activity of 5% was detected towards the natural substrate 3,4-DHBA. In the case of the enzymes (4,5-DHPD) from C. testosteroni and V.

eiseniae neither the enzyme overexpressed in E. coli BL21 (DE3) nor the enzyme which was overexpressed together with a chaperone showed any significant activity.

Table 6: Results of the carboxylation using the natural substrate

enzymes

conversion - overexpressed organism host substrate

carboxylation in E. coli

2,3-DHBD Aspergillus oryzae BL21 (DE3) Catechol 25%

4,5-DHPD Bordetella pertussis BL21 (DE3) 3,4-DHBA 5%

4,5-DHPD Comamonas BL21 (DE3) 4,5-DHPA - testosteroni

Comamonas BL21 (DE3) +

4,5-DHPD 4,5-DHPA

testosteroni chaperone

Verminephrobacter

4,5-DHPD BL21 (DE3) 4,5-DHPA

eiseniae

Verminephrobacter BL21 (DE3) +

4,5-DHPD 4,5-DHPA

eiseniae chaperone

Beside the natural substrate a range of various nonnatural aromatic and hetero- aromatic substrates were applied for the carboxylation reaction. The expected position of the carboxylation is highlighted with an arrow, only in the case of 2,3-DHBA no data are available in literature.

Phenol Catechol Resorcin 2,3-DHBA 3,4-DHBA Pyrrol

The 2.3-DHBD accepted resorcin as a substrate yielding 2,6-dihydroxybenzoic acid as the expected product (conv. = 15%, Table 7). Catechol was converted in a low to moderate rate of the 4,5-DHPD from B. pertussis and C. testosteroni obtaining 2,3- DHBA as the expected products (conv. = 5% and 25%, Table 7). By applying phenol and pyrrol as substrate for none of the enzymes a carboxylation activity was detected. The carboxylation activity was also tested for 2,3- and 3,4-DHBA as substrate using all enzymes.

Table 7: Results of the carboxylation using a nonnatural substrate

enzymes

conversion - overexpressed organism host substrate

carboxylation in E. coli

2,3-DHBD Aspergillus oryzae BL21 (DE3) Resorcin 15%

4,5-DHPD Bordetella pertussis BL21 (DE3) Catechol 5%

Comamonas

4,5-DHPD BL21 (DE3) Catechol 25%

testosteroni Example 5: Conditions - carboxylation process

A range of different experiments were carried out in order to get more insight of the optimal reaction conditions. For this purpose the carboxylation reaction of catechol employing the 2,3-DHBD from A. oryzae was investigated in more details applying various reaction conditions. The results are summarized in table 5:

i) various concentrations of bicarbonate (0.1 - 3M) (entry 1 -3 and 5);

ii) increasing the reaction time from 24 h to 48 h (entry 4);

iii) addition of Ammonium acetate (NH₄Ac) and the reducing agent

Dithiothreitol (DTT) for both it was described in literature to enhance the enzymatic carboxylation activity (entry 6-9);

iv) applying various non-natural substrates using bicarbonate (1 M) as

carbon source (entry 10-12);

v) the use of bicarbonate and C0₂ gas as carbon dioxide source (0.1 -3M

bicarbonate and 2 bar C0₂-pressure) (entry 13-16);

vi) the use of only C0₂ gas as carbon dioxide source (2 bar C0₂-pressure) (entry 17);

vii) up scaling using 100 mM substrate concentration (entry 18);

viii) decarboxylation of the natural substrate as a control experiment (entry

19);

Table 8: Various reaction condition for the carboxylation reaction employing 2.3-

DHBD enzymes co₂- conversion

KHC0₃ NH₄AC/DTT

entry overexpressed substrate pressure carboxycomment

EM] [mM]

in E. coli [bar] lation

2,3-DHBD from A.

1 Catechol 0.1 - - <1 %

oryzae

2,3-DHBD from A.

2 Catechol 0.5 - - 8%

oryzae

2,3-DHBD from A.

3 Catechol 1 - - 26%

oryzae

2,3-DHBD from A. 48 h reaction

4 Catechol 1 - - 27%

oryzae time

2,3-DHBD from A.

5 Catechol 3 - - 32%

oryzae

2,3-DHBD from A.

6 Catechol 0.1 140/10 - - oryzae

7 2,3-DHBD from A. Catechol 0.5 140/10 - 5% enzymes co₂- conversion

KHC0₃ NH₄Ac/DTT

entry overexpressed substrate pressure carboxy- comment

[M] [m ]

in E. coli [bar] lation

oryzae

2,3-DHBD from A.

8 Catechol 1 140/10 - 21 %

oryzae

2,3-DHBD from A.

9 Catechol 3 140/10 - - oryzae

2,3-DHBD from A.

10 Phenol 1 - - side reaction

oryzae

2,3-DHBD from A.

1 1 Resorcin 1 - - 15%

oryzae

2,3-DHBD from A.

12 Pyrrol 1 - - - oryzae

2,3-DHBD from A.

13 Catechol 0.1 - 2 <1 %

oryzae

2,3-DHBD from A.

14 Catechol 0.5 - 2 12%

oryzae

2,3-DHBD from A.

15 Catechol 1 - 2 23%

oryzae

2,3-DHBD from A.

16 Catechol 3 - 2 18%

oryzae

2,3-DHBD from A.

17 Catechol - - 2 - oryzae

2,3-DHBD from A. l OOmM

18 Catechol 1 - - 33%

oryzae substrate

2,3-DHBD from A. decarboxylat

19 2,3-DHBA - - - >99%

oryzae ion

Increasing the amount of bicarbonate was going in hand with a continuous enhancement of the carboxylation activity.

Neither a longer reaction time nor the addition of NH₄Ac and DTT could

increase the conversion of the carboxylation reaction.

Regarding the source of carbon dioxide it turned out that C0₂ gas (2 bar pressure) was not applicable for the carboxylation employing the 2,3-DHBD from A.

oryzae. An alternative source of carbon dioxide could be the use of carbonic

anhydrase which is generating C0₂ from bicarbonate:

HCO₃ + H⁺ H₂C0₃→- C0₂ + H₂0 A much lower concentration of bicarbonate could be used since the high concentration of bicarbonate which is necessary for a good conversion makes the adjustment of the pH very difficult.

The standard substrate concentration used in these experiments was 10 mM, however by applying a 10fold increase of the substrate concentration (100 mM) an equally good conversion was obtained.

Example 6: Time study

For this time study the conversion of the carboxylation reaction of catechol employing the 2,3-DHBD from A. oryzae was determined after 1 . 3, 6, 10 and 24 hours. Within the first hour a conversion of 22% was reached however no further increase was obtained till a total reaction time of 24 h. Product inhibition or more likely the position of the equilibrium might cause the problem which needs to be further investigated in order to push the conversion to higher rates.

Example 7: Standard reaction conditions for carboxylation:

All carboxylation reactions were performed in glass vials capped with septums.

Lyophilized whole cells [30mg, either wild-type microorganisms (preparation B) or E. coli host cells containing the corresponding overexpressed enzyme (preparation A)] were resuspended in phosphate buffer (900 μΙ_, pH 5.5, 100 mM) and were re h yd rated for 30 min. The substrate was added as a stock solution (100 μΙ_) to yield a final con- centration of 10 mM, followed by addition of KHC0₃ (3M, 300 mg). After the addition of bicarbonate the vials were immediately tightly closed and the mixture was shaken at 30 °C and 120 rpm. After 24h the reaction mixture was centrifuged (13000 rpm, 15 min), an aliquot of 100 μΙ_ of each sample was diluted in 1 mL of an H₂0/acetonitrile mixture (v/v = 50%) and TFA (30 μΙ_) was added. After incubation at room temperature for 5 min, the samples were again centrifuged (13000 rpm, 15 min) and filtered before analyzed on reverse-phase HPLC to determine the conversion.

For biotransformations with 1 -naphthol, acetonitrile (10% v/v) was added to the phosphate buffer solution (100 mM, pH 5.5) to overcome solubility problems. Table 9: Carboxylation of nonnatural substrates

Substrates

Conversion [%] (arrows indicate observed

Enzyme Organism Preparation

carboxylation site)

Aspergillus

2.3DHBD A 30 51 ³ 43 58 ^b oryzae

2.6DHBD Rhizobium sp. A 35 31 ^d 0 0

Trlchosporon

SAD A 21 72 ^a 44 29 ^b moniliiforme

Lactobacillus

PAD A 0 0 0 0 plantarum

Bacillus

PAD A 0 1 ^d 0 18 ° amyloliquefaciens

NCIMB Mycobacterium

B 1 0 0 33 ° 10420 paraffinicum

Bacillus subtilis

DSM10 B n.t. n.t. n.t. 20 °

168

FCC073" Fusarium solani B n.t. n.t. n.t. 1 1 °

PrepaConversion [%] (arrows indicate observed

Enzyme Organism

ration carboxylation site)

2.3DHBD Aspergillus oryzae A 65 ^b 16 24 13 54 ,6DHBD Rhizobium sp. A 0 0 0 0 46

Trichosporon

SAD A 67 ^b 30 25 15 62 moniliiforme

Lactobacillus

PAD A 35 ° 0 0 0 0 plantarum

Bacillus

PAD A 35 ° 0 0 0 0 amyloliquefaciens

NCIMB Mycobacterium

B 12 ° 0 0 0 0 10420 paraffinicum

Bacillus subtilis

DSM 10 B 18 n.t. n.t. n.t. n.t.

168

FCC073^* Fusarium solani B 13 ° n.t. n.t. n.t. n.t. a two carboxylated products: 2,6-DHBA and 2,4-DHBA;

bring-carboxylation in o-position to the OH-group,

c carboxylation on the side chain, ^d pure 2,6-DHBA.

Preparation A: E. coli whole cells containing the overexpressed enzyme;

Preparation B: whole wild-type microbial cells;

n.t....not tested;

* FCC stands for your in-house Fab-Crew-Collection, the strain was obtained from G. Gubitz (TU Graz) in 2003.

Example 8: Pushing the Equilibrium via Addition of Organic Co-Solvents

In order to push the equilibrium towards the carboxylation reaction, the use of organic cosolvents was investigated (table 10). For this experiment, 2,6-DHBD from Rhizobium sp. was screened with its natural substrate (resorcinol) by applying various water-miscible and -immiscible organic solvents in various amounts (5 and 20% v/v) to the biotransformation reaction.

2,6-DHBD f rom f°^OH ic

Scheme 3: Regio-selective carboxylation of resorcinol to 2,6-dihydroxybenzoic acid

The addition of water-immiscible organic solvents had only minor effects on the conversion, whereas water-miscible co-solvents led to significant improvements in the conversion of the carboxylation. The best results were obtained with acetone (20%), acetonitrile (20%), NMP (A/-methyl-2-pyrrolidone, 20%) as well as DME (1 ,2- dimethoxyethane, 20%). With all four solvents the conversion was increased by almost 50% (relative, i.e. 47% compared to only 31 % in the absence of co-solvent). Table 10: Comparison of the conversion of resorcinol using 2,6-DHBD from R. sp. in different amounts of various organic solvents

Conversion [%]

Organic solvent

5 % 20 %

THF 35 i

DMSO 34 39

DMF 37 44

iso-propanol 34 30

ethylene glycol 33 34

DME 34 46 Conversion [%]

Organic solvent

5 % 20 %

MeOH 32 32

EtOH 34 41

iso-butanol 33 32

acetone 33 46

acetonitrile 34 46

NMP 38 47

glycerine 31 36

form amide 33 39

1 ,4-dioxane 34 45

toluene 31 32

n-heptane 33 34

dichlormethane 35 33

tert-butylmethyl ether 38 25

no organic solvent 32

To test the general viability of organic co-solvents their influence on the carboxylation of catechol and 2-methoxy-4-vinyl phenol employing salicylic acid decarboxylase from Trichosporon moniliiforme (SAD Tm) and phenolic acid decarboxylase from Bacillus amyloliquefaciens (PAD Ba), respectively, was investigated (Table 1 1 -13). For this study only solvents showing a significant enhancement of the enzymatic carboxylation activity during the biotransformation of resorcinol were selected.

Furthermore, additional solvents were tested (amines, PFC, etc.) and concentration studies with the best organic solvents were performed (Table 1 1 , 12 and 13).

Table 11 : Influence of various organic solvents (20% v) on the enzymatic carboxylation of resorcinol, catechol and 2-methoxy-4-vinyl phenol employing 2,6- DHBD, SAD and PAD Ba, respectively, (n.t. = not tested)

Conversion [%]

2,6-DHBD SAD PAD

Trichosporon Bacillus

Organic solvent Rhizobium sp.

me amyloliquefaciens

acetone 46 38 27 acetonitrile 46 46 20 Conversion [%]

2,6-DHBD SAD PAD

Trichosporon Bacillus

Organic solvent Rhizobium sp.

ens

/V-methyl-2-pyrrolidone (NMP) 47 36 2

/V-methylmorpholine 35 1 39

1 ,2-dimethoxyethane (DME) 46 40 28

DMF 44 35 26

1 ,4-dioxane 45 36 26

EtOH 41 38 30 methyl-isobutylketone n.t. 25 27

toluene 32 30 37 n-hexane 34 30 39 dichloromethane 33 35 50 ferf-butyl-methyl ether 26 29 44

n-octane n.t. 29 33

/V-methyl diethanolamine 20 <1 57

AMP <1 <1 34

2-(isopropylamino)ethanol <1 <1 5

Perfluorodecalin (PFC) 33 30 35

PFC + TWEEN80 33 29 32 no organic solvent 31 26 35

Table 12: Solvent concentration study employing SAD to the carboxylation of catechol

Organic co-solvent

Enzyme Substrate

A _{rv £} 10 33 35 33 43 SAD from

F 20 37 46 38 40 Trichosporon f "

30 32 30 37 7 moniliiforme

40 20 19 24 1

50 10 9 13 0 Table 13: Solvent concentration study employing PAD Ba to the carboxylation of 2-MeO-4-vinylphenol

Organic solvent

Enzyme Substrate N-methyl-

% v/v toluene n-hexane diethanolamine

0 35 35 35

5 43 36 37

40 53 26 32

50 51 21 30

Example 9: Organic cosolvent study

The carboxylation of resorcinol (natural substrate) applying 2,6- dihydroxy- benzoic acid decarboxylase from Rhizobium sp. (2,6-DHBD Rsp) as biocatalyst using various organic cosolvents was investigated under standard screening condition only differing in the addition of 5% and 20% (v/v) organic solvent, respectively. The applied organic cosolvents were categorized into four groups: i) water-miscible solvents, ii) water-immiscible solvents, iii) CO₂ capturing solvents and iv) PEG^'S (polyethylene glycols) (Fig. 1 and Fig. 7).

A significant increase in conversion to the corresponding carboxylated product 2,6-dihydroxybenzoic acid was obtained applying water-miscible organic solvent or PEG^'S at a ratio of 20% (v/v), whereas for the lower ratio (5% v/v) of organic solvent the obtained conversion was only minor effected in comparison to the biotrans- formation without any addition of cosolvent. The addition of water-immiscible organic solvents and a number of solvents which were supposed to capture CO₂ had no influence on the conversion or resulted even in an inactivation of the biocatalyst. The results obtained for the organic cosolvent study were summarized in Fig. 7 and table 14 showing the conversion at a ratio of 5% and 20% (v/v) organic cosolvent,

accordingly in comparison to blank reaction for which no cosolvent was added. Overall, the presence of organic cosolvents pushed the conversion by ca. 50%. Table 14: Carboxylation of resorcinol using 2,6-DHBD Rsp applying various organic cosolvents at different amounts.

conversion [%]

organic cosolvent 5% v/v 20% v/v

THF 35 1

DMSO 34 39

DMF 37 44

/-propanol 34 30

ethylene glycol 33 34

1 ,2-dimethoxyethane (DME) 34 46

MeOH 32 32

EtOH 34 41

i-butanol 33 32

acetone 33 46

acetonitrile 34 46

A/-methyl-2-pyrrolidone (NMP) 38 47

glycerine 31 36

formamide 33 39

1 ,4-dioxane 34 45

toluene 31 32

n- heptane 33 34

dichlorormethane 35 33

f-butylmethyl ether 38 25

V-me†hyl diethanol amine 25 19

2-(isopropylamino)-ethanol 23 < 1

2-amino-2-methyl-propanol 23 < 1

/V-methylmorpholine 30 35

Perfluorodecalin (RFC) 31 33

PFC+Tween80 31 33

Polyethylene glycol-200 34 42

Polyethylene glycol-300 35 45

Polyethylene glycol-400 36 46

no organic solvent 32

The dependence of the concentration of the organic cosolvent towards the effect of shifting the equilibrium has been studied in more detail (Fig. 2). For this purpose SAD Tm as biocatalyst and catechol as substrate was used applying various water-miscible organic solvents which were found to push the conversion of the carboxylation product (see above). Various concentrations of the cosolvent were used ranging from 0% until 50% (v/v). The optimum of enhancement was obtained within a range of 15% to 25% of cosolvent, confirming that a ratio of 20% (v/v) cosolvent which was applied in the standard screening procedure was the optimum.

Example 10: Bicarbonate source study

In order to elucidate the correlation between the carboxylation and the source of bicarbonate, a set of experiments using various bicarbonate sources (3M) differing in the type of the cation were performed applying resorcinol as substrate and 2,6- DHBD Rsp as biocatalyst (Fig. 3). For standard conditions KHC0₃ (3M) was used as bicarbonate source obtaining 36% conversion of the expected carboxylated product 2,6-dihydroxybenzoic acid. The use of a carbonate-based ionic liquid (1 -butyi-3- methylimidazolium hydrogen carbonate solution -50% in methanol/water 2/3; 200 μΙ_) in addition to bicarbonate (3 , KHC0₃) resulted in a strong increase of conversion (conv. 46%) whereas the use of the carbonate-based ionic liquid alone (without addition of bicarbonate) led to a significant decrease of conversion (conv. 25%). The use of NH₄HCO₃ displayed similar results concerning conversion (conv. 33%) as under standard conditions. In case of Na⁺ and Cs⁺ bicarbonate source, a conversion of 25% was obtained which is within the same range as the carbonate-based ionic liquid without additional KHCO3.

Example 11 : Substrate concentration study

In order to examine the potential of the enzymatic carboxylation reaction a substrate concentration study was performed especially in view of a potential industrial applicability employing resorcinol as substrate and 2,6-DHBD Rsp as biocatalyst (Fig. 4). The conversion of resorcinol to the corresponding 2,6-dihydroxybenzoic acid stayed almost constant within a range of 5 mM until 100 mM with only a slightly continuous decrease of conversion from 36% to 31 % when the substrate concentration was increased from 5 mM to 100 mM, indicating the potential for industrial use. A significant decrease of conversion was observed applying 200 mM of resorcinol which led to a conversion of 1 1 % to the corresponding product. For even higher substrate

concentrations (300 mM to 500 mM) no conversion was obtained. Example 12: Temperature study

In order to examine the correlation between the carboxylation and reaction temperature, a set of experiments using temperatures ranging from 10°C to 60 °C were performed applying resorcinol as test substrate and 2,6-DHBD Rsp as biocatalyst (Fig. 5). The carboxylation exhibit a very broad operational window concerning temperature with an almost constant conversion of resorcinol within a range of 10°C to 50 °C showing an optimum at 30 °C. Beyond 50 °C the activity of the biocatalyst dropped rapidly resulting in complete inactivation at 60°C.

Example 13: pH study

In order to elucidate the correlation between the carboxylation and the pH, a set of experiments using various pH values ranging from 6.5 to 9.5 were performed applying resorcinol as test substrate and 2,6-DHBD Rsp as biocatalyst (Fig. 6). The optimum pH for the carboxylation lies between a range of 7.5 and 8.5 whereas a more acidic buffer system (pH 6.5) as well as more basic conditions (pH 9.5) led to a decrease in conversion. The lower final conversion in comparison to the other engineering studies (temperature, organic cosolvent, substrate concentration, bicarbonate source) is due to less bicarbonate used for this experiments in order to enable the adjustment of the pH to lower values (7.5 and 6.5). References:

(1 ) (a) P. G. Jessop, T. Ikariya, R. Noyori, Chem. Rev. 1999, 99, 475-93; (b) M.

Yoshida, N. Hara, S. Okuyama, Chem. Commun. 2000, 151 -2; (c) T. Sakakura, J.-C.

Choi, Y Saito, T. Masuda, T. Sako, T. Oriyama, J. Org. Chem. 1999, 64, 4506-8; (d) H.

Kawanami, Y. Ikushima, Chem. Commun. 2000, 2089-90; (e) R. J. Sowden, M. F. Sellin, N. D. Blasio, D. J. Cole-Hamilton, Chem. Commun. 1999, 251 1 -12; (f) T.

Mizuno, N. Okamoto, T. Ito, T. Miyata, Tetrahedron Lett. 2000, 41 , 1051 -53; (g) M.

Kunert, M. Brauer, O. Klobes, H. Gorls, E. Dinjus, Eur. J. Inorg. Chem. 2000, 1803-9;

(h) Y. Wada, T. Kitamura, S. Yanagida, Res. Chem. Intermed. 2000, 26, 153-9; (i) M.

Tokuda, T. Kabuki, Y. Kato. H. Suginome, Tetrahedron Lett. 1995, 36, 3345-48.

(2) (a) R. J. De Pasquale, J. Chem. Soc. Chem. Commun. 1973, 1 157-8. (b) S. Inoue, in: Carbon Dioxide as a Source of Carbon, M. Aresta, G. Forti (eds.), Elsevier,

Amsterdam, 1987, pp. 206-331 ; (c) D. J. Darensbourg, M. W. Holtcamp, Coord. Chem.

Rev. 1996, 153, 155-74. (3) M. Aresta, E. Quaranta, I. Tommasi, New J. Chem. 1994, 18, 133-42.

(4) F. C. Hartman, M. R. Harpel, Annu. Rev. Biochem. 1994, 63, 197-234.

(5) M. Hugler, H. Huber, K. O. Stetter and G. Fuchs. Arch. Microbiol. 2003, 179, 160- 173; R. K. Thauer, Science 2007, 318, 1732-1733.

(6) M. Calvin, Nature 1961 , 192, 799; F. C. Hartman and M. R. Harpel, Annu. Rev. Biochem. 1994, 63, 197-234.

(7) M. C. Evans, B. B. Buchanan and D. I. Arnon, Proc. Natl. Acad. Sci. USA 1966, 55, 928-934.

(8) S. W. Ragsdale, Ann. N. Y. Acad. Sci. 2008, 1 125, 129-136; H. L. Drake, A. S. GoBner and S. L. Daniel, Ann. N. Y. Acad. Sci. 2008, 1 125, 100-128; S. W. Ragsdale and E. Pierce, Biochim. Biophys. Acta 2008, 1784, 1873-1898.

(9) S. Herter, G. Fuchs, A. Bacher and W. Eisenreich, J. Biol. Chem. 2002, 277, 20277-20283; B. Alber, M. Olinger, A. Rieder. D. Kockelkorn. B. Jobst, M. Hugler and G. Fuchs, J. Bacteriol. 2006, 188, 8551 -8559.

(10) I. A. Berg, D. Kockelkorn, W. Buckel and G. Fuchs, Science 2007, 318, 1782- 1786.

(1 1 ) H. Huber, M. Gallenberger, U. Jahn, E. Eylert, I. A. Berg, D. Kockelkorn, W.

Eisenreich and G. Fuchs, Proc. Natl. Acad. Sci. USA 2008, 105, 7851 -7856.

(12) T. J. Erb, I. A. Berg, V. Brecht, M. Muller, G. Fuchs and B. E. Alber, Proc. Natl. Acad. Sci. USA 2007, 104, 10631 -10636.

(13) S. Kuwabata, R. Tsuda, K. Nishida, H. Yoneyama, Chem. Lett. 1993, 1631 -4.

(14) K. Sugimura, S. Kuwabata, H. Yoneyama, J. Am. Chem. Soc. 1989, 1 1 1 . 2361 -2.

(15) Y. F. Cheung. C. H. Fung, C. Walsh, Biochemistry 1975, 14, 2981 -6.

(16) T. Yoshida and T. Nagasawa, J. Biosci. Bioeng. 2000, 89, 1 1 1 -1 18.

(17) T. Matsuda, T. Harada and K. Nakamura, Green Chem. 2004, 6, 440-444.

(18) A. Liese and M. V. Filho, Curr. Opinin Biotechnol. 1999, 10, 595-603.

(19) This activity seems to be widespread among the microbial community and was reported from Aspergillus niger, A. oryzae, Pseudomonas testosteroni, Klebsiella pneumoniae, Bacillus megaterium, Serratia spp., and even sulfur-metabolisers: see refs. 1 12, 1 19, 121 and: J. C. Patel and D. J. W. Grant, Antonie van Leeuvenhoek 1969, 35, 53-64.

(20) A. Dibenedetto, R. Lo Noce, C. Pastore, M. Aresta and C. Fragale, Environ.

Chem. Lett. 2006, 3, 145-148.

Claims

Claims:

1 . A process for the regioselective carboxylation of a nonnatural substrate compound using a decarboxylase.

2. The process according to claim 1 , wherein the carboxylation is an o- or β- carboxylation.

3. The process according to claim 1 or 2, wherein at least 25% of the nonnatural substrate are converted, preferably at least 30%.

4. The process according to any one of claim 1 to 3, wherein the

decarboxylase is selected from the group consisting of dihydroxybenzoic acid decarboxylase (DHBD), dihydroxyphthaiic acid decarboxylase (DHPD), salicylic acid decarboxylase (SAD), pyruvate decarboxylase (PDC), phenolic acid decarboxylase (PAD), and ferulic acid decarboxylase (FDC).

5. The process according to any one of claim 1 to 4, wherein the nonnatural substrate is compound of formula I,

wherein

A is a mono- or bicyclic aromatic or heteroaromatic ring; and

each R¹ is independently from one another selected from the group consisting of halogen, carboxy, -(CH₂)_nOH, -NH₂, Ci.₃alkoxy, d.₃alkyl, Ci.₃alkenyl, and

n is 0, 1 , 2, 3 or 4.

6. The process according to claim 5, wherein the nonnatural substrate is hydroxystyrene.

7. The process according to claim 6, wherein the side chain R¹ = -CH=CH₂ of said hydroxystyrene is carboxylated.

8. A process for the regioselective carboxylation of natural and nonnatural substrate compounds using a decarboxylase, wherein an organic cosoivent is added to enhance the solubility of the substrate and/or product(s) and to increase the

conversion.

9. The process according to claim 8, wherein the organic cosoivent is selected from the group consisting of DMSO, DMF, EtOH, acetone, acetonitrile,

A/-methyl-2-pyrrolidone, glycerine, formamide, and 1 ,4-dioxane.

10. A process for manufacturing a carboxylated compound, comprising the steps of:

i) resuspending host cells which overexpress the decarboxylase in a buffer,

ii) adding the non-natural substrate compound;

iii) adding a carbon source; and

iv) obtaining the carboxylated compound.

1 1 . The process according to claim 10, wherein said carbon source is CO₂ or a carbonate buffer.

12. A compound obtained in a process according to any one of claims 1 to 1 1 for use in pharmaceutical, cosmetic and/or food industry.