ARTICLE IN PRESS
Metabolic Engineering 8 (2006) 196?C208
Metabolic engineering of the E. coli L-phenylalanine pathway for the production of D-phenylglycine (D-Phg)
Ulrike Mullera,b,, Friso van Assemac, Michele Gunsiord,1, Sonja Orfa, ?? Susanne Kremera, Dick Schipperb, Anja Wagemansc, Craig A. Townsendd, Theo Sonkec, Roel Bovenberga,b, Marcel Wubboltsa,c
DSM Biotech GmbH, Karl-Heinz-Beckurts-Str. 13, 52428 Julich, Germany ?? b DSM Anti-Infectives, P.O. Box 425, 2600 AK Delft, The Netherlands c DSM Pharma Chemicals, Advanced Synthesis, Catalysis & Development, P.O. Box 18, 6160 MD Geleen, The Netherlands d Department of Chemistry, The Johns Hopkins University, 3400 North Charles Street, Baltimore, Maryland 21218, USA Received 6 July 2005; received in revised form 2 December 2005; accepted 5 December 2005 Available online 7 February 2006
D-phenylglycine (D-Phg) is an important side chain building block for semi-synthetic penicillins and cephalosporins such as ampicillin and cephalexin. To produce D-Phg ultimately from glucose, metabolic engineering was applied. Starting from phenylpyruvate, which is the direct precursor of L-phenylalanine, an articial D-Phg biosynthesis pathway was created. This three-step route is composed of the enzymes hydroxymandelate synthase (HmaS), hydroxymandelate oxidase (Hmo), and the stereoinverting hydroxyphenylglycine aminotransferase (HpgAT). Together they catalyse the conversion of phenylpyruvate via mandelate and phenylglyoxylate to D-Phg. The corresponding genes were obtained from Amycolatopsis orientalis, Streptomyces coelicolor, and Pseudomonas putida. Combined expression of these activities in E. coli strains optimized for the production of L-phenylalanine resulted in the rst completely fermentative production of D-Phg. r 2006 Elsevier Inc. All rights reserved.
Keywords: Aromatic amino acids; D-Amino acids; Renewable; Sustainable; Fine chemicals
1. Introduction The D-amino acids, D-phenylglycine (D-Phg) and D-4hydroxyphenylglycine (D-HPG) are important precursors used for the preparation of semi-synthetic cephalosporins and penicillins. Currently, D-HPG is obtained in a two-step chemo-enzymatic synthesis based on hydantoinase technology (Liese et al., 2000), and D-Phg is
made by classical or enzymatic resolution of a racemic mixture (reviewed in Wegman et al., 2001) derived from petrochemical feedstocks. As an alternative to current production methods, a
Corresponding author. DSM Anti-Infectives, P.O. Box 425, 2600 AK Delft, The Netherlands. E-mail address: email@example.com (U. Muller). ?? 1 Current address: National Cancer Institute, Bldg 37, 9000 Rockville Pike, Bethesda, Maryland 20892, USA.
complete fermentative route based on renewable and sustainable resources would be of commercial interest. Several successful examples of metabolic engineering for microbial production of aromatic amino acids like L-phenylalanine, or L-tryptophan and derived compounds like indigo are known (LaDuca et al., 1999; Bongaerts et al., 2001). Broadening the aromatic amino acid pathway of Escherichia coli by metabolic pathway engineering to DPhg would result in a microbial strain that allows the production of D-Phg in a single fermentation step. Thus far, no direct biosynthetic pathway to free D-Phg has been identied in nature. Nevertheless several actinomycete peptide antibiotics like the streptogramin (Cocito, 1979) and vancomycin groups of antibiotics (Nagarajan, 1991) contain either Phg or hydroxylated derivatives thereof such as D-HPG and L-3,5-dihydroxyphenylglycine (L-DHPG), as building block.
1096-7176/$ - see front matter r 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ymben.2005.12.001
ARTICLE IN PRESS
U. Muller et al. / Metabolic Engineering 8 (2006) 196?C208 ?? 197
Recently the pathways for L-HPG and L-DHPG biosynthesis have been elucidated illustrating that the biosynthesis of each compound follows completely different routes. While the pathway to DHPG originates with malonyl-CoA and acetyl-CoA, and involves a type III polyketide synthase (Li et al., 2001b; Pfeifer et al., 2001), LHPG is synthesized from tyrosine using a hydroxymandelate synthase (HmaS), a novel dioxygenase, as the pivotal enzyme (Choroba et al., 2000; Hubbard et al., 2000). In both cases the last biosynthetic step is a transamination catalysed by an L-hydroxyphenylglycine aminotransferase converting either 4-hydroxyphenylglyoxylate or 3,5-dihydroxyphenylglyoxylate to the corresponding L-amino acids L-HPG (Choroba et al., 2000) or L-DHPG (Pfeifer et al., 2001). Orthologous genes encoding the enzymes involved in the biosynthesis of L-HPG have recently been identied in several actinomycete strains, e.g., A. orientalis (Choroba et al., 2000), Streptomyces lavendulae (Chiu et al., 2001), Streptomyces coelicolor (Hojati et al., 2002), and Nocardia uniformis (Gunsior et al., 2004), which, respectively, produce chloroeremomycin, complestatin, calcium dependent antibiotic, and nocardicin A. These actinomycetes synthesize L-HPG from 4-hydroxyphenylpyruvate through the action of three enzymes,
HmaS (decarboxylation and oxidation to L-4-hydroxymandelate), hydroxymandelate oxidase Hmo (oxidation to 4hydroxyphenylglyoxlate) and L-4-hydroxyphenylglycine aminotransferase (transamination to L-HPG). The latter enzyme, L-4-hydroxyphenylglycine aminotransferase uses L-tyrosine as an amino-donor thereby releasing
4-hydroxyphenylpyruvate as the starting molecule again (Choroba et al., 2000; Hubbard et al., 2000). In order to construct a biosynthetic pathway to the free D-amino acids D-Phg or D-HPG, part of the L-HPG pathway needs to be combined with a D- rather than an L-4-hydroxyphenylglycine aminotransferase. The presence of such a D-4-hydroxyphenylglycine and D-phenylglycine aminotransferase has previously been described for Pseudomonas putida LW-4 (Van den Tweel et al., 1986, 1988) and Pseudomonas stutzeri (Wiyakrutta and Meevootisom, 1997). We investigated whether the combination of the rst two steps of the L-HPG actinomycete biosynthetic pathway and
a D-4-hydroxyphenylyglycine aminotransferase results in a new metabolic pathway for D-Phg starting from the intermediate phenylpyruvate (Fig. 1), the direct precursor of L-phenylalanine. The introduction of this ??articial?? pathway into an L-phenylalanine producing E. coli strain resulted in fermentative production of D-Phg. To our knowledge, this is the rst report of multi-step engineering to produce free D-Phg from renewable resources such as glucose. 2. Material and methods 2.1. Bacterial strains and culture media Amycolatopsis orientalis NRRL 18098 (US patent 5,843,437) was obtained from the ARS (Agricultural Research Service) Patent Culture Collection, Peoria, Illinois, USA. A. orientalis was cultivated in 10 g L1 glucose, 5 g L1 yeast extract, 20 g L1 starch, 1 g L1 casamino acids, pH 7.5 with NaOH at 28 C. Streptomyces coelicolor A3(2) M145, kindly obtained from Professor M.J. Bibb of the John Innes Institute, Norwich (UK), was cultivated at 28 C in YE-ME medium containing 3 g L1 yeast extract, 5 g L1 peptone, 3 g L1 malt extract, 10 g L1 glucose, 340 g L1 sucrose. 10 g L1 glycine and 5 mM MgCl2 were added after sterilization. Pseudomonas putida LW-4 (NCIMB 12565) was obtained from the National Collection of Industrial and Marine Bacteria, Aberdeen, Scotland, UK. P. putida was grown at 30 C in LB medium. Escherichia coli strains and plasmids used in this study are listed in Tables 1 and 2. DH5a, and TOP10 were used as hosts for plasmid construction. During strain construction cultures were grown at 30, 33, 37, or 42 C in LB broth. Antibiotics were used when needed at the following concentrations: kanamycin (25 or 50 mg L1 ), chloramphenicol 25 mg L1 , ampicillin 100 mg L1 . To test in vivo D-Phg production, the recombinant E. coli strains were cultivated in different mineral media of the following compositions: Mineral salt medium A: Na citrate 3H2 O 1:0 g L1 , MgSO4 7H2 O 0:3 g L1 , KH2 PO4 3:0 g L1 , 1 1 K2 HPO4 12:0 g L , NaCl 0:1
g L , NH4 2 SO4
Fig. 1. Biosynthesis of D-phenylglycine (D-Phg). 1.
Hydroxymandelate synthase (HmaS) from A. orientalis hmaSAo or S. coelicolor hmaS Sc ; 2. Hydroxymandelate oxidase (Hmo) from A. orientalis hmoAo or S. coelicolor hmoSc ; 3. D-(4-Hydroxy)phenylglycine aminotransferase (HpgAT) of Pseudomonas putida (hpgAT).
ARTICLE IN PRESS
198 Table 1 Escherichia coli strains used in this study Strain TOP10 DH5a LJ110 F25 F26 KB532 K7 K8 K9 K10 Relevant properties or genotype araD139 D(ara?Cleu)7697 lacZDM15 recA endA lacZDM15 recA endA W3110; F ; l ; fnr LJ110 DtyrB::cat CmR LJ110 DaspC::cat CmR DpheA tyrA aroF fbr DtyrR thiA hsdR17 endA glnV44(AS) KB532 DtyrB::FRT-cat-FRT, KB532 P1F25 KB532 DtyrB::FRT KB532 DaspC::FRT-cat-FRT, KB532 P1F26 KB532 DtyrB::FRT DaspC::FRT-cat-FRT, K8 P1F26 Source or reference Invitrogen Bethesda Research Laboratory (Zeppenfeld et al., 2000) This study This study DSM, proprietary strain This study This study This study This study U. Muller et al. / Metabolic Engineering 8 (2006) 196?C208 ??
Table 2 Plasmids used in this study Plasmid pBAD/Myc-HisC pZerOTM -2 pJF119EH pCRs -Blunt II-TOPOs pBAD-hmaSAo pBAD-hmoAo pBAD-hmaSSc pBAD-hmoSc pZErO-hpgAT pBAD-hpgAT pJF-hmaSAo pJF-hmoSc pJF-hpgAT Pp pJF-hmaSAo -hmoSc -hpgAT pCR-bl-pheAfbr pJF-hmaSAo -hmoSc -hpgAT-pheAfbr pKD3 pKD46 pCP20 Relevant characteristics bla araC ParaBAD pBR322ori c-Myc epitope polyhistidine 6 His; expression vector kan Plac lacZa ccdB; cloning vector Ptac bla bla kan; TOPO TA cloning vector pBAD/Myc-HisC derivative; A. orientalis hmaS gene pBAD/Myc-HisC derivative; A. orientalis hmo gene pBAD/Myc-HisC derivative; S. coelicolor hmaS gene pBAD/Myc-HisC derivative; S. coelicolor hmo gene pZerOTM -2 derivative; P. putida NCIMB 12565 chromosomal DNA with complete hpgAT gene and adjacent DNA pBAD/Myc-HisC derivative; P. putida NCIMB 12565 hpgAT gene pJF119EH derivative; hmaSAo including RBS of pBAD-hmaSAo pJF119EH derivative; hmoSc including RBS of pBAD-hmoSc with change of codon 2, to a more frequently used one in E. coli pJF119EH derivative; hpgAT of pBAD-hpgAT including RBS pJF119EH derivative; hmaSAo of pJF-hmaSAo , hmoSc of pJF-hmoSc , hpgAT of pJF-hpgAT pCRs -Blunt II-TOPOs derivative; pheAfbr gene of E. coli including the original RBS pJF119EH derivative; hmaSAo , hmoSc , hpgAT, pheAfbr bla FRT-cat-FRT bla ab exo (red recombinase), temperature conditional replicon bla cat, temperature sensitive replicon, temperature inducible FLP recombinase Source Invitrogen Invitrogen (Furste et al., 1986) ?? Invitrogen This study This study This study This study This study This study This study This study This study This study This study This study (Datsenko and Wanner, 2000) (Datsenko and Wanner, 2000) (Cherepanov and Wackernagel, 1995)
5:0 g L1 , CaCl2 2H2 O 15:0 mg L1 , FeSO4 7H2 O 75:0 mg L1 , thiamine
HCl (vitamin B1) 5:0 mg L1 , and L-tyrosine 0:05 g L1 . Additional minerals were added in the form of a trace element solution 1 ml L1 composed of Al2 SO4 3 18H2 O 2:0 g L1 , CoCl2 6H2 O 0:7 g L1 , CuSO4 5H2 O 2:5 g L1 , H3 BO3 1 0:5 g L , MnCl2 4H2 O 20:0 g L1 Na2 MoO4 2H2 O 3:0 g L1 , NiSO4 6H2 O 2:0 g L1 , ZnSO4 7H2 O 15:0 g L1 . A stock solution of glucose 30 g L1 was autoclaved separately and added to the sterilized medium to a nal concentration of 4 g L1 . Mineral salt medium B: Identical to mineral salt medium A, but with higher amounts of L-tyrosine 0:1 g L1 , and glucose 10 g L1 , and addition of L-aspartate 3 g L1 .
Mineral salt medium C: Identical to mineral salt medium B, but with higher amounts of Na citrate 3H2 O 1:5 g L1 , MgSO4 7H2 O 0:9 g L1 , NaCl 1:0 g L1 , FeSO4 7H2 O 1:125 mg L1 , thiamine HCl 75:0 mg L1 , 1 1 L-tyrosine 0:2 g L , and glucose 20 g L . Furthermore, medium C did not contain K2 HPO4 . 2.2. Genetic methods Standard methods were used for PCR amplication, plasmid construction, transduction with phage P1, analyses of DNA fragments, and DNA sequencing (Sambrook et al., 1989). Plasmid DNA was isolated using a Qiaprep Spin Miniprep kit (Qiagen, Hilden, Germany). A QIAquick Gel
ARTICLE IN PRESS
U. Muller et al. / Metabolic Engineering 8 (2006) 196?C208 ?? 199
extraction kit was used to isolate DNA fragments from agarose gels (Qiagen, Hilden, Germany). DNA sequencing was provided by GATC AG, Konstanz, Germany. 2.3. Isolation of the hpgAT gene from P. putida NCIMB 12565 2.3.1. Construction of gene library Chromosomal DNA of P. putida NCIMB 12565 was extracted from exponentially growing cells (OD620nm 1.9) using standard protocols (Ausubel et al., 1995), treated with RNase 20 mg L1 , and subsequently with phenol/ chloroform/isoamyl alcohol (25:24:1) to remove proteins. The chromosomal DNA was then partially digested with Sau3AI and run on a 0.6% agarose gel. DNA fragments between 4 and 10 kb in size were isolated. Vector DNA was prepared by the digestion of 1 mg of pZerOTM -2 with BamHI according to the Invitrogen protocol. Vector DNA and P. putida chromosomal DNA fragments were ligated with T4 DNA ligase and transformed into chemically competent E. coli TOP10 cells. Transformants were plated onto LB medium with 50 mg L1 kanamycin. In total 5000 colonies were obtained forming the primary gene library. All colonies were pooled in LB medium supplemented with 50 mg L1 kanamycin. After addition of glycerol to a nal concentration of 15 vol%, the DNA library was stored in aliquots of 1 ml at 80 C. 2.3.2. Screening for clones with D-HpgAT activity Cultures of 1,800 colonies were prepared in 150 ml LB medium supplemented with 50 mg L1 kanamycin in microtiter plates. The cultures were grown overnight at 28 C and harvested by centrifugation in an Eppendorf 5804 R centrifuge (Eppendorf, Hamburg, Germany). The cells were washed with 50 mM KPO4
buffer, pH 7.0, and re-suspended in 180 ml reaction mix (100 mM potassium phosphate, pH 7.0, 15 mM a-ketoglutarate, 0.1 mM pyridoxal phosphate and 0.5 vol% Triton X-100). The reaction was started by adding D-4-hydroxyphenylglycine to a nal concentration of 5 mM. The absorption at 340 nm in each well was monitored during 20 min using an Optimax microtiter plate reader (Molecular Devices, Sunnyvale, California, USA). A negative control (non-transformed E. coli TOP10) and a positive control (P. putida NCIMB 12565) were treated accordingly. Of 1,800 clones screened, one showed signicant increase in the absorption at 340 nm relative to the negative control due to the formation of 4-hydroxyphenylglyoxylate. This clone contained the P. putida HpgAT encoding gene on a 12 kb plasmid, pZErO-hpgAT. 2.4. Plasmid constructions Several individual and multiple gene constructs in various expression plasmids were prepared to assay enzyme activity before nal assembly into an operon for the production of D-Phg. Chromosomal DNA of A. orientalis,
S. coelicolor and plasmid pZErO-hpgAT were used as templates for PCR amplication in which unique restriction sites were appended to each gene as indicated in the primer list (Table 3). The resulting amplicons were transferred individually into the appropriately digested pBAD/Myc-HisC plasmid to yield pBAD-hmaS Ao , pBADhmoAo , pBAD-hmaS Sc , pBAD-hmoSc , and pBAD-hpgAT. Fusions to the Myc epitope and His6 tag were prevented by including the original stop codon of the genes amplied. Furthermore, all genes were cloned via a translation-start (ATG) fusion. To enable in-frame cloning in the NcoI site without alteration of the second codon restriction enzymes differing in recognition and cleavage site such as BsaI and BspHI or the cloning procedure as described by Dietmaier et al. (1993) were utilized. Cloning of the genes without undesired mutations was veried by sequencing. The pBAD plasmids were used as templates for construction of the articial D-Phg operon. Unique restriction sites and the optimized Shine Delgarno sequence present in pBAD/Myc-HisC were added to the genes as indicated in the primer list (Table 3) by PCR amplication methods. The resulting products were transferred individually into the appropriately digested pJF119EH plasmid to yield pJF-hmaS Ao , pJF-hmoSc , and pJF-hpgAT. The clone was sequence veried and enzyme activity conrmed active expression of the single genes following IPTG induction (data not shown). Cloning of the hmoSc gene as an XbaI/SphI fragment and the hpgAT gene as an HindIII fragment behind the hmaS Ao gene in the appropriate sites of plasmid pJF-hmaS Ao resulted in the construction of plasmid pJF-hmaS Ao -hmoSc hpgAT. The pheAfbr gene, a truncated and feedback inhibitionresistant version of pheA (Backman and Balakrishnan, 1988), was PCR amplied with the primer pair listed in Table 3 from chromosomal DNA of E. coli LJ110 as template. The resulting amplicon retained the
original Shine Delgarno sequence but possessed a new stop codon intentionally deleting a region of the pheA gene encoding amino acid residues 338?C387. Cloning of the PCR fragment in pCRs -Blunt II-TOPOs resulted in the construction of pCR-Bl-pheAfbr . Sequence analyses veried the correct cloning. Subsequently, the pheAfbr gene was cloned as an HindIII fragment behind the hpgAT gene in plasmid pJFhmaS Ao -hmoSc -hpgAT, which was partially digested with HindIII, resulting in plasmid pJF-hmaS Ao -hmoSc -hpgATpheAfbr . Correct cloning was conrmed by restriction enzyme digestions. Complementation of the phenylalanine auxotrophy of E. coli KB532 conrmed active expression of pheAfbr . 2.5. Construction of deletion mutants The mutants with an aspC or tyrB gene knockout were obtained by a one-step inactivation protocol (Datsenko and Wanner, 2000) using PCR primers as given in Table 3. The FRT-cat-FRT cassettes were amplied by PCR using
ARTICLE IN PRESS
200 U. Muller et al. / Metabolic Engineering 8 (2006) 196?C208 ?? Table 3 List of synthetic oligonucleotides used in this study Primer name A hmaSAo_fowA hmaSAo_revA hmoAo_fowA hmoAo_revA hmaSSc_fowA hmaSSc_revA hmoSc_fowA hmoSc_revA hpgAT_fowA hpgAT_revA hmaSAo_fowB hmaSAo_revB hmoSc_fowA hmoSc_revB hpgAT_fowB hpgAT_revB pheA_fowA pheA_revA B sense tyrB as tyrB sense aspC as aspC Nucleotide sequence Primer used for gene cloning 50 -GTCCAC GGTCTC C CATG CAGAATTTCGAGAT-30 50 -ACATCCC AAGCTT CACGTTCGAGGTC-30 50 -CGCTCGG TCATG
ACGTACGTTTCCCTG-30 50 -ACGAAG AAGCTT ATCAAACAACCCCCAG-30 50 -ATGCCGCCCAGTGACATCGCGTACGC-30 50 -CCCTC GGTACC
AGGTCATCGGCCGGCCACTTCC-30 50 -ATGCGGGAGCCGCTCACGCTCGAC-30 50 -CCAACT GGTACC TGGTCATCCGTGGCTCCTGTCTCG-30 50 -GTGCAC GGTCTC G CATG TCTATTTATAGCGATTATGAACGTAAAAC-30 50 -GTGCAC GGTCTC C TCGA GTTAGCCCAGGAGGTTTTCTTCAGC-30 50 -TGG GAATTC AGGAGGAATTAACCATGCAG-30 50 -CGGCCAGG TCTAGA TACGTCATCGCCG-30 50 -TGGG TCTAGA
GGAGGAATTAACCATGCGcGAGCCG-30 a 50 -GAATTCCCATA GCATGC
CTGGTCATCCGTGGCTCC-30 50 -TTTCCC AAGCTT ACAGGAGGAATTAACCATG-30 50 -GTACCAGCTGCA AAGCTT GAGTTAGCCCAG-30 50 -CGCCTA AAGCTT
CGGTACCATTTGATAACAAAAAGGC-30 50 -CGCCTA AAGCTT
GAATTCATGGATTACCGTGAATCGG-30 Restriction site BsaI (NcoI) HindIII BspHI (NcoI) HindIII (NcoI) KpnI (NcoI) KpnI BsAI (NcoI) BsAI (XhoI) EcoRI XbaI XbaI SphI HindIII HindIII HindIII HindIII
Primer used for gene deletion
CATATGAATATCCTCCTTAG CCTGATAGCGGACTTCCCTTCTGTAACCATAATGGAACCTCG TGTGTAGGCTGGAGCTGCTTCG AATGCTTACAGCACTGCCACAATCGCTTCGCACAGCGGAG CATATGAATATCCTCCTTAG
A: Italics represent the Shine Delgarno (SD) sequences, underlined text denotes the restriction site sequences, and bold type indicates the start/stop codons. B: Italics represent the sequence homologous to the gene to be deleted, underlined text denotes the sequence homologous to plasmid pKD3 for amplication of the resistance cassette. a changing codon 2, indicated by non-capital letters, to a more frequently used one in E. coli.
these primers and pKD3 (Datsenko and Wanner, 2000) as template. After purication, amplied DNA was transformed by electroporation into E. coli LJ110 pKD46. The resulting chloramphenicol-resistant recombinant strain containing FRT-cat-FRT in the deleted region of tyrB was designated F25; recombinant strain F26 contained FRT-cat-FRT in the deleted region of aspC. A phage P1kc lysate prepared from F25 (DtyrB::FRT-cat-FRT) was used to transfer this mutation to E. coli KB532 to produce K7. The cat gene was removed from the chromosome with FLP recombinase by using a temperature-conditional helper plasmid (pCP20). After removal of the helper plasmid, the resulting
chloramphenicol-sensitive strain (DtyrB::FRT) was designated strain K8. A phage P1kc lysate prepared from F26 (DaspC::FRT-cat-FRT) was used to transfer this mutation to E. coli KB532 to produce K9 (DaspC::FRT-cat-FRT) and in K7 to produce K10 (DtyrB::FRT DaspC::FRT-cat-FRT). Deletion mutants were veried by analysis of PCR products and antibiotic resistance. 2.6. Enzyme activity determinations 2.6.1. Preparation of cell free extracts for enzyme activity determinations Single colonies of the E. coli TOP10 strains harbouring the pBAD-hmaS Ao , pBAD-hmoAo , pBAD-hmaS Sc , pBAD-
hmoSc , and pBAD-hpgAT were cultivated in 50 ml LB medium containing 100 mg L1 carbenicillin at 30 C. At OD620nm 1.2, the cells were induced by the addition of 0.002% (nal concentration) L-arabinose. Single colonies of the E. coli DH5a strains harbouring the plasmids pJF-hmaS Ao , pJF-hmoSc , pJF-hpgAT, pJFhmaS Ao -hmoSc -hpgAT were used to inoculate 10 ml of LB medium containing ampicillin 100 mg L1 and incubated at 30 C for 16 h. 1 ml of these cultures was subsequently used to inoculate 50 ml of the same medium. Cells were grown at 30 C at 180 rpm. At OD620nm 0.8, the cells were induced by the addition of 0.1 mM IPTG. After 4 h, the cells were harvested and washed with 100 mM potassium phosphate buffer, pH 7.5. Aliquots of washed cells were frozen at 20 C for later use. As a control, E. coli TOP10 harbouring plasmid pBAD/Myc-HisC or E. coli DH5a harbouring plasmid pJF119EH was treated accordingly. Crude extracts were prepared either by sonication or with B-PERTM (in phosphate buffer) (Pierce, Rockford, Illinois, USA) immediately before use. The cell pellet from 25 ml of cultivation was resuspended in 1 ml 200 mM potassium phosphate buffer pH 7.5, and subjected to sonication (30 10 s bursts with 10 s rests). The lysate
was centrifuged for 30 min at 16,000 g to pellet cell debris. The claried extract was used for activity determination.
ARTICLE IN PRESS
U. Muller et al. / Metabolic Engineering 8 (2006) 196?C208 ?? 201
Cells from 12.5 ml of cultivation were resuspended in 1 ml B-PERTM Bacterial Protein Extraction Reagent (in phosphate buffer) (Pierce, Rockford, Illinois, USA) by vigorously vortexing until the cell suspension was homogenous. After additional vortexing for 1 min, the suspension was centrifuged at 16,000 g for 10 min and the supernatant was collected for immediate use. The protein concentration of the crude extracts was determined according to Bradford (1976) by using the Rotis -Quant protein assay reagent (Roth, Karlsruhe, Germany), with bovine serum albumin as a standard. 2.6.2. Activity assay of HmaS The reaction mixture contained 200 mM potassium phosphate buffer pH 7.5, 5 mM 4-hydroxyphenylpyruvate or 5 mM phenylpyruvate, 44 mM ascorbate, 0.3 mM FeSO4 , and crude extract in a nal concentration of 0:6 g L1 of soluble protein. In the case of 4-hydroxyphenylpyruvate the reaction mixture also contained 9.6% ethanol because a stock solution of 50 mM 4-hydroxyphenylpyruvate in 96% ethanol was used. The reaction was started by the addition of the crude extract and allowed to proceed for a xed time at 28 C. Following the addition of 1 N HCl (0.16 N HCl nal concentration) to stop the reaction, the solution was centrifuged for 20 min at 16,000 g. Products were analysed by HPLC (vide infra). 2.6.3. Activity assay of Hmo towards
4-hydroxyphenylglyoxylate (340 nm) of 3; 510 M1 cm1 was used to calculate the enzymatic (specic) activity. To analyse the activity of HpgAT towards L- and Dglutamate, the reverse reaction?ªconversion of 4-hydroxyphenylglyoxylate to HPG?ªwas analysed
spectrophotometrically by the decrease in absorption at 340 nm caused by utilization of 4-hydroxyphenylglyoxylate. 1 ml reaction mixture containing 100 mM potassium phosphate buffer pH 7.0, 60 mM L- or D-glutamate, 0.1 mM pyridoxal phosphate and crude extract at a nal concentration of 0:1 g L1 of soluble protein was incubated in a cuvette at 20 C. The reaction was started by the addition of
4-hydroxyphenylglyoxylate (0.67 mM nal concentration) into the reaction mixture. 2.6.6. Activity of HpgAT towards D-Phg and phenylglyoxylate The reaction mixture contained 100 mM potassium phosphate buffer pH 8.0, 15 mM a-ketoglutarate, 0.1 mM pyridoxal phosphate and crude extract at a nal concentration of 0:01 g L1 of soluble protein. The reaction was started by the addition of 4 mM D-Phg and allowed to proceed for a xed time period at 30 C. At certain time intervals aliquots of 1 ml were taken and transferred into 0.4 ml 1 M H3 PO4 to stop the reaction. Samples were analysed by HPLC (vide infra).
To examine the enantioselectivity of HpgAT for D-Phg formation, the reaction mixture contained 10 mM phenylglyoxylate, 100 mM L- or D-glutamate, 10 mM pyridoxal phosphate, 0:05 g L1 protein solution (cell free extracts of E. coli TOP10 pBAD-hpgAT) in 100 mM potassium phosphate buffer pH 8.0. 1 ml samples were taken after 0, 5, 10, 15, 30, 60, and 120 min of incubation at 35 C. The reaction was stopped by addition of 2 ml 0.2 mM H3 PO4 . The amount of L- and D-Phg and the amount of L- and D-glutamate in the samples were determined by HPLC. 2.7.
100 ml of the crude extract containing 0.5?C0.8 mg protein was incubated with 100 mM potassium phosphate buffer pH 7.5, 2 mM DL-4-hydroxymandelate, and 20 mg L1 catalase in a total volume of 1 ml. The oxidation of 4hydroxymandelate was monitored
spectrophotometrically at 340 nm. To correct for non-specic oxidation of 4hydroxymandelate, control assays were run using an assay mixture without cell free extract. 2.6.4. Activity assay of Hmo towards L- or D-mandelate The reaction mixture contained 100 mM potassium phosphate buffer pH 7.5, 2 mM L- or D-mandelate, 20 mg L1 catalase, and cell free extract at a nal concentration of 0:6 g L1 of soluble protein. The reaction was started by the addition of substrate and allowed to proceed for a xed time at 28 C. Following the addition of 1 N HCl (0.16 N HCl nal concentration) to stop the reaction, the solution was centrifuged for 20 min at 16,000 g. Products were analysed by HPLC (vide infra). 2.6.5. Activity of HpgAT towards D- or L-HPG and D- or Lglutamate The same spectrophotometric assay as described for whole cells during the screening procedure was used to test the activity of HpgAT towards D-HPG and L-HPG. Instead of whole cells, crude extracts of hpgAT overexpressing E. coli cells were added to nal concentration of 0:0120:1 g L1 total protein. The molar absorptivity of
formation in vitro
To test in vitro product formation, a crude extract of E. coli DH5a pJF-hmaS Ao -hmoSc -hpgAT was prepared with B-PER buffer. This crude extract at a nal concentration of 0:3 g L1 soluble protein was used in a reaction mixture containing 200 mM potassium phosphate buffer pH 8.0, 5 mM phenylpyruvate, 44 mM ascorbate, 20 mg L1 catalase, 40 mM L-glutamate, and 0.1 mM pyridoxal phosphate. The assay was started by the addition of phenylpyruvate. Following the addition of 1 N HCl (0.16 N HCl nal concentration) to stop the reaction after 20 h, the solution was centrifuged for 20 min at 16,000 g. The amounts of mandelate, phenylglyoxylate and D-Phg produced were determined by HPLC (vide infra). 2.8.
production in vivo
2.8.1. Shake ask cultivations A single colony of KB532 pJF-hmaS Ao