Metabolic Engineering of Escherichia coli for l-Tyrosine Production by Expression of Genes Coding for the Chorismate Mutase Domain of the Native Chorismate Mutase-Prephenate Dehydratase and a Cyclohexadienyl Dehydrogenase from Zymomonas mobilis

Abstract

The expression of the feedback inhibition-insensitive enzyme cyclohexadienyl dehydrogenase (TyrC) from Zymomonas mobilis and the chorismate mutase domain from native chorismate mutase-prephenate dehydratase (PheA_CM) from Escherichia coli was compared to the expression of native feedback inhibition-sensitive chorismate mutase-prephenate dehydrogenase (CM-TyrA_p) with regard to the capacity to produce l-tyrosine in E. coli strains modified to increase the carbon flow to chorismate. Shake flask experiments showed that TyrC increased the yield of l-tyrosine from glucose (Y_l-Tyr/Glc) by 6.8-fold compared to the yield obtained with CM-TyrA_p. In bioreactor experiments, a strain expressing both TyrC and PheA_CM produced 3 g/liter of l-tyrosine with a Y_l-Tyr/Glc of 66 mg/g. These values are 46 and 48% higher than the values for a strain expressing only TyrC. The results show that the feedback inhibition-insensitive enzymes can be employed for strain development as part of a metabolic engineering strategy for l-tyrosine production.

l-Tyrosine (l-Tyr) is an aromatic amino acid with several applications. It is used as precursor in the synthesis of some drugs (39), biodegradable polymers (10), melanin (12), and phenylpropanoids (38). The pathway leading to l-Tyr biosynthesis starts with the condensation of phosphoenolpyruvate (PEP) and erythrose-4-phosphate (E4P) in a reaction catalyzed by 3-deoxy-d-arabino-heptulosonate-7-phosphate synthase (DAHPS). Seven reactions are necessary to produce chorismate (CHA), the common precursor for the synthesis of the three aromatic amino acids (Fig. 1). For l-phenylalanine (l-Phe) and l-Tyr biosynthesis, CHA is converted to prephenate (PPA) in a reaction catalyzed by chorismate mutase (CM). In nature, there are two distinct pathways that synthesize l-Tyr. In enteric bacteria, prephenate dehydrogenase (TyrA_p) catalyzes the conversion of PPA to p-hydroxyphenylpyruvate (HPP). l-Tyr is formed by transamination of HPP. In plants, PPA is converted to l-arogenate (l-AGN) by transamination. Conversion of l-AGN into l-Tyr is catalyzed by arogenate dehydrogenase (TyrA_a). Zymomonas mobilis and other organisms possess enzymes called cyclohexadienyl dehydrogenases (TyrA_c), which are capable of using either PPA or l-AGN as a substrate for l-Tyr biosynthesis (47). TyrA_a, TyrA_p, and TyrA_c belong to the TyrA protein family, which includes the group containing all known dehydrogenases involved in l-Tyr biosynthesis, and use NAD⁺ or NADP⁺ or both dinucleotides as cofactors. Members of the TyrA protein family are frequently fused to other catalytic or regulatory domains. Furthermore, compounds like l-Tyr and HPP inhibit the activity of some TyrA proteins (8, 9, 40, 44). In Escherichia coli, l-Tyr inhibits the activity of three enzymes: the AroF DAHPS isozyme and 3-dehydroquinate synthase in the common aromatic pathway and the bifunctional enzyme chorismate mutase-prephenate dehydrogenase (CM-TyrA_p), in the l-Tyr biosynthetic pathway (4, 35) (Fig. 1).

FIG. 1.

l-Tyr biosynthetic pathways. The common aromatic pathway begins with the reaction between PEP and E4P, which yields 3-deoxy-d-arabino-heptulosonate-7-phosphate (DAHP). CHA is the branch point of the pathway. PPA is formed from CHA in a reaction catalyzed by CM. l-AGN is produced from PPA by the action of an aminotransferase (AT). l-AGN is the substrate of TyrA_a or TyrA_c. HPP is formed from PPA by the activity of either TyrA_p or TyrA_c, and HPP is transaminated by the aminotransferase. In l-Tyr biosynthetic pathways, 1 mol of NAD⁺ or NADP⁺ is used. The dashed lines indicate two or more reactions.

The general strategy used for the development of aromatic amino acid-overproducing strains has involved the alleviation of control mechanisms in key pathways (7, 41). Several different microorganisms have been modified for l-Tyr production. Corynebacterium glutamicum, Arthrobacter globiformis, and Brevibacterium lactofermentum l-Tyr-overproducing strains were developed by classical mutagenesis methods, and they were selected based on their capacity to grow on toxic aromatic amino acid analogs (20, 25, 36). Biochemical analyses of C. glutamicum strains showed that enzymes like DAHPS and CM were not inhibited by the aromatic amino acids (Fig. 1), while the TyrA_a of this organism is naturally insensitive to feedback inhibition by l-Tyr (18, 21). The best C. glutamicum strain obtained by this procedure produced 26 g/liter of l-Tyr in 80 h in a fed-batch fermentation (24). Metabolic engineering and protein-directed evolution strategies have been used to construct E. coli l-Tyr-producing strains (31, 32; A. Takai, R. Nishi, Y. Joe, and H. Ito, U.S. patent application 20050277179). In these studies, feedback inhibition-resistant variants of CM-TyrA_p were obtained by error-prone PCR of the encoding gene tyrA and were selected using the toxic analog 3-fluoro-dl-tyrosine. Using E. coli as a host strain, Takai et al. (U.S. patent application 20050277179) and Lütke-Eversloh and Stephanopoulos (32) constructed E. coli l-Tyr producers by overexpression of feedback inhibition-resistant variants of DAHPS and CM-TyrA_p (13, 29, 31, 32; Takai et al., U.S. patent application 20050277179). One of the E. coli l-Tyr producers, which had additional central metabolism genetic modifications, produced 9.7 g/liter of l-Tyr, with an l-Tyr yield on glucose (Y_l-Tyr/Glc) of 0.10 g/g and an l-Tyr-specific production rate (q_l-Tyr) of 73 mg/g (dry weight)/h in a fed-batch fermentation (32). Olson et al. (33) employed a different approach for construction of an E. coli l-Tyr producer. An l-Phe-producing strain, which was developed by a classical mutagenesis procedure, was converted into an l-Tyr-producing strain by replacing the native promoter of tyrA with the trc promoter. This strain produced 55 g/liter of l-Tyr, with a Y_l-Tyr/Glc of 0.3 g/g and a q_l-Tyr of around 57 mg/g (dry weight)/h (33, 34).

The cyclohexadienyl dehydrogenase (TyrC) of Z. mobilis is not inhibited by l-Tyr or the other two aromatic amino acids (47). Due to this property, this enzyme was used recently for the development of a tunable selection system for mutant prephenate dehydratases. In this system, TyrC diverted PPA into the l-Tyr pathway to avoid the nonenzymatic conversion of PPA into phenylpyruvate in l-Phe biosynthesis (26). On the other hand, CM-prephenate dehydratase (P-protein) from E. coli is involved in l-Phe biosynthesis. When the CM domain (residues 1 to 109) of the P-protein is expressed alone (PheA_CM), it retains catalytic activity and becomes insensitive to l-Phe inhibition (45, 46). In contrast, the CM domain from CM-TyrA_p loses its catalytic activity when it is not fused to the corresponding TyrA_p domain (14).

Considering the information described above, an approach not dependent on random mutagenesis for generating a CHA-to-HPP pathway insensitive to feedback inhibition by l-Tyr was evaluated in this work. The strategy reported here was based on expressing genes encoding enzyme domains expected to be insensitive to feedback inhibition by l-Tyr. The tyrC gene from Z. mobilis, which codes for TyrC, was expressed in an E. coli strain previously modified to increase carbon flow to CHA. In addition, to reconstitute the complete CHA-to-HPP pathway, an operon was constructed with the tyrC gene and the region of the pheA gene (pheA_CM) that codes for the PheA_CM domain.

Details about the construction of plasmids pTrctyrC, pTrctyrCpheA_CM, and pTrctyrCpheA_CMINV are shown in Tables 1 and 2 and Fig. 2. In several experiments, a strain expressing the tyrA gene was used as a control. tyrA was cloned from E. coli JM101 chromosomal DNA (6) into pCR-Blunt II-TOPO to generate plasmid pTOPOtyrA. The tyrA gene was isolated from this plasmid and cloned into plasmid pTrc99A to construct pTrctyrA. The nucleotide sequences of the three genes were determined from plasmids and were found to be identical to those reported previously (Table 2).

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Relevant features	Reference or source
Z. mobilis ATCC 31821	Wild-type strain	Laboratory stock
E. coli strains
FA114	FA113 ΔtyrA ΔpheA	5
JM101	supE thi Δ(lac-proAB) F′	6
PB12	JM101 Δ(ptsHI crr) Glc+	19
PB12G	PB12 transformed with pJLBaroGfbr.	This study
PB12A	PB12 transformed with pJLBaroGfbr and pTrctyrA	This study
PB12C	PB12 transformed with pJLBaroGfbr.and pTrctyrC	This study
PB12CP	PB12 transformed with pJLBaroGfbr.and pTrctyrCpheACM	This study
Plasmids
pTrc99A	Cloning vector, carries bla and lacIq genes and trc promoter	1
pCR-BluntII-TOPO	Cloning vector, carries kan gene	Invitrogen (Carlsbad, CA)
pTOPOtyrA	tyrA gene cloned in pCR-BluntII-TOPO	This study
pTrctyrA	tyrA gene under control of the trc promoter	This study
pTOPOtyrC	tyrC gene cloned in pCR-Blunt II-TOPO	This study
pTrctyrC	tyrC gene under control of the trc promoter	This study
pTrcpheAfbra	pheAfbr gene codes for CM and prephenate dehydratase domains of P protein under control of the trc promoter	3
pTrctyrCpheACM	tyrC and pheACM under control of the trc promoter; pheACM codes for the CM domain of P-protein	This study
pTrctyrCpheACMINV	Same as pTrctyrCpheACM but pheACM is not in the correct orientation for its expression	This study
pJLBaroGfbr	aroGfbr under control of the lacUV5 promoter; carries lacIq and tet genes	3

^aThe superscript fbr indicates feedback inhibition resistance.

TABLE 2.

Sequences of the primers used in this study

Primer	Nucleotide sequencea
Cloning of tyrA
tyrA3	5′ GGCTTAAGAGGTTTACCATGGTTGCTGAATTG 3′
tyrA5	5′ CCCCAAGCTTGATGAAAAGGTGCCGGATGATGTG 3′
Cloning of tyrC
tyrC5b	5′ GCAGGCGCTCTCCATGGCCGTCTTTAAG 3′
tyrC3	5′ CCCCAAGCTTGTTCATGCTGCGATCAATCG 3′
Cloning of pheACM
CM5	5′ CCCCAAGCTTCACACAGGAAACAGACCATGG 3′
CM3c	5′ CCCCAAGCTTTTATCAGAGAAAAGCGATGCGTGC 3′
Sequencing
M13 reversed	5′ CAGGAAACAGCTATGAC 3′
M13 forwardd	5′ GTAAAACGACGGCCAG 3′
tyrA1500	5′ CTGTTTTATCAGACCGCTTCTGC 3′
SectyrA5	5′ GTGTGCCAATCCACGTTACTG 3′
pTrc2	5′ CAGCTTATCATCGACTGCAC 3′

^aIn the nucleotide sequences key restriction enzyme sites are underlined.

^bTo clone the tyrC gene in pTrc99A, the GTG start codon in the wild-type gene was changed to ATG in order to create an NcoI site.

^cStop codons are indicated by bold type.

^dObtained from Invitrogen (Carlsbad, CA).

FIG. 2.

Construction of plasmids pTrctyrC and pTrctyrCpheA_CM. The tyrC gene was amplified using PCR from chromosomal DNA of Z. mobilis ATCC 3182 and cloned into plasmid pCR-Blunt II-TOPO (Invitrogen, Carlsbad, CA) to generate plasmid pTOPOtyrC. Then the tyrC gene was obtained by digestion of pTOPOtyrC and cloned into pTrc99A (1) to generate plasmid pTrctyrC. To construct the pTrctyrCpheA_CM operon, the CM domain from P-protein was amplified by PCR using plasmid pTrcpheA^fbr (3) as the template and cloned into pTrctyrC to generate plasmid pTrctyrCpheA_CM (Table 2). In plasmid pTrctyrCpheA_CMINV, the pheA_CM gene was inserted in the incorrect orientation for its expression. PDT, prephenate dehydratase domain; R, regulatory domain; RBS, ribosome binding site.

Functional complementation assays were performed with E. coli FA114 (Table 1), which lacks CM, TyrA_p, and prephenate dehydratase activities. FA114 was transformed with each of the following plasmids: pTrc99A, pTrctyrA, pTrctyrCpheA_CM, and pTrctyrCpheA_CMINV (Table 1). Assays were performed at 30°C in liquid M9 minimal medium (37) supplemented with (per liter) 2 g glucose, 20 mg l-Phe, 20 mg l-tryptophan, 20 mg l-leucine, 20 mg l-isoleucine, 10 mg p-aminobenzoic acid, 10 mg p-hydroxybenzoic acid, 10 mg 2,3-dihydroxybenzoic acid, 200 μg ampicillin, and 23.83 mg isopropyl-β-d-thiogalactopyranoside (IPTG). Only the cells transformed with plasmid pTrctyrA or pTrctyrCpheA_CM were able to grow in minimal medium (data not shown). This experiment indicated that the genes were expressed and the enzyme activity levels of TyrC and PheA_CM were sufficient to give FA114 the metabolic ability to synthesize l-Tyr.

In order to determine if l-Tyr inhibited TyrC and PheA_CM activities, enzymatic assays were performed in the presence or absence of l-Tyr. Strain FA114/pTrctyrCpheA_CM was grown in LB medium with 23.83 mg/liter of IPTG and 100 mg/liter of carbenicillin. Cultures were started at an optical density at 600 nm (OD₆₀₀) of 0.03 and incubated at 30°C for 12 h. Cells were washed once with lysis buffer (0.1 M Tris, 1 mM sodium EDTA, 1 mM dithiothreitol, 50 mM KCl; pH 7.5) and disrupted by sonication. PheA_CM activity was assayed at 30°C by following the consumption of CHA at 290 nm (ɛ₂₉₀ for CHA, 1.85 × 10³ M⁻¹ cm⁻¹). TyrC activity was measured at 30°C by monitoring the appearance of NADH at 340 nm (ɛ₃₄₀ for NADH plus HPP, 6.4 × 10³ M⁻¹ cm⁻¹). The reaction mixtures contained 50 mM Tris, 0.5 mM sodium EDTA, and 0.1 mg/ml bovine serum albumin (pH 7.5) at 30°C. The following concentrations of substrates were used: 2 mM NAD⁺, 0.2 mM PPA (for TyrC activity), and 1 mM CHA (for PheA_CM activity) (15, 27). Protein concentrations were determined by the Bradford method (11). One unit of TyrC activity was defined as the amount of enzyme that consumed 1 μmol of CHA per min at 30°C, and 1 U of PheA_CM activity was defined as the amount of enzyme that produced 1 μmol of NADH per min at 30°C. TyrC had a specific activity of 1.54 ± 0.08 IU/mg protein, while in the presence of 0.6 mM l-Tyr the specific activity was 1.34 ± 0.47 IU/mg protein. These results show that TyrC is not sensitive to l-Tyr inhibition. This is in agreement with the results obtained by Zhao et al. (47), who determined that 2 mM l-Tyr did not inhibit the activity of TyrC. Previous reports have determined that the activity of PheA_CM is not inhibited by l-Phe (46). However, l-Tyr was not assayed as an inhibitor, because it does not inhibit P-protein activity. To determine the possible sensitivity of the PheA_CM domain to l-Tyr, assays were carried out as described above. The specific activity of PheA_CM was 29.4 ± 5.39 IU/mg protein without l-Tyr. The specific activities with 0.6 and 1.1 mM l-Tyr were 34.4 ± 3.09 and 21.0 ± 0.23 IU/mg protein, respectively. In the case of CM-TyrA_p, it has been reported that at concentrations higher than 0.5 mM, l-Tyr inhibits TyrA_p activity by 90% and CM activity by 45% (23, 43). Our results showed that both TyrC and PheA_CM are less sensitive to l-Tyr inhibition than corresponding activities of CM-TyrA_p.

The effect of expressing the aroG^fbr, tyrA, tyrC, and pheA_CM genes on l-Tyr production was evaluated in shake flasks cultures. The aroG^fbr gene codes for a feedback inhibition-resistant DAHPS mutant; its expression causes an increase in the carbon flux directed into the common aromatic amino acid pathway (3). These experiments were carried out with E. coli strain PB12, which lacks the phosphotransferase system operon genes; therefore, it does not use PEP to transport and phosphorylate glucose (19). In this way, PEP availability for aromatic biosynthesis is increased (2). PB12 was transformed with plasmid pJLBaroG^fbr to generate strain PB12G. Strain PB12G was then transformed with plasmid pTrctyrA (PB12A), pTrctyrC (PB12C), or pTrctyrCpheA_CM (PB12CP) (Table 1). Shake flask cultures were grown at 30°C in M9 liquid medium supplemented with (per liter) 10 g glucose, 23.83 mg IPTG, 30 μg tetracycline, and 100 μg carbenicillin. Growth was monitored by measuring the OD₆₀₀, and the results were converted to dry weight of cells by assuming that 1 OD₆₀₀ unit was equivalent to 0.37 g/liter (22). l-Tyr and l-Phe in culture supernatants were quantified using an Agilent 1100 high-performance liquid chromatography system (Agilent Technologies, Palo Alto, CA) equipped with a Phenomenex Synergy Hydro RP18 column (150 by 4.6 mm; 4 μm) attached to an Agilent 1100 electrospray mass spectrometry detection system. Samples were eluted with 10% methanol in 0.1% acetic acid in water at an isocratic flow rate of 0.5 ml/min. UV quantitation was performed at 220 nm. The mass spectrometry conditions for amino acid determination were as follows: electrospray ionization in the positive mode; fragmentor voltage, 90 V; drying gas temperature, 300°C; drying gas flow rate, 13 liters/min; capillary voltage, 4,000 V; and nebulizer pressure, 30 lb/in². For total scans, (M+1)⁺ ions were detected and identified for standards and also for samples. The m/z ratios for l-Phe and l-Tyr are 166 and 182, respectively. The glucose content was determined from culture supernatants using an enzymatic analyzer (2700 biochemistry analyzer; YSI, Yellow Springs, OH.).

Compared with PB12G, PB12A had a q_l-Tyr and a Y_l-Tyr/Glc that were 1.6- and 1.1-fold higher, respectively. On the other hand, the q_l-Tyr for PB12C and PB12CP were 6.8- and 5.8-fold higher, respectively, than the values for PB12A (Fig. 3 and Table 3). Compared with PB12A, PB12C and PB12CP had Y_l-Tyr/Glc that were 6.6- and 6.4-fold higher, respectively. These results showed that overexpression of wild-type tyrA (CM-TyrA_p) resulted in only a small increase in the l-Tyr production capacity in a strain already expressing a feedback inhibition-resistant version of DAHPS. On the other hand, expression of tyrC resulted in a significant increase in l-Tyr production, and no further enhancement was observed when both tyrC and pheA_CM were expressed. These results showed that expression of a feedback inhibition-insensitive variant of DAHPS is sufficient to cause E. coli to overproduce l-Tyr. This capacity was not increased considerably by overexpression of CM-TyrA_p, most likely because it is subject to feedback inhibition by l-Tyr. This regulatory constraint was overcome in the strains expressing feedback inhibition-insensitive TyrC protein, resulting in a large increase in l-Tyr production. The observed differences in l-Tyr production between PB12A and PB12C indicated that the conversion of PPA to HPP is a limiting step in the pathway that can be overcome by introducing the activity of a feedback inhibition-insensitive TyrA protein, like TyrC.

FIG. 3.

Results of the shake flask experiments. ▾, PB12G; ▪, PB12A; ⧫, PB12C; •, PB12CP.

TABLE 3.

Kinetic and stoichiometric parameters determined in shake flask experiments^a

Strain	Yl-Tyr/Glc (mg/g)	ql-Tyr (mg/g/h)
PB12G	6.3 ± 0.32	0.99 ± 0.24
PB12A	10.0 ± 3.0	1.1 ± 0.22
PB12C	66.6 ± 15.8	7.6 ± 0.55
PB12CP	64.3 ± 6.0	6.4 ± 0.82

^aParameters were calculated from 0 to 52 h.

To further characterize the best l-Tyr-producing strains, PB12C and PB12CP were grown using media and fermentation conditions similar to those employed for biotechnological processes. Cultures were grown in 1-liter bioreactors (BioFlo 110 modular fermentor system; New Brunswick Scientific, New Jersey). Operation parameters were controlled by the AFS-Biocommand bioprocessing software (New Brunswick Scientific, New Jersey). The culture conditions were as follows: working volume, 0.6 liter; temperature, 37°C; airflow rate, 0.6 liter/min; dissolved oxygen level maintained above 30% air saturation by increasing the stirrer speed (500 to 1200 rpm); and pH controlled at 7.0 by addition of 15% NH₄OH. The culture medium contained (per liter) 50 g glucose, 6 g Na₂HPO₄, 3 g KH₂PO₄, 5 g (NH₄)₂SO₄, 2 g sodium citrate·2H₂O, 10 g yeast extract, 1.5 g MgSO₄·7H₂O, 550 mg CaCl₂, 40 mg thiamine-HCl, 30 μg tetracycline, 100 μg carbenicillin, and 23.83 mg IPTG. A trace element solution contained (per liter) 8.74 μg Na-EDTA·2H₂O, 1.55 μg CoCl₂·6H₂O, 9.30 μg MnCl₂, 0.93 μg CuCl₂, 1.30 μg Na₂MoO₄·2H₂O, 20.96 μg zinc acetate·H₂O, 62.49 μg Fe(III) citrate, and 1.86 μg H₃BO₃. The cell glucose, l-Tyr, and l-Phe contents were determined as described above. Acetic acid, 3-dehydroshikimic acid (DHS), and shikimic acid (SHIK) were quantified as described elsewhere (3). l-Tyr is insoluble in water at concentrations above 0.4 g/liter at neutral pH; for this reason, 1-ml portions of fermentation samples were treated with 50 μl of 6 N HCl, mixed, and incubated 30 min at 42°C, samples were centrifuged at 12,000 rpm to remove the cells (Eppendorf 5410), and supernatants were filtered and diluted for high-performance liquid chromatography analysis.

After 21 and 27 h of incubation, strains PB12C and PB12CP reached final biomass concentrations of 13.9 ± 0.8 and 11.1 ± 2.1 g (dry weight)/liter, respectively. Differences in several parameters related to l-Tyr production were observed between these strains (Table 4). PB12CP, which expresses the tyrC and pheA_CM genes, produced an l-Tyr titer that was 46% higher than that of PB12C. For PB12CP the amount of l-Tyr produced per gram (dry weight) of cells and the Y_l-Tyr/Glc were 85 and 48% higher than the values for PB12C, respectively. The Y_l-Tyr/Glc for PB12C and PB12CP corresponded to 7.7 and 11.9%, respectively, of the maximum theoretical yield (0.553 g/g) (32). Finally, the q_l-Tyr for PB12CP was 36% higher than that for PB12C. Under these culture conditions, strain PB12CP displayed the highest l-Tyr yields and titer. Based on these results, it is clear that a strain expressing both tyrC and pheA_CM is a better l-Tyr producer than a strain expressing only tyrC. However, considering the results observed in shake flask experiments, the positive effect of expressing both tyrC and pheA_CM appears to be dependent on the culture conditions.

TABLE 4.

Parameters determined from data generated in bioreactor experiments using strains PB12C and PB12CP

Strain	Specific glucose consumption rate (g/g/h)	Final dry wt of cells (g/liter)	Final l-Tyr titer (g/liter)	Yl-Tyr/Glc (mg/g)	Amt of l-Tyr produced (mg/g [dry wt] of cells)	ql-Tyr (mg/g/h)	Final SHIK titer (g/liter)	Final DHS titer (g/liter)
PB12C	0.17 ± 0.01	13.9 ± 0.8	2.0 ± 0.00	44.6 ± 8.1	153.5 ± 10.7	7.30 ± 0.51	1.2 ± 0.04	1.2 ± 0.01
PB12CP	0.14 ± 0.01	11.1 ± 2.1	3.0 ± 0.29	66.0 ± 0.1	284.0 ± 4.8	9.96 ± 0.29	2.5 ± 1.2	1.7 ± 0.04

Metabolites of l-Phe and common aromatic pathways were also measured (l-Phe, DHS, and SHIK) (Fig. 1 and Table 4). l-Phe was not detected in culture supernatants from either of the two strains (data not shown). On the other hand, accumulation of DHS and SHIK in the culture medium indicated the presence of rate-limiting steps in the common aromatic pathway (3, 16, 30). This suggests that the TyrC and PheA_CM enzymes efficiently compete for carbon with the l-Phe pathway. The amounts of l-Tyr produced by PB12C and PB12CP represent 45 and 42%, respectively, of the sum of the amounts of the measured metabolites (DHS, SHIK, and Tyr). Finally, acetic acid was not detected in culture supernatants. Strains with an inactive phosphotransferase system display very low acetic acid production, even with elevated glucose concentrations (28).

Previous studies of microbial production of l-Tyr employed classical mutagenesis methods for the generation of strains that overproduce this amino acid. More recently, mutagenesis of genes encoding specific enzymes that are feedback inhibited by l-Tyr has proven to be a successful approach for generating production strains. Despite the utility of these approaches, their application involves a requirement to generate and screen a large number of mutants to find strains with an increased capacity to overproduce l-Tyr (or l-Phe) or feedback inhibition-resistant enzymes that retain their catalytic activity. In this work, we took advantage of the natural diversity in the TyrA family of enzymes and designed a rational strategy to overcome feedback inhibition regulation in the l-Tyr pathway in E. coli. TyrC from Z. mobilis is not inhibited by l-Tyr; thus, by expressing TyrC in E. coli it was possible to reduce considerably the negative control that l-Tyr exerts in its terminal pathway. Our results showed that this approach greatly increased carbon flow to the l-Tyr pathway. On the other hand, in enzymatic assays it was found that the PheA_CM domain of P-protein was weakly inhibited by l-Tyr. In bioreactor cultures it was determined that expression of this domain increases the l-Tyr production capacity in a strain also expressing TyrC. The strains generated in this study are not final production strains; they can be further improved with additional genetic alterations. If all the metabolic bottlenecks in the common aromatic pathway could be alleviated, it would be possible to increase the amount l-Tyr produced by PB12C and PB12CP twofold. The modifications needed to meet this goal are overexpression of the aroL and ydiB genes in the common aromatic pathway (30; Takai et al., U.S. patent application 20050277179). In addition, it should be possible to further improve these strains by expression of the tktA gene to increase E4P availability (17) and by inactivation of the transcriptional regulator TyrR (32; Takai et al., U.S. patent application 20050277179). The approach described here is one of the strategies that can be used to eliminate the regulation in the l-Tyr biosynthetic pathway, with the purpose of producing l-Tyr and compounds derived from l-Tyr. Finally, the results obtained in this study show that a metabolic engineering strategy, based on the utilization of natural diversity, is a viable rational alternative to approaches based on random mutagenesis.