Simultaneous improvement of catalytic activity and thermal stability of tyrosine phenol-lyase by directed evolution Eugene Rha 1 , Sujin Kim 1 , Su-Lim Choi 1 , Seung-Pyo Hong 2 , Moon-Hee Sung 2 , Jae J. Song 3 and Seung-Goo Lee 1 1 Industrial Biotechnology & Bioenergy Research Center, KRIBB, Daejeon, South Korea 2 BioLeaders Corp., Daejeon, South Korea 3 Molecular Bioprocess Research Center, KRIBB, JungUp, South Korea Introduction Tyrosine phenol-lyase (TPL) (EC4.1.99.2) is a tetra- meric enzyme that catalyzes the a,b-elimination and b-replacement of l-tyrosine, with pyridoxal-5¢-phos- phate (PLP) as the cofactor [1–3]. At high concentra- tions of ammonium pyruvate, the enzyme catalyzes the synthesis reaction of aromatic amino acids from phenolic substrates [4]. The resulting amino acids can be used as precursors for several neurotransmitters, such as l-DOPA (DOPA, 3,4-dihydroxyphenylala- nine), dopamine, epinephrine and norepinephrine [5]. In most living organisms, l-tyrosine is principally syn- thesized from l-phenylalanine. Yet, for the industrial production of l-tyrosine and its derivatives, attention has been focused on enzymatic synthesis using TPL Keywords N-terminal arm; protein engineering; structural relevance; Symbiobacterium toebii; tyrosine phenol-lyase Correspondence S G. Lee and J. J. Song, 111, Gwahangno, Yuseong, Daejeon 305-806, South Korea Fax: +82 42 860 4379 Tel: +82 42 860 4373 E-mail: sglee@kribb.re.kr; jjsong@kribb.re.kr (Received 17 April 2009, revised 11 August 2009, accepted 24 August 2009) doi:10.1111/j.1742-4658.2009.07322.x The tyrosine phenol-lyase from Symbiobacterium toebii was engineered to improve both its stability and catalytic activity by the application of ran- dom mutagenesis and subsequent reassembly of the acquired mutations. Activity screening of the random library produced four mutants with a two-fold improved activity, whereas parallel screening after heat treatment at 65 °C identified three mutants with half-inactivation temperatures improved by up to 5.6 °C. The selected mutants were then reassembled using the staggered extension PCR method, and subsequent screening of the library produced seven mutants with up to three-fold improved activity and half-inactivation temperatures improved by up to 11.2 °C. Sequence analyses revealed that the stability-improved hits included A13V, E83K and T407A mutations, whereas the activity-improved hits included the additional T129I or T451A mutation. In particular, the A13V mutation was propagated in the hits with improved stability during the reassembly– screening process, indicating the critical nature of the N-terminal moiety for enzyme stability. Furthermore, homology modeling of the enzyme structure revealed that most of the stability mutations were located around the dimer–dimer interface, including the N-terminus, whereas the activity- improving mutations were located further away, thereby minimizing any interference that would be detrimental to the co-improvement of the stabil- ity and catalytic activity of the enzyme. Abbreviations DOPA, 3,4-dihydroxyphenylalanine; LB, Luria–Bertani; PLP, pyridoxal-5¢-phosphate; StEP PCR, staggered extension PCR; T 1 ⁄ 2, half-inactivation temperature; TPL, tyrosine phenol-lyase. FEBS Journal 276 (2009) 6187–6194 ª 2009 The Authors Journal compilation ª 2009 FEBS 6187 from phenolics, such as hydroxylated or halogenated derivatives of phenol, 4-chlorophenol, 4-nitrophenol and catechol [1,6]. Although most structure–function studies of TPL have focused on enzymes from enteric bacteria, includ- ing Citrobacter freundii and Erwinia herbicola, thermo- philic enzymes are also considered to provide benefits as biocatalysts for enzymatic processes. For example, a thermostable TPL from Symbiobacterium species can maintain stability with high concentrations of phenolic substances, whereas enzymes from enteric bacteria are inactivated under the same conditions [7,8]. Therefore, a TPL with improved thermal stability or catalytic activity may be very useful for the development of an ideal enzymatic process for aromatic amino acids [9–11]. On a molecular level, protein stabilization is related to increased rigidity in an unstable structural unit, whereas improved enzyme activity is related to increased flexibility of the catalytic residues in the active site [12,13]. Therefore, many studies have reported the concomitant occurrence of increased sta- bility and compromised catalytic activity. Nonetheless, several recent studies have been successful in simulta- neously improving both activity and stability using directed evolution technology [14–16]. Thus, it would appear that such co-improvements will most probably occur when the two properties are combined with weak structural and functional interference [17]. Accordingly, to improve the potential of TPL as a biocatalyst, this study used random mutagenesis, fol- lowed by a staggered extension PCR (StEP PCR) to reassemble beneficial mutations. StEP PCR is an alter- native DNA shuffling technology that is based on a short-cycle PCR [12]. Three-dimensional modeling analysis of the consequent hits was also performed to investigate the simultaneous improvement of the activ- ity and stability of Symbiobacterium toebii TPL, as a better structural understanding of how proteins respond to mutations and recombination can help in the development of more ambitious enzyme engineer- ing strategies, such as increasing the probability of mutant sequences having the desired properties [18,19]. Results and Discussion Mutagenesis and characterization of mutant enzymes The error-prone PCR using S. toebii TPL as the template produced a library of TPL mutants, each containing two to six mutations distributed over the entire sequence (approximately 2.7 mutations in 1377 bp). The mutation frequency was determined by sequencing 10 clones randomly picked from the naive library (1.2 · 10 5 colonies). When cultivating 12 000 colonies from the mutagenesis library in Luria–Bertani (LB) medium and assaying them on microtiter plates at 37 °C, four ‘activity’ mutants, A1–A4, were identi- fied (Table 1). Meanwhile, parallel screening after heat treatment at 65 °C for 10 min highlighted three other ‘stability’ mutants, S1–S3 (Table 1). To quantify the stability of the selected mutants, the remaining activity was measured after incubation for 30 min at various temperatures between 37 and 75 °C. Table 1 shows a comparison of the apparent half- inactivation temperatures (T 1 ⁄ 2 ), which were up to 5.6 °C higher for the stability mutants than for the wild- type enzyme. Meanwhile, the activities of A1 and A2 were two-fold greater than that of the wild-type enzyme, whereas the activity of the stability mutants was rela- tively unchanged. The specific mutations in the activity and stability mutants are summarized in Table 1. Co-improvement of activity and thermal stability To reassemble the acquired mutations, StEP PCR was conducted using the hits from the random mutagenesis as a mixed template. From the resulting library (1.4 · 10 6 colonies), 10 colonies were randomly selected and their sequences analyzed. The number of crossovers was approximated to be 2.2 times along the gene size (1377 bp) when counting the crossovers as the recombination of mutations from different tem- plates. Next, after examining 1200 colonies from the Esc- herichia coli library for the remaining activity, mutants AS1–AS7 were selected on the basis of a co-improve- ment in stability and activity. For example, the T 1 ⁄ 2 values for AS4 and AS6 were 6.7 and 11.2 °C higher, respectively, and the catalytic activities were improved by 2.8- and 2.0-fold, respectively (Table 1). Isolates AS1–AS7 were analyzed to have fewer mutations (1.2 per gene, in contrast with 2.7 per gene in the A and S mutants), implying that many of the original mutations had been deleted during the reas- sembly process. Aligning each mutation against the sequences of A1–A4 and S1–S3 allowed the parental origin of the mutations to be estimated. For example, AS4, composed of A13V, E83K and T451A, was ana- lyzed to be a reassembly of S3 (harboring the A13V mutation), S2 (harboring the E83K mutation) and A2 (harboring the T451A mutation). A more comprehen- sive representation of the correlation between the mutations is shown in Fig. 1 (redrawn from the data in Table 1). Interestingly, mutants AS1–AS3, showing Directed evolution of tyrosine phenol-lyase E. Rha et al. 6188 FEBS Journal 276 (2009) 6187–6194 ª 2009 The Authors Journal compilation ª 2009 FEBS no significant improvement in activity, only contained mutations originating from the S1–S3 mutants (the gray-shaded ellipse in Fig. 1). Meanwhile, the four other mutants (AS4–AS7) included mutations from both the stability and activity mutants, and demon- strated a significant improvement in both stability and catalytic activity (Fig. 1). As such, the activity and sta- bility mutation combinations in Table 1 represent the synergistic recruitment of the original characteristics of the parental mutants. The wild-type protein and two mutants, AS4 and AS7, were purified to homogeneity with purification yields of over 40%, and investigated for their stability and activity at temperatures between 40 and 80 °C. As a result, the specific activities of the mutants were at least two-fold higher than that of the wild-type across a broad temperature range (Fig. 2A). When heated for 30 min in 0.1 m potassium phosphate buffer (pH 8.0), the thermal stability of both mutants was confirmed to be up to 10 °C higher than that of the wild-type (Fig. 2B). When the pH properties of the purified enzymes were investigated, the AS4 and AS7 enzymes displayed an alkaline shift for their maximum activity (Fig. 2C). In addition, the mutant enzymes exhibited a higher remaining activity than the wild-type after incu- bation for 36 h at alkaline pH (Fig. 2D). Extraction of structural information from the evolutionary process Directed evolution has instigated a new enzyme engi- neering paradigm for improving enzyme properties without reliance on structural data [18,20]. This tech- nology takes advantage of the natural process in which Table 1. Genetic and catalytic changes of S. toebii TPL during evolutionary engineering. Library Screening Name Activity mutations Stability mutations Relative activity a (fold) Stability change b (°C) Random mutation (first screening) High activity A1 E42D c , T129I 2.04 )1.3 A2 A196T, T451A 1.99 )1.3 A3 T129I 1.69 )2.8 A4 T129I, V262A 1.47 +1.5 High stability S1 T407A 0.88 +5.6 S2 E83K 1.05 +3.1 S3 A13V, I457F 1.27 +5.5 Reassembly (second screening) High activity and stability AS1 A13V, T407A 1.12 +10.1 AS2 A13V, E83K, I457F 1.42 +5.6 AS3 A13V, E83K 1.42 +6.7 AS4 T451A A13V, E83K 2.84 +6.7 AS5 T451A A13V, E83K, T407A 1.87 +11.2 AS6 T129I A13V, E83K, T407A 1.99 +11.2 AS7 T129I A13V 2.10 +7.5 a Fold increase in activity of mutant enzymes at 37 °C compared with that of wild-type TPL. b Increase in T 1 ⁄ 2 when heated for 30 min under standard assay conditions compared with that of wild-type TPL (T 1 ⁄ 2 =63°C). c Italic letters indicate mutations that were erased during the reassembly process. 15 1 5 10 2 3 5 6 2 3 1 4 7 4 1 –5 0 0.5 1 2 3 Activity change (fold) Stability change (ºC) 3 2 Fig. 1. Schematic map of the activity and thermal stability during the evolutionary engineering of S. toebii TPL. Open symbols ( ) indicate activity-related mutants and filled symbols ( ) indicate stability mutants obtained during the screening of random muta- genesis libraries. Gray symbols ( ) indicate reassembled hits obtained during subsequent shuffling experiments, where the full and broken arrows show the trajectory of stability and activity mutations, respectively. The gray-shaded ellipse encircles S1–S3 and AS1–AS3 hits that only include stability mutations. E. Rha et al. Directed evolution of tyrosine phenol-lyase FEBS Journal 276 (2009) 6187–6194 ª 2009 The Authors Journal compilation ª 2009 FEBS 6189 beneficial mutations accumulate, whereas deleterious mutations are simultaneously removed through the recombination of homologous DNA fragments [18,19]. As such, certain mutations tend to appear more often with the progression of the evolutionary engineering process. In this study, the A13V, E83K and T407A mutations originating from the S1–S3 hits were repeat- edly detected during the screening of the reassembly library (Table 1) and, of these mutations, A13V was detected most frequently, indicating its critical nature for stability. This mutation has already been shown to increase the thermal stability of the enzyme by 4 °C, with a slight compromise in enzyme activity [21]. Meanwhile, the four hits AS4–AS7 that exhibited a co-improvement of activity and stability included the additional T129I or T451A mutation (Table 1). Therefore, to understand the structure and function of the mutated residues, hypothetical structures of S. toebii TPL were generated by comparative model- ing using the open and closed structures of the TPLs from E. herbicola [1C7G in Protein Data Bank (PDB) entries] and C. freundii as templates [2,22–24]. The open and closed models were deposited in PMDB under accession numbers PM0075863 and PM0075934, respectively. When the open conformation was super- imposed on the open and closed templates (1C7G and 2VLH, respectively), the rmsd values were 0.37 and 1.97 A ˚ for the C a traces, respectively. The homology model consisted of four identical subunits, with PLP molecules near the catalytic lysine residue (K258). Each subunit comprised an N-terminal arm (M1–T21), a small domain (R22–S58, D312–I457) and a PLP-binding large domain (D59–V311). The active sites were located in clefts between the two domains, each constituting a catalytic dimer with the adjacent subunit. The two dimers were then tightly coupled via intertwined N-terminal arms near a hydrophobic cluster (M57–E70) at the center of the tetramer (Fig. 3A). When marking the stability mutations (A13V, E83K and T407A) in the three-dimensional model of S. toebii TPL, shown in Fig. 3A, they were all distributed around the dimer–dimer interface of the tetrameric assembly (red spheres). Meanwhile, the activity muta- tions (T129I and T451A) that survived the reassembly process were positioned far away from the dimer– dimer interface (cyan spheres). Interestingly, when the activity mutations that were deleted during the reas- sembly process (E42D, A126T and V262A) were marked in the three-dimensional model (Fig. 3), they were located near the dimer–dimer interface of the three-dimensional structure (gray spheres). To further investigate the stabilizing effect of Ala13, various N-terminal homologs were retrieved from the blast website and aligned, as shown in Table 2. The position corresponding to Ala13 was found to vary in sequence and identified as a serine in many known microbial species, including C. freundii and E. herbicola, yet was a branched amino acid (Val, Thr or Met) in b-tyrosinases (TPL) from anaerobic bacteria and 120 4 0 20 40 60 80 100 1 0 2 3 4 5 6 Temperature (ºC) 60 80 100 120 Temperature (ºC) p H Specific activity (units·mg –1 ) Specific activity (units·mg –1 ) Relative activity (%) Relative activity (%) 0 1 2 3 p H 40 50 60 70 80 40 50 60 70 80 6 7 8 9 10 11 5 6 7 8 9 10 11 0 20 40 AB CD Fig. 2. Effect of temperature and pH on stability and catalytic activity of purified TPL proteins. (A) Specific activity assayed at various temperatures in 100 m M potassium phosphate buffer (pH 8.0). (B) Relative activ- ity assayed to evaluate thermal stability. Enzymes were pre-incubated in 100 m M potassium phosphate buffer for 30 min at various temperatures. (C) Specific activity assayed in 50 m M potassium phos- phate ⁄ glycine buffer at the indicated pH values. (D) Relative activity assayed to evaluate pH stability. Enzymes were pre-incubated in 50 m M potassium phos- phate ⁄ glycine buffer for 36 h at ambient temperature. Filled circles indicate S. toebii TPL, and open circles and triangles indicate AS4 and AS7, respectively. Directed evolution of tyrosine phenol-lyase E. Rha et al. 6190 FEBS Journal 276 (2009) 6187–6194 ª 2009 The Authors Journal compilation ª 2009 FEBS l-tryptophanases. Notwithstanding, the flanking sequences of Ala13 were highly conserved and symmet- rical in each N-terminal homolog examined (Table 2). Together with this sequence observation, a magnified picture of the subunit interfaces (Fig. 4) revealed that the symmetric sequence consisted of two possible elec- trostatic interactions, K10–E15 * –K12, that flanked the contact between the Ala13 residues and the interacting subunits, where the asterisked E15 indicates the inter- acting residue contributed by the other intertwined subunit. Therefore, the stabilizing effect of the Ala13 to Val mutation could be related to a tighter assembly with the interacting subunits. In a previous study, the current authors have shown that the mutation of Thr15 to Ala in the vicinity of the hydrophobic core induces a tighter binding of the cofactor in C. freundii TPL, thereby reducing the decomposition rate of the cofactor by a Pictet–Spen- gler reaction [9]. In the S. toebii enzyme, Pro16 occu- pies the Thr15 position (Table 2, Fig. 4) and is located close to Ala13, which is very important for the stabil- ity of the enzyme in this study. In addition, the structural rationale for the other stability (E83K and T407A) and activity (T129I and T451A) mutations was investigated on the basis of the open and closed conformation models. As a result, the Table 2. Comparison of N-terminal sequences between homologous proteins and S. toebii TPL. Italic letters highlight residues correspond- ing to Ala13 in S. toebii TPL, and bold letters mark highly conserved residues in homologous N-termini. Strains (% identity) N-terminal sequence Source SC1(100) -MQRPWAEPYKIKAVEPIRMTT Symbiobacterium toebii gi:1805293 b-tyrosinase Sth (99) 1-MQRPWAEPYKIKAVEPIRMTT S. thermophilum gi:55977750 b-tyrosinase Fnu (67) 8-AEPFRIKSVETVKMID Fusobacterium nucleatum gi:27887209 b-tyrosinase Dno (66) 13-AEPFKIKSVEPVKMIS Dichelobacter nodosus gi:146233320 b-tyrosinase Cte (66) 11-AEPFKIKSVEPVKMIS Clostridium tetani gi:28202972 b-tyrosinase Pag (63) 5-AEPFRIKSVETVSMIS Pantoea agglomerans gi:260188 b-tyrosinase Cfr (63) 5-AEPFRIKSVETVSMIP Citrobacter freundii gi:401201 b-tyrosinase Ehe (63) 5-AEPFRIKSVETVSMIS Erwinia herbicola gi:1351283 b-tyrosinase Cag (69) 17-RRSWAEPWKIKTVEPLRIIS Chloroflexus aggregans gi:117996139 b-tyrosinase Dha (68) 6-AEPFRIKVVEPVRSMK Desulfitobacterium hafniense gi:109642283 b-tyrosinase Rca (70) 15-RRSWAEPWKIKMVEPLRVTT Roseiflexus castenholzii gi:156232199 b-tyrosinase Cau (68) 16-RRSWAEPWKIKMVEPLRVTS Chloroflexus aurantiacus gi:76167194 b-tyrosinase Tde (66) 8-AEPFRIKVVETVKMID Treponema denticola gi:42526628 b-tyrosinase Tte (58) 9-AEPYKIKMVEPLKITT T. tengcongensis gi:20808031 tryptophanase Dre (54) 5-QPKAEPFRIKMVEPIKMIS Desulfotomaculum reducens gi:134052564 tryptophanase Cno (55) 6-EPFKIKMVEPLTITT Clostridium novyi gi:118442982 tryptophanase Stt (54) 2-PKGEPFKIKMVEPIRLIP S. thermophilum gi:2842553 tryptophanase Pvu (49) 1-MAKRIVEPFRIKMVEKIRVPS Proteus vulgaris gi:2914379 tryptophanase AB Glu83 Thr129 Thr451 Thr407 Ala13 N Thr451 Thr407 Ala13 Val262 Glu42 Ala13 N Glu83 Thr129 Ala196 * Fig. 3. Structural assignments of stability- and activity-improving mutations in homo- logy model framework (A) and subunit structure (B) of S. toebii TPL. Red and cyan letters indicate stability and activity mutations, respectively, whereas black letters indicate the activity mutations deleted during the reassembly process. PLP was adopted from the 1C7G PDB file and is indicated by yellow sticks. E. Rha et al. Directed evolution of tyrosine phenol-lyase FEBS Journal 276 (2009) 6187–6194 ª 2009 The Authors Journal compilation ª 2009 FEBS 6191 T407A mutation in the small domain was correlated with the stability of the substrate-binding site, as T407 was in van der Waals’ contact with important substrate-binding residues, T50 and R405, in the same domain. Meanwhile, E83 was located on the surface of the large domain near the hydrophobic core, but no direct interaction with other residues was detected in this study. By contrast, the T129I and T451A muta- tions were located in the large and small domains, respectively, constituting the active site cleft between the domains (Fig. 3B). In the closed conformation, it has been proposed that the small domain undergoes an extraordinary motion towards the large domain, closing the active site cleft and bringing the cataly- tically important residues R382 and F449 into the active site [3,24]. Consequently, the modeling studies revealed that most of the stability mutations were located around the dimer–dimer interface, including the N-terminus. Meanwhile, the activity-improving mutations were found further away from the interface, as the activity- related mutations near the interface were seemingly deleted during the reassembly–screening steps, thereby allowing the directed co-evolution of the stability and catalytic activity of S. toebii TPL. Materials and methods Materials Sodium pyruvate and PLP were obtained from Musashino Shoji (Tokyo, Japan) and Fluka (Seelze, Germany), respec- tively, and yeast extract and bacto-casitone were purchased from BD (Franklin Lakes, NJ, USA). The other chemicals, including l-tyrosine, phenol and NH 4 Cl, were all purchased from Sigma-Aldrich (St Louis, MO, USA). The restriction endonucleases, T4 DNA ligase and Vent DNA polymerase were purchased from New England Biolabs (Beverly, MA, USA), and the Taq DNA polymerase was obtained from Takara (Otsu, Japan). The oligonucleotides were synthe- sized at Bioneer Co. (Daejeon, South Korea), and the DNA sequencing was performed by Solgent Co. (Daejeon, South Korea). Random mutagenesis and DNA shuffling The plasmid pHCE IIB-TPL, harboring the S. toebii TPL gene [7], was used as template for an error-prone PCR employing a Genemorph II Random Mutagenesis kit (Stratagene, La Jolla, CA, USA) with the following primers: 5¢-CTCAAGACCCGTTTAGAGGCCC-3¢ (forward) and 5¢-ATGCGTCCGGCGTAGAGGAT-3¢ (reverse). Thermal cycling was performed using a DNA Thermal Cycler (Bio-Rad, Hercules, CA, USA). The amplified PCR products were digested with NdeI and HindIII to yield a 1.377 kb DNA fragment. The plasmid pHCE IIB (Biolead- ers, Daejeon, South Korea) was also digested with NdeI and HindIII, and dephosphorylated with shrimp alkaline phosphatase (Roche, Mannheim, Germany). The plasmid and insert were then ligated at 4 °C for 24 h with a T4 DNA ligase, following which the products were electro- transformed into E. coli JM83 (ATCC #35607) and spread on LB–ampicillin plates. After incubation overnight at 37 °C, the ampicillin-resistant colonies were transferred to fresh LB–ampicillin plates using toothpicks. A mixture of plasmids selected from the random muta- genesis library was utilized as the DNA template for StEP PCR [16,25] with Vent DNA polymerase and the following primers: 5¢-AACGGCGGCATGTCTTTCTATA-3¢ (for- ward) and 5¢-ATGCCTGGCAGTTCCCTACTCT-3¢ (reverse). PCR was carried out at 95 °C for 5 min, followed by 40 cycles of 30 s at 94 °C and 5 s at 55 °C, plus an addi- tional 15 cycles of 30 s at 95 °C, 30 s at 55 °C and 1.5 min at 72 °C. The amplified DNA was then cloned into pHCE IIB and transformed into E. coli JM83 cells. Expression and screening of mutant library Escherichia coli JM83 cells harboring the TPL library within the constitutive expression system pHCE IIB were inoculated manually using toothpicks into 96 deep-well plates containing an LB–ampicillin medium (500 lL) and cultivated in a well plate culture system (Bioneer Co., Daejeon, South Korea) for 20 h at 37 °C. Protein expres- sion from the constitutive expression system did not require the addition of an inducer [26]. The cultivated cells (450 lL) were centrifuged for 20 min using a well plate cen- trifuge (Hanil Sci., Incheon, South Korea), washed in 50 mm Tris ⁄ HCl buffer (pH 8.0) and treated with 200 lL P16 A13 A13 K10 K12 E15 V14 P16 K12 E15 V14 [39–49] N–TER Hydrophobic Core [57–69] [310–322] [411–420] K10 K12 Fig. 4. Structural symmetry and putative electrostatic interactions within intertwined N-terminal arm. Broken lines represent electro- static interactions in the vicinity of A13 in the homology model of S. toebii TPL. Directed evolution of tyrosine phenol-lyase E. Rha et al. 6192 FEBS Journal 276 (2009) 6187–6194 ª 2009 The Authors Journal compilation ª 2009 FEBS Cellytic BÔ (Sigma-Aldrich) for 30 min at 37 °C. The cell lysate (100 lL) was then transferred using a multichannel pipette (Eppendorf, Hamburg, Germany) into 96-well PCR plates and mixed with an equal volume of a substrate solu- tion (described in the assay conditions below). After incu- bation at 37 °C, the reaction solutions were heated for 3 min at 94 °C, centrifuged to remove any insoluble aggre- gates and analyzed for phenol production. Following the activity analysis, the remaining cells with positive hits (50 lL) were transferred to fresh LB–ampicillin medium and preserved as glycerol stocks at )20 °C. Purification and characterization The mutant cells were cultivated at 37 °C for 24 h in 200 mL of LB medium containing 100 lgÆmL )1 of ampicil- lin. The cells were then suspended in 5 mL of NaCl ⁄ P i , mixed with an equal volume of Cellytic BÔ, treated for 30 min with 20 lL of DNase (10 UÆmL )1 ; Roche) and 100 lL of lysozyme (100 mgÆmL )1 ; Sigma, St Louis, MO, USA) and sonicated. Thereafter, the solution was centri- fuged at 24 000 g for 30 min, and the supernatant was incu- bated at 48 °C for 30 min to precipitate heat-labile E. coli proteins. The solution was then centrifuged again, loaded onto a Resource Q ion exchange column (Pharmacia, Uppsala, Sweden), washed with a standard buffer and eluted using a 0–0.5 m KCl gradient. Most of the active fractions were pooled, adjusted to include 1.7 m (NH 4 ) 2 SO 4 , loaded onto a hydrophobic Phenyl Superose column (Pharmacia) and eluted using a reverse gradient of (NH 4 ) 2 SO 4 from 1.7 to 0 m. The active fractions were then dialyzed against 100 mm Tris ⁄ HCl (pH 8.0) and stored at 4 °C. All the column procedures were carried out using an AKTA system (Amersham Bioscience, Uppsala, Sweden) at room temperature. Homology modeling and structural analysis To examine the structural and functional effects of the detected mutations, three-dimensional models were gener- ated of both the wild-type and mutant S. toebii TPL. The wild-type model of S. toebii was produced using ProModII and optimized using Gromos96 from SWISS-MODEL [27], an automated comparative protein modeling server. The tetrameric structure of TPL from E. herbicola (1C7G in PDB entry), sharing a 63% sequence identity with S. toebii TPL, was used as the template for comparative modeling. The coenzyme PLP was adopted from the 1C7G PDB file and fitted into the PLP-binding site of each monomer. Meanwhile, an open conformation model and mutant models of S. toebii TPL were constructed using the Build Model module from Discovery Studio (Accelrys, San Diego, CA, USA). The model with the best loop conformations was then selected using the Profiles-3- D verification method, and the structure was optimized on the basis of energy minimization in the DS CHARM module employing the steepest descent method followed by the conjugate gradient method. During the minimiza- tion process, the protein backbone was restrained using an harmonic constraint. Database Model data are available in the PMDB database under the accession numbers PM0075863, PM0075934, PM0075854 and PM0075847. Enzyme assay The enzyme activity in solution was measured by incubat- ing 5 lg of the enzyme with 1 mml-tyrosine and 10 lm PLP in 50 mm Tris ⁄ HCl buffer (200 lL, pH 8) for 15 min at 37 °C. After heating at 94 °C for 3 min, the amount of phenol in solution was measured colorimetrically using a microplate reader (Bio-Rad) based on the 4-aminoantipyrin method [28]. One unit of enzyme was defined as the amount of enzyme able to catalyze the formation of 1 l mol of phenol in 1 min at 37 °C. The protein concentration was determined using the Bradford assay (Bio-Rad) with BSA as the standard. Acknowledgements This project was supported by the Bio R&D program and the pioneer research program through the KOSEF (Korea Science and Engineering Foundation) and a grant from the KRIBB Research Initiative Program. References 1 Kumagai H, Yamada H, Matsui H, Ohkishi H & Ogata K (1970) Tyrosine phenol lyase. I. Purification, crystalli- zation, and properties. J Biol Chem 245, 1767–1772. 2 Phillips RS, Demidkina TV & Faleev NG (2003) Struc- ture and mechanism of tryptophan indole-lyase and tyro- sine phenol-lyase. Biochim Biophys Acta 1647, 167–172. 3 Milic ´ D, Matkovic ´ -C ˇ alogovic ´ D, Demidkina TV, Kuli- kova VV, Sinitzina NI & Antson AA (2006) Structures of apo- and holo-tyrosine phenol-lyase reveal a catalytically critical closed conformation and suggest a mechanism for activation by K + ions. Biochemistry 45, 7544–7552. 4 Yamada H & Kumagai H (1975) Synthesis of l-tyrosine related amino acids by b-tyrosinase. Adv Appl Microbiol 19, 249–288. 5 Gelenberg AJ & Gibson CJ (1984) Tyrosine for the treatment of depression. Nutr Health 3, 163–173. 6Lu ¨ tke-Eversloh T, Santos CN & Stephanopoulos G (2007) Perspectives of biotechnological production of E. Rha et al. Directed evolution of tyrosine phenol-lyase FEBS Journal 276 (2009) 6187–6194 ª 2009 The Authors Journal compilation ª 2009 FEBS 6193 l-tyrosine and its applications. Appl Microbiol Biotechnol 77, 751–762. 7 Lee SG, Hong SP, Choi YH, Chung YJ & Sung MH (1997) Thermostable tyrosine phenol-lyase of Symbio- bacterium sp. SC-1: gene cloning, sequence determina- tion, and overproduction in Escherichia coli. Protein Expr Purif 11, 263–270. 8 Suzuki S, Horinouchi S & Beppu T (1988) Growth of a tryptophanase-producing thermophile, Symbiobacterium thermophilum gen. nov., sp. nov., is dependent on cocul- ture with a Bacillus sp. J Gen Microbiol 134, 2353–2362. 9 Lee SG, Hong SP, Kim do Y, Song JJ, Ro HS & Sung MH (2006) Inactivation of tyrosine phenol-lyase by Pictet–Spengler reaction and alleviation by T15A muta- tion on intertwined N-terminal arm. FEBS J 273, 5564– 5573. 10 Burton SG, Cowan DA & Woodley JM (2002) The search for the ideal biocatalyst. Nat Biotechnol 20, 37–45. 11 Kim DY, Rha E, Choi SL, Hong SP, Sung MH & Lee SG (2007) Development of bioreactor system for l-tyro- sine synthesis using thermostable tyrosine phenol-lyase. J Microbiol Biotechnol 17, 116–122. 12 Lee SG, Hong SP, Song JJ, Kim SJ, Kwak MS & Sung MH (2006) Functional and structural characterization of thermostable d-amino acid aminotransferases from Geobacillus spp. Appl Environ Microbiol 72, 1588–1594. 13 Hoseki J, Okamoto A, Takada N, Suenaga A, Futatsugi N, Konagaya A, Taiji M, Yano T, Kuramitsu S & Kagamiyama H (2003) Increased rigidity of domain structures enhances the stability of a mutant enzyme created by directed evolution. Biochemistry 42, 14469– 14475. 14 Song JK & Rhee JS (2000) Simultaneous enhancement of thermostability and catalytic activity of phospholi- pase A(1) by evolutionary molecular engineering. Appl Environ Microbiol 66, 890–894. 15 Bae E & Phillips GN Jr (2006) Roles of static and dynamic domains in stability and catalysis of adenylate kinase. Proc Natl Acad Sci USA 103, 2132–2137. 16 Zhao H, Giver L, Shao Z, Affholter JA & Arnold FH (1998) Molecular evolution by staggered extension process (StEP) in vitro recombination. Nat Biotechnol 16, 258–261. 17 Moore JC, Jin HM, Kuchner O & Arnold FH (1997) Strategies for the in vitro evolution of protein function: enzyme evolution by random recombination of improved sequences. J Mol Biol 272, 336–347. 18 Bloom JD, Meyer MM, Meinhold P, Otey CR, MacMillan D & Arnold FH (2005) Evolving strategies for enzyme engineering. Curr Opin Struct Biol 15, 447– 452. 19 Voigt CA, Mayo SL, Arnold FH & Wang ZG (2001) Computational method to reduce the search space for directed protein evolution. Proc Natl Acad Sci USA 98, 3778–3783. 20 Olsen M, Iverson B & Georgiou G (2000) High- throughput screening of enzyme libraries. Curr Opin Biotechnol 11, 331–337. 21 Kim J-H, Song JJ, Kim B-G, Sung MH & Lee SC (2004) Enhanced stability of tyrosine phenol-lyase from Symbiobacterium toebii by DNA shuffling. J Microbiol Biotechnol 14, 153–157. 22 Antson AA, Demidkina TV, Gollnick P, Dauter Z, von Tersch RL, Long J, Berezhnoy SN, Phillips RS, Haru- tyunyan EH & Wilson KS (1993) Three-dimensional structure of tyrosine phenol-lyase. Biochemistry 32, 4195–4206. 23 Sundararaju B, Antson AA, Phillips RS, Demidkina TV, Barbolina MV, Gollnick P, Dodson GG & Wilson KS (1997) The crystal structure of Citrobacter freundii tyrosine phenol-lyase complexed with 3-(4¢-hydroxyphe- nyl)propionic acid, together with site-directed mutagen- esis and kinetic analysis, demonstrates that arginine 381 is required for substrate specificity. Biochemistry 36, 6502–6510. 24 Milic ´ D, Demidkina TV, Faleev NG, Matkovic ´ -C ˇ alog- ovic ´ D & Antson AA (2008) Insights into the catalytic mechanism of tyrosine phenol-lyase from X-ray struc- tures of quinonoid intermediates. J Biol Chem 283, 29206–29214. 25 Zhao H & Arnold FH (1997) Optimization of DNA shuffling for high fidelity recombination. Nucleic Acids Res 25, 1307–1308. 26 Poo H, Song JJ, Hong SP, Choi YH, Yun SW, Kim JH, Lee SC, Lee SG & Sung MH (2002) Novel high-level con- stitutive expression system, pHCE vector, for a conve- nient and cost-effective soluble production of human tumor necrosis factor-a. Biotechnol Lett 24, 1185–1189. 27 Schwede T, Kopp J, Guex N & Peitsch MC (2003) SWISS-MODEL: an automated protein homology- modeling server. Nucleic Acids Res 31, 3381–3385. 28 Otey CR & Joern JM (2003) High throughput screen for aromatic hydroxylation. In Directed Enzyme Evolution: Screening and Selection Methods (Arnold FH & Georgiou G, eds), pp. 141–148. Humana Press, Totowa, NJ. Directed evolution of tyrosine phenol-lyase E. Rha et al. 6194 FEBS Journal 276 (2009) 6187–6194 ª 2009 The Authors Journal compilation ª 2009 FEBS . Simultaneous improvement of catalytic activity and thermal stability of tyrosine phenol-lyase by directed evolution Eugene Rha 1 ,. the stability and activity mutants, and demon- strated a significant improvement in both stability and catalytic activity (Fig. 1). As such, the activity and