CHEMBIOCHEM FULL PAPERS DOI: 10.1002/cbic.201300433

Protein Engineering of the Antitumor Enzyme PpADI for Improved Thermal Resistance Leilei Zhu, Feng Cheng, Victoria Piatkowski, and Ulrich Schwaneberg*[a] Arginine deiminase (ADI, EC 3.5.3.6) is a potential antitumor drug for the treatment of arginine-auxotrophic tumors such as hepatocellular carcinomas (HCCs) and melanomas. Studies in human lymphatic leukemia cell lines have confirmed the antiangiogenic activity of ADI. Activity and thermal resistance limit the efficacy of ADI in treatment of auxotrophic tumors. Previously, we reengineered ADI from Pseudomonas plecoglossicida (PpADI) for improved activity under physiological conditions (37 8C, PBS buffer, pH 7.4) by two rounds of directed evolution and combination of beneficial substitutions through site-directed mutagenesis. The best variant, PpADI M6 (K5T/D38H/D44E/ A128T/E296K/H404R), showed a 64.7-fold improvement in kcat value and a 37.6 % decreased S0.5 value under physiological

conditions. However, M6 lost rapidly its activity (half-life of ~ 2 days at 37 8C). Here we report the re-engineering of PpADI M6 for improved thermal resistance by directed evolution in order to increase its half-life under physiological conditions. Directed evolution and recombination of the two most beneficial positions yielded variant PpADI M9 (K5T/D38H/D44E/A128T/ V140L/E296K/F325L/H404R), for which the Tm value increased from 47 (M6) to 54 8C (M9); this corresponds to an increased half-life from ~ 2 days (M6) to ~ 3.5 days (M9) under physiological conditions. Structure analysis of the homology model of M9 showed that the beneficial substitutions V140L and F325L likely promote the formation of tetrameric PpADI, which has greater thermal resistance than dimeric PpADI.

Introduction ADI catalyzes the conversion of arginine to citrulline and ammonia and exerts antitumor activity by depletion of the nonessential amino acid arginine from human blood and other extracellular fluids. ADI-based arginine depletion is a specific therapy for arginine-auxotrophic tumors (e.g., hepatocellular carcinomas [HCC] and melanomas), and PEGylated ADI is currently under clinical trials for HCC treatment with promising results.[1] HCC is the third leading cause of cancer mortality worldwide, with a five-year survival rate of less than 10 % in the EU and USA)[2] due to low responses of hepatoma to chemotherapeutic treatments. In 1999, the US Food and Drug Administration (FDA) designated ADI-PEG-20 as an orphan drug for treating HCC and invasive malignant melanomas. In 2005, the European Agency for the Evaluation of Medicinal Products (EMEA) also granted ADI-PEG-20 orphan drug status for the treatment of HCC. Phase II trials of ADI-PEG-20 in patients with metastatic HCC have shown that the drug is safe, well tolerated, and could benefit patients with unresectable HCC. Izzo and coworkers[1] proposed that combination therapies based on arginine deprivation and growth inhibition through kinase inhibitors might further improve the survival rate. Recent studies showed that the combination of taxane and ADI-PEG20, which induces caspase-independent apoptosis, is more effective than a taxane monotherapy for prostate cancer. Additionally, ADI [a] Dr. L. Zhu, F. Cheng, V. Piatkowski, Prof. Dr. U. Schwaneberg Lehrstuhl fr Biotechnologie, RWTH Aachen University Worringerweg 1, 52056 Aachen (Germany) E-mail: [email protected] Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cbic.201300433.

 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

can enhance radio sensitivity of human mammary adenocarcinoma (MCF-7) cells,[3] serve as a potential therapeutic agent for inflammatory bowel disease (IBD)[4] and glioblastoma multiforme,[5] exert anti-proliferative and anti-angiogenic activities, and inhibit the growth of viruses such as HIV-1 and hepatitis.[6] Current research efforts are focused on the in vivo inhibitory effect of ADIs towards leukemia, melanoma, prostate cancer, renal cell carcinoma, and human umbilical vein endothelial cells (HUVEC),[7] as well as clinical trials focusing on HCC (Phase II),[1] melanoma (Phase I/II),[8] and inhibitory mechanisms of ADI on tumors.[9] ADIs catalyze the first step of the arginine dehydrolase pathway by hydrolyzing arginine to citrulline and ammonium. Arginine deprivation is believed to be a main molecular reason for the inhibitory effects of ADI on arginine auxotrophic tumors that do not express argininosuccinate synthetase (ASS). ADIs have been identified, purified, and characterized from bacteria, Archaea, and some eukaryotes, excluding mammalian cells.[10] Enzymatic properties of several recombinant ADIs have been examined, such as specific activity, functional expression level, optimal temperature, optimal pH, substrate affinity, and halflife in human plasma. Only the ADI from Mycoplasma hominis has been developed thus far into an orphan drug for HCC and melanoma (Phoenix Pharmacologics, Inc.).[11] High activity and thermal resistance at 37 8C and at physiological low arginine concentrations is an important prerequisite for the success of ADI in medical applications and to reduce dose requirements per treatment. We previously reported that after two rounds of directed evolution, the specific activity of PpADI was significantly imChemBioChem 2014, 15, 276 – 283

276

CHEMBIOCHEM FULL PAPERS proved, especially at low arginine concentrations.[12, 21] The best variant with high activity at low arginine concentrations, PpADI M6 (K5T/D38H/D44E/A128T/E296K/H404R), showed a 64.7-fold improved kcat value and 37.6 % decreased S0.5 value (ligand concentration occupying half of the binding sites) under physiological conditions (pH 7.4, 37 8C, PBS buffer). Until now, thermal resistance of ADIs has not been systematically studied, and no ADI has been improved by protein engineering with regard to thermal resistance. For clinical applications, enhanced thermal resistance is often favorable in order to extend the shelf life of therapeutic proteins in reagents and to reduce dosages through prolonged lifetimes. For instance, fructosyl peptide oxidase[13] and kanamycin nucleotidyltransferase variants[14] revealed an increased thermal resistance as well as an increased resistance to proteolytic digestion. The correlation between thermal resistance and prolonged lifetime in vivo might be a general principle. As state-of-the-art PEGylation is commonly used to prolong the circulation half-life of therapeutic proteins such as ADI and to decrease immunogenicity. PEGylated ADI has a 50 % reduced specific activity, whereas the circulation half-life increases from 5 h to 6 days in mice.[15] Protein engineering by rational design and directed evolution offers opportunities to tailor enzymatic properties to meet therapeutic requirements in iterative rounds of diversity generation and screening. In case of the antitumor enzyme, ADI, two reports were published in which the optimal pH was shifted by 0.5 unit to pH 7.0, the kcat value improved 64.7-fold, and the S0.5 value decreased by 37.6 % under physiological conditions.[12, 21] There has been no report in which improving the thermal resistance of any ADI prolonged the circulation halflife in vivo to complement a PEGylation strategy. In this manuscript, we report the first directed ADI evolution experiment for improved thermal resistance of ADI. Based on the reengineered PpADI M6 variant, a directed evolution campaign employing the citrulline detection system in microtiter plate format yielded PpADI variant M9 with significantly improved thermal resistance after a single round of directed evolution and subsequent site-directed mutagenesis.

Results Improved citrulline detection assay in 96-well microtiter plate format To identify PpADI variants with improved thermal resistance at low arginine concentrations, the previously reported colorimetric screening protocol for citrulline detection[12, 21] in 96-well microtiter plate format was employed with modified substrate concentrations, reaction time, and color development. An emphasis was put on advancing the citrulline quantification system to measure citrulline concentrations < 0.8 mm, which was achieved by adjusting the incubation time for color development (elongated from 15 to 30 min) and measuring absorbance at 492 nm (see Figure 1 A). Reproducible citrulline quantification at concentrations < 0.5 mm (Figure 1 A) were achieved with a 30 min incubation for color development, conditions that were subsequently employed in the screening of an  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chembiochem.org

Figure 1. A) Calibration curve of the citrulline colorimetric detection system in 96-well microtiter plate format. Incubation time for color development: 15 min (^), 20 min (&), 25 min (~), and 30 min (*). B) Activity values of a 96well plate assay with PpADI parent M6 in descending order of arginine conversion, which was quantified through citrulline formation by using the colorimetric citrulline detection system. The apparent coefficient of variation was calculated without subtracting the background, and the true coefficient of variation was calculated after background subtraction.

epPCR mutant library. In addition, the permeabilizer cetyltrimethylammoniumbromide (CTAB) was replaced by lysozyme, as CTAB inhibits PpADI. As the activity of PpADI M6 is significantly improved (> 64.7-fold) compared to PpADI WT, the volume of cell lysate used for microtiter plate screening was reduced to 30 % with a PpADI conversion time of 3 min. The coefficient of variation is an important performance criterion of a screening system to determine its accuracy. The optimized citrulline detection assay has a true coefficient of variation of 16.8 % (Figure 1 B). Lysozyme treatment of cell pellets involved freezing, thawing, and centrifugation steps which contribute to this true coefficient of variation. Microtiter plate screens with a true coefficient of variation ranging between 8 and 20 % have been successfully employed in directed enzyme evolution experiments.[16] EpPCR library construction and screening of PpADI variants with improved thermal resistance M6 harbors six substitutions (K5T/D38H/D44E/A128T/E296K/ H404R) and was used as a parent for epPCR library generation. Two concentrations of MnCl2 (0.08 and 0.15 mm) were used in epPCR library generation and resulted in 66 % to 76 % inactive variants, respectively. One to four nucleotide mutations per gene were observed. Five hundred clones from each epPCR library were screened with the modified citrulline quantification ChemBioChem 2014, 15, 276 – 283

277

CHEMBIOCHEM FULL PAPERS

www.chembiochem.org

assay. The coverage of the library was ~ 16 %, estimated by using PEDEL[17] (program for estimating diversity in error-prone PCR libraries, http://guinevere.otago.ac.nz/cgi-bin/aef/pedel.pl). PpADI conversions were performed with fresh cell lysate and heat-treated cell lysate (55 8C for 30 min) in order to eliminate expression mutants. The ratio of the activity of heat-incubated cell lysate versus the activity of nonincubated cell lysate is termed residual activity and was used to select variants with improved thermal resistance and preserved activity. Two variants with significantly improved thermal resistance, M7 (K5T/ D38H/D44E/A128T/E296K/F325L/H404R) and M8 (K5T/D38H/ D44E/A128T/V140L/E296K/H404R), were identified and carried one additional substitution. Furthermore, M9 (K5T/D38H/D44E/ A128T/V140L/E296K/F325L/H404R) was generated by combing both beneficial substitutions (V140L was introduced in M7). M9 harbors both substitutions (F325L and V140L), which contribute in an additive manner to the thermal resistance of M9, whereas the kcat value is comparable to the M6 parent (Table 2).

Figure 2. Thermal inactivation profiles of PpADI WT and variants M6–M9. Residual activities of PpADI WT (*), parent M6 (&), and variants M7 (~), M8 (^), and M9 (*) were measured after heat treatment (4 h, pH 7.4, 50 mm potassium phosphate buffer) at varied temperatures (37–56 8C).

Table 1. Percentage of a-helical content of PpADI variants M6–M9 after heat inactivation (4 h, pH 7.4, 50 mm potassium phosphate buffer) at varied temperatures.

Variants

RT

37 8C

M6 M7 M8 M9

55.5 60.3 51.6 60.2

49.5 45.8 46.6 56.2

a-helical content [%] 42 8C 47 8C 43.9 45.7 45.0 50.3

28.8 36.6 36.8 47.0

55 8C 8.3 11.9 18.3 18.1

Thermal resistance profile of parent M6 and improved variants The thermal resistance of purified PpADI WT, parent M6, and variants M7, M8, and M9 were assessed by measuring their residual activity after heat treatment for 4 h at varied temperatures (see Figure 2). The half-inactivation temperatures (Tm) were found to be 42 8C (PpADI WT), 47 8C (M6), 50 8C (M7), 52 8C (M8), and 54 8C (M9), respectively. The most thermally resistant variant, M9, has a Tm value that is 7 8C higher than parent M6 and 12 8C higher than PpADI WT.

Figure 3. Thermal inactivation at 37 8C (pH 7.4; 50 mm potassium phosphate buffer) as determined through residual activity measurements of PpADI WT (*), parent PpADI M6 (&), and M9 (*) after 1–6 days of incubation.

CD measurements of parent M6 and variants (M7, M8, and M9) after heat incubation at varied temperatures The secondary structural integrity of PpADI variants was investigated by comparing CD spectra of each variant at room temperature and after heat incubation (4 h) at varied temperatures (room temperature, 37, 42, and 47 8C; Table 1). The a-helical content in PpADI variants M6–M9 was monitored and calculated with the K2D3 software (http://www.ogic.ca/projects/k2d3/ )[25] at 222 nm. With increased temperature, the absorbance of

Long-term measurements to determine thermal resistance of PpADI WT, parent M6, and M9 The heat-inactivation profiles of PpADI WT, parent M6, and the thermally resistant variant M9 are given in Figure 3. Significant differences in activity retention were observed after 2 days of incubation at 37 8C. M9 retained ~ 90 % of its activity, whereas parent M6 and PpADI WT only retained ~ 40 and ~ 20 % of their activity, respectively. After 4 days of incubation at 37 8C, M9 still retained ~ 40 % of its activity, whereas parent M6 and PpADI WT lost more than 90 % of their activity. A comparison of half-lives at 37 8C showed that M9 has a half-life of ~ 3.5 days compared to ~ 2 day half-life of parent M6 and the ~ 1 day half-life of PpADI WT.  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Table 2. Kinetic constants (kcat and S0.5) of PpADI M6 (parent) and variants M7, M8, and M9, which were calculated based on the sigmoidal model[a] by using the Hill equation.[12]

1

kcat [s ] S0.5 [mm] Hill coefficient

M6 (parent)

M7

M8

M9

3.90  2.09 1.04  0.10 1.45  0.17

15.13  2.27 1.88  0.18 1.74  0.17

7.49  1.12 2.18  0.51 1.21  0.15

13.68  2.05 1.21  0.09 1.83  0.18

[a] Initial velocity data were fitted to the sigmoidal model of the kinetics equation v = vmax  Kn/(S0.5n+Kn), where v is the initial velocity, vmax is the maximum velocity, K is the substrate concentration, and S0.5 is the ligand concentration at which half of the active sites are occupied (concentration of half saturation).[19]

ChemBioChem 2014, 15, 276 – 283

278

CHEMBIOCHEM FULL PAPERS the CD spectra (from 190 to 240 nm) was reduced, thus indicating at least a partial unfolding of a-helical structures. CD spectra (see Figure S1 in the Supporting Information) and an a-helical content of PpADI variants at room temperature (51.6– 60.3 %) and at 55 8C (8.3–18.1 %; see Table 1) differ significantly. As a general trend, one could observe that the loss of a-helical content became more dramatic at elevated temperatures, reaching particularly large differences at 47 8C (M6: 28.8 %; M9: 47.0 %). Overall M7, M8, and M9 at 47 and 52 8C showed, as expected, a lower loss in a-helical content than parent M6 (Table 1). Determination of kinetic constants (kcat and S0.5) of PpADI M6 (parent), M7, M8, and M9 PpADI variants M7, M8, M9, and parent M6 were subject to detailed kinetic characterization at 37 8C in PBS buffer (pH 7.4, 137 mm NaCl, 2.7 mm KCl, 10 mm Na2PO4, 2 mm KH2PO4 ; regarded as isotonic[18]). PBS buffer maintains a constant pH (7.4), and its osmolarity and ion concentrations mimic physiological conditions. Table 2 summarizes the kcat and S0.5 values of parent M6 (K5T/D38H/D44E/A128T/E296K/H404R) and variants M7 (K5T/D38H/D44E/A128T/E296K/F325L/H404R), M8 (K5T/ D38H/D44E/A128T/V140L/E296K/H404R), and M9 (K5T/D38H/ D44E/A128T/V140L/E296K/F325L/H404R), which were expressed in shaking flask cultures and purified, as previously reported, by using a two-step procedure (anion exchange followed by gel filtration).[21] The S0.5 value for arginine conversion increased for variants M7 (S0.5 = 1.88 mm) and M8 (S0.5 = 2.18 mm) as compared to parent M6 (S0.5 = 1.04 mm). The S0.5 value of M9 (1.21 mm) was similar to that of M6. The kcat values of M7 (15.13 s 1) and M9 (13.68 s 1) were comparable to M6 (13.90 s 1), whereas M8 showed a 47 % decrease in kcat value (7.49 s 1).

Discussion ADI is a promising drug for the treatment of HCC and melanoma[11] that has orphan drug status.[11] A main challenge for effective treatment is that ADI is highly active and stable under physiological conditions (e.g., 100–120 mm arginine; pH 7.4). Thus far, only three protein engineering campaigns by directed evolution have been reported for ADIs[12, 20, 21] in which the PpADI was successfully improved by boosting activity (64.7fold[12] at pH 7.4), shifting the optimum pH towards the physiological pH (from 6.5 to 7.0[21]), and reducing S0.5 values (from 1.30 to 0.81 mm[21]). None of these directed evolution experiments was aimed at improving the thermal resistance of PpADI in order to prolong circulation half-life and minimize dosage as well as immunogenicity response. Directed evolution has successfully been used to improve the thermal resistance of therapeutic enzymes (e.g., asparaginase[22] and cocaine esterase[23]), and highly thermally resistant enzymes often have improved plasma half-lives.[23] Reports on efforts to improve the in vivo circulation half-life by combined protein engineering and a PEGylation strategy have not been reported, to the best of our knowledge.  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chembiochem.org In order to maintain the high activity of PpADI M6 (K5T/ D38H/D44E/A128T/E296K/H404R) and to improve its thermal resistance, a third round of directed PpADI evolution with M6 as the parent was performed and yielded the thermally improved variants M7 (K5T/D38H/D44E/A128T/E296K/F325L/ H404R) and M8 (K5T/D38H/D44E/A128T/V140L/E296K/H404R) with one additional substitution each (F325L or V140L). Positions V40 and F325 have not been reported to influence the thermal resistance of PpADI or any other ADIs. The two substitutions F325L and V140L were combined by using site-directed mutagenesis to yield the variant M9 (K5T/D38H/D44E/A128T/ V140L/E296K/F325L/H404R). The thermal resistance of M9 significantly exceeds the thermal resistance of M7 and M8 (Figure 2). In detail, the Tm values of M9 increased by 12 8C (compared to PpADI WT), 7 8C (compared to parent M6), 4 8C (compared to M7), and 2 8C (compared to M8). The half-life of M9 at 37 8C increased by 2.5 days (compared to PpADI WT), and 1.5 days (compared to PpADI M6). Unfortunately, there are no Tm values or half-lives reported for other purified ADIs to compare. The kcat value of M9 is comparable to the highly active variant M6 (13.68 vs. 13.90 s 1, respectively; WT: 0.18 s 1 [12]); this is one of the rare cases[24] in which thermal resistance and activity of an enzyme were improved simultaneously by three rounds of directed evolution. Notably, MhADI (ADI from M. hominis) commercialized by Phoenix Pharmacologic, Inc.[11] has a kcat value of 0.15 s 1, whereas the PpADI variant M9 has a kcat value approximately 91 times higher (13.68 s 1). To gain deeper insights into the structural changes of thermally resistant PpADI M9, a molecular model of M9 was constructed (Figure 4) based on crystal structures of PaADI, which has an 84.9 % sequence identity with PpADI (http://embnet.vital-it.ch/software/ClustalW.html). The overall fold of PaADI and PpADI consists of five bbab modules in a cyclical arrangement, generating a pseudo-fivefold symmetrical barrel, and an additional 85-residue a-helical domain is inserted between the first and second bbab modules. The 85-residue insertion includes five a-helices (a’-1–5, colored black in Figure 4), a short helix (3/10), and a short b-strand. The a-helical domains are at the dimer–dimer interface and govern tetramer formation.[25] CD measurements of PpADI M6–M9 showed that the a-helical content decreases dramatically after heat inactivation (Table 1) so that the tetrameric structure is most likely lost; the latter was confirmed by native polyacrylamide gels (Figure S2 in the Supporting Information). Figure 4 shows the tetrameric model of PpADI M9 with the four monomers indicated by the letters A, B, C, and D. The V140L substitution is located in a’-5, which belongs to the five a-helices at the dimer–dimer interface of PpADI (Figure 4). Substitution F325L is in the a-helix (residues 323–334), close to the dimer–dimer interface but not in direct contact with a neighboring subunit (Figure 4). Interestingly, the side chains of both positions (V140L and F325L) are engaged in hydrophobic interaction networks within the same monomer or with the neighboring monomer at the dimer interface. It is therefore likely that changes in the hydrophobic network are relayed between monomers at the dimer–dimer interface, which has been reported to be a main driving force for ChemBioChem 2014, 15, 276 – 283

279

CHEMBIOCHEM FULL PAPERS

www.chembiochem.org

Figure 4. Model of PpADI M9 generated by the YASARA software package, based on crystal structures of PaADI (2a9g, 2aaf, and 1rxx from Pseudomonas aeruginosa). Substitutions (V140L and F325L) are shown in CPK style. The tetrameric PaADI is built by two dimer subunits (dimer A–B in dark gray; dimer C–D in light gray). PaADI consists of four ADI molecules which are individually labeled as monomers A, B, C, and D. Monomers A and B form the first dimer, and monomers C and D form the second. The V140L substitution is located directly at the dimer–dimer interface, which is indicated by a dashed line.

monomer interactions.[26] Repositioning of the a-helical domain or fostering of dimer–dimer interactions might therefore contribute to enhanced thermal resistance. Calculation of DDG values (using FoldX) for dimer– dimer interactions caused by the V140L substitution showed a negative value, 2.02 kcal mol 1, which supports the stabilizing effect of V140L from our experimental results. However the FoldX calculation of the DDG values for F325L showed a positive value, 0.09 kcal mol 1, which is inconsistent with the experimental results. One possible explanation is that the calculation of FoldX is based on a fixed protein backbone and does not involve rearrangement, which might be the case for F325L. Figure 5 shows the interactions between the V140L substi-

tution and Val318 and Phe304 at the interface of monomer B and monomer C. Position 140 is located directly at the interface in helix a’-5 (residues 136–149), and the V140L substitution fosters hydrophobic interactions with Val318 and Phe304, which are within a distance of 5  from the substituted residue. This likely strengthens dimer–dimer interactions (Figure 5, left) and increases the thermal resistance of PpADI. Position 325 is located in the a-helix (residues 323–334, Figure 5) which is about 10  from the dimer–dimer interface of PpADI. F325L interacts with hydrophobic residues Ile266, Ala268, Ile301, and Pro303 in the neighboring b-strands (residues 263– 270 and residues 303–310). All hydrophobic interactions are intramolecular in each ADI monomer. The Hill coefficients of M7 (1.74) and M9 (1.83) were increased compared to parent M6 (1.45), possibly indicating a stronger cooperativity among the subunits due to stronger interactions among subunits of PpADI. However, this is not the case for M8, which has a slightly decreased Hill coefficient (1.21). In essence, both substitutions (V140L and F325L) enhance the thermal resistance of PpADI remarkably, and further in vivo studies are required to determine expected increases in circulation half-lives of PpADI variants M7–M9. Additionally, crystallization experiments are required to gain structural insights into interactions that govern dimer–dimer interface formation and stabilization of PpADI monomers to understand the complex interplay that determines activity and thermal resistance of PpADI WT and its variants (M6–M9).

Figure 5. Left: Structure model of variant M8 (K5T/D38H/D44E/A128T/V140L/E296K/H404R). V140L (in helix a’-5) is located in the a-helical domain which mediates tetramer formation. V140L is involved in inter-dimer interactions between monomer B (dark gray) and C (light gray) with Val318 and Phe304. Right: Structure model of variant M7 (K5T/D38H/D44E/A128T/E296K/F325L/H404R). The F325L substitution and the residues involved in the interaction with F325L are presented: F325L (blue), Ile301 (green), Ala268 (violet), Pro303 (yellow) and Ile266 (pink). The dashed line shows the interface between monomers B (dark gray) and C (light gray).

 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

ChemBioChem 2014, 15, 276 – 283

280

CHEMBIOCHEM FULL PAPERS Conclusions In summary, PpADI variants with significantly improved thermal resistance that retained specific activity were obtained for the first time by employing a modified citrulline quantification system in microtiter plate format. Two novel positions (Val140 and Phe325) were identified that can individually enhance the thermal resistance of PpADI significantly (3 8C for F325L and 5 8C for V140L). Structural analysis of the homology model of M9 indicates that thermal resistance is likely improved through hydrophobic interactions within a monomer (F325L) and at the dimer–dimer interface (V140L). The important role of dimer and tetramer formation for PpADI stability was confirmed by CD spectroscopy and native polyacrylamide gel analysis. This successful directed PpADI evolution proves that thermal resistance can be improved without reducing specific activity.

Experimental Section All chemicals were of analytical reagent grade or higher quality and were purchased from Fluka, Sigma–Aldrich, and Applichem (Darmstadt, Germany), except the resins for purification, which were purchased from TOSOH (Stuttgart, Germany). All enzymes were purchased from New England Biolabs, Fermentas, and Sigma–Aldrich, unless stated otherwise. A thermal cycler (Mastercyler gradient; Eppendorf) and thin-wall PCR tubes (Multi-ultra tubes; 0.2 mL; Carl Roth, Germany) were used for all PCR reactions. The PCR volume was always 50 mL except for Megaprimer PCR of the whole plasmid (MEGAWHOP, 25 mL); larger volumes were prepared in multiple 50 mL PCR reactions. The amount of DNA in cloning experiments was quantified by using a NanoDrop photometer (NanoDrop Technologies, Germany). Construction of PpADI error-prone (epPCR) library: An EpPCR library was generated by the standard epPCR method using improved variant M6 (K5T/D38H/D44E/A128T/E296K/H404R) as template. For the mutagenic PCR (95 8C for 2 min, 1 cycle; 95 8C, 30 s/ 55 8C, 30 s/72 8C, 30 s, 29 cycles; 72 8C for 3 min, 1 cycle), Taq DNA polymerase (2.5 U), dNTP mix (0.20 mm), template (50 ng, pET42b(+) harboring the PpADI M6 gene), and MnCl2 (0.08 and 0.15 mm), and primer (5’-TACATA TGTCCG CTGAAA CACAGA AG-3’ and 5’-GTGCTC GAGTTA GTAGTT GATCGG-3’, 10 pmol each) were used. The PCR products were purified by using a QIAquick PCR purification kit (Qiagen). The purified epPCR products were cloned into expression plasmid pET42b(+) by MEGAWHOP.[27] For MEGAWHOP (72 8C for 5 min, 1 cycle; 98 8C for 1 min 30 s, 1 cycle; 98 8C, 45 s/55 8C, 45 s/72 8C, 4 min, 24 cycles; 72 8C for 10 min, 1 cycle), PfuS DNA polymerase (1.2 U), dNTP mix (0.20 mm), and epPCR products (500 ng) together with template (50 ng, pET42b(+) harboring the PpADI gene of variant M3) were used. Following PCR, DpnI (40 U; New England Biolabs) was added, and the mixture was incubated overnight at 37 8C. The MEGAWHOP products were transformed into Escherichia coli BL21-Gold (DE3) cells for expression and screening. Site-directed mutagenesis for substitution V140L in PpADI M7: Site-directed mutagenesis for substitution V140L in PpADI M7 was performed according to the published method[28] by using plasmid (K5T/D38H/D44E/A128T/E296K/F325L/ pET42b(+)-PpADI-M7 H404R). The following oligonucleotides were used for mutagenesis of V140L: 5’-AGCGAA GGTGCC AGCCTG GTCAAG ATGTAC AAC-3’  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chembiochem.org and 5’-GTTGTA CATCTT GACCAG GCTGGC ACCTTC GCT-3’ (substituted nucleotide is underlined). Conditions for mutagenic PCR were as described previously.[21] The PCR products were purified by using a QIAquick PCR purification kit (Qiagen) and transformed into E. coli BL21-Gold (DE3) cells for expression. Cultivation and expression of PpADI in 96-well plates: Cultivation and expression in 96-well plates (V-bottom, polystyrene plates; Greiner Bio-One GmbH, Frickenhausen, Germany) were performed as previously described,[21] except that V-bottom plates were used. After expression, the cell culture was centrifuged in V-bottom 96-well plates with an Eppendorf centrifuge (4 8C, 3400 rpm, A-4-81 rotor, 15 min). The supernatant was discarded, and the pellet was frozen at 20 8C overnight. Improved screening system in 96-well plate to identify PpADI variants with improved thermal resistance: The previously employed screening assay based on citrulline detection[12] was further optimized for higher sensitivity to detect low concentrations of citrulline. Instead of 15 min, the reaction mixture was incubated for 30 min at 55 8C for color development. After freezing the cell pellet overnight, lysozyme solution (80 mL, 0.8 mg mL 1, in PBS buffer, lysozyme from chicken egg white, ~ 70 000 units mg 1, Fluka) was added to each well of the 96-well plates. The lysozyme solution was mixed homogenously with the cell pellet and then incubated at 37 8C for 30 min. After incubation, 140 mL PBS buffer was added to each well to dilute the cell lysate, followed by centrifugation at 4 8C and 3400 rpm, A-4-81 rotor for 15 min. A screening assay was performed for heat-treated (55 8C, 30 min) cell lysate and non-heattreated cell lysate. Diluted cell lysate (20 mL) was transferred into a 96-well microtiter plate. The enzyme reaction was initiated by the addition of arginine solution (100 mL, 1 mm arginine in PBS buffer, pH 7.4), and the mixture was incubated for 3 min at 37 8C. The subsequent color development procedure was carried out as described previously[21] except that 30 min was used for color development. Standard deviation measurements were performed in 96-well plate format by using lysed culture from BL21-Gold (DE3) cells lacking PpADI and, in a separate experiment, containing PpADI variant M6. Apparent standard deviation was based on the absolute absorbance values obtained from the PpADI variant M6 plate. The true standard deviation was calculated by subtracting the background absorbance value of (BL21Gold (DE3) lacking PpADI) from the apparent values. Expression of PpADI in shaking flask and purification: Expression of PpADI in shaking flasks and purification by anion exchange chromatography and gel filtration were performed as previously described.[21] Thermal resistance profile of parent M6 and improved variants: Thermal inactivation of PpADI M6 (parent) and variants M7, M8, and M9 were monitored by activity measurements after heat treatment. Purified PpADI M6 (parent) and variants M7, M8, and M9 in phosphate buffer (Na2HPO4 and NaH2PO4, 50 mm, pH 7.4), were incubated at a range of twelve temperatures from 37 to 60 8C for 4 h. Subsequently, the samples were assayed for residual activity. The Tm values were determined from the plots of relative inactivation [%] versus temperature [8C]. The Tm value is defined as the temperature at which 50 % of the initial enzyme activity is lost after heat treatment. Thermal resistance of parent M6 and improved variants at 37 8C: Purified PpADI M6 (parent) and variants M7, M8, and M9 in phosphate buffer (Na2HPO4 and NaH2PO4, 50 mm, pH 7.4) were inChemBioChem 2014, 15, 276 – 283

281

CHEMBIOCHEM FULL PAPERS cubated at physiological temperature (37 8C). The residual activities were determined at regular time intervals for up to seven days by using the citrulline detection assay. Structural integrity of PpADI M6 and variants at various temperatures by CD measurements: CD spectra (from 190 to 240 nm) were measured with an Olis CD Spectrophotometer model SDM 17 (On/Line Instrument Systems, Inc., Bogart, USA) under a constant nitrogen flow, with spectra recorded from 190–240 nm by using a 1 mm path-length cell and a bandwidth of 1 nm. Each spectrum shown is the average of three individual measurements. Raw collected CD data were converted into delta epsilon (De, molar circular dichroism). Purified PpADI M6 and variants M7, M8, and M9 (180 mg mL 1) were first incubated for 4 h at five temperatures: RT, 37.0, 42.6, 47.6, and 55.6 8C, and then subjected to CD measurement. Baseline measurements were determined in phosphate buffer (Na2HPO4 and NaH2PO4, 50 mm, pH7.4) and subtracted from the experimental data. CD data were analyzed by using K2D3 software (http://www.ogic.ca/projects/k2d3/)[29] to derive the a-helical content. Determination of kinetic constants of PpADI M6 (parent) and variants M7, M8, and M9: The kcat and S0.5 values were determined from initial velocity data measured as a function of substrate concentration. Enzyme reactions were carried out at 37 8C in a water bath, as described previously,[21] by using purified enzyme (50 mL, 0.65–0.75 mm) and substrate solution (200 mL, 0.5–6 mm of arginine, PBS buffer, pH 7.4) in deep-well plates. The initial velocity data obtained were fitted to the sigmoidal model of kinetics equation v = vmax Kn/(S0.5n+Kn) (where v is the initial velocity, vmax is the maximum velocity, K is the substrate concentration, and S0.5 is the ligand concentration producing half occupation)[19] by using GraphPad Prism 6. The kcat value was calculated from the ratio of vmax and enzyme concentration. Native polyacrylamide gel electrophoresis of PpADI M6, M7, M8, and M9: polyacrylamide native gels (10 %) were used for protein electrophoresis to analyze the native form of PpADI variants before and after incubation at 50 8C. Incubated and nonincubated purified protein samples of PpADI variants (7 mL of 180 ng mL 1 protein and 3 mL of loading dye) were loaded into each lane of a native gel. The molecular marker used in the gel was the NativeMark unstained protein standard (Invitrogen). Electrophoresis was performed with Tris/glycine buffer (pH 8.3), and the electrophoresis chamber was incubated in an ice-water bath. Protein bands were stained with Coomassie brilliant blue solution. Homology modeling: An homology modeling approach was used for the structure prediction of PpADI M6, M7, M8, and M9. The model was built by using YASARA 12.2.22.[30] The crystal structure of PaADI (PDB ID: 2a9g) was used as template by the server and has 84.9 % sequence identity with PpADI (http://blast.ncbi.nlm.nih.gov/Blast.cgi). All tetramers were energy minimized in a vacuum by using the AMBER03 force field.[31] FoldX analysis of PpADI variants M7 and M8: The changes in Gibbs free energy (DDG) induced by the mutations at positions 140 and 325 were calculated by FoldX (version 3.0 beta3).[32] The structure of the homology model of PpADI M6 was rotamerized and energy minimized by using the “RepairObject” command to correct the residues with nonstandard torsion angles. Then, single mutations were built by using the “mutate (multiple) residue” command, and the DG values were calculated by using the YASARA FoldX-plugin.[33] Temperature, pH, and ionic strength were assigned as 298 K, 7.0, and 0.05 m, respectively.  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chembiochem.org Acknowledgements The authors acknowledge financial support from RWTH Aachen University. Keywords: arginine deiminase · directed evolution · protein engineering · thermal resistance

[1] E. S. Glazer, M. Piccirillo, V. Albino, R. Di Giacomo, R. Palaia, A. A. Mastro, G. Beneduce, G. Castello, V. De Rosa, A. Petrillo, P. A. Ascierto, S. A. Curley, F. Izzo, J. Clin. Oncol. 2010, 28, 2220 – 2226. [2] R. Capocaccia, M. Sant, F. Berrino, A. Simonetti, V. Santi, F. Trevisani, Am. J. Gastroenterol. 2007, 102, 1661 – 1670 (Am. J. Gastroenterol. 2007, 102, 1660; J. S Scolapio, K. R. S. Gill, Am. J. Gastroenterol. 2007, 102, 1671). [3] H. Park, J. B. Lee, Y. J. Shim, Y. J. Shin, S. Y. Jeong, J. Oh, G. H. Park, K. H. Lee, B. H. Min, Mol. Cells 2008, 25, 305 – 311. [4] H. S. Oz, J. Zhong, W. J. de Villiers, Mediators Inflammation 2012, 2012, 813892. [5] N. Syed, J. Langer, K. Janczar, P. Singh, C. Lo Nigro, L. Lattanzio, H. M. Coley, E. Hatzimichael, J. Bomalaski, P. Szlosarek, M. Awad, K. O’Neil, F. Roncaroli, T. Crook, Cell Death Dis. 2013, 4, e458. [6] a) M. Kubo, H. Nishitsuji, K. Kurihara, T. Hayashi, T. Masuda, M. Kannagi, J. Gen. Virol. 2006, 87, 1589 – 1593; b) F. Izzo, M. Montella, A. P. Orlando, G. Nasti, G. Beneduce, G. Castello, F. Cremona, C. M. Ensor, F. W. Holtzberg, J. S. Bomalaski, M. A. Clark, S. A. Curley, R. Orlando, F. Scordino, B. E. Korba, J. Gastroenterol. Hepatol. 2007, 22, 86 – 91. [7] a) J. M. Kwan, A. M. Fialho, M. Kundu, J. Thomas, C. S. Hong, T. K. Das Gupta, A. M. Chakrabarty, Leuk. Res. 2009, 33, 1392 – 1399; b) W. B. Tsai, I. Aiba, S. Y. Lee, L. Feun, N. Savaraj, M. T. Kuo, Mol. Cancer Ther. 2009, 8, 3223 – 3233; c) R. H. Kim, J. M. Coates, T. L. Bowles, G. P. McNerney, J. Sutcliffe, J. U. Jung, R. Gandour-Edwards, F. Y. Chuang, R. J. Bold, H. J. Kung, Cancer Res. 2009, 69, 700 – 708; d) K. Beloussow, L. Wang, J. Wu, D. Ann, W. C. Shen, Cancer Lett. 2002, 183, 155 – 162; e) C. Y. Yoon, Y. J. Shim, E. H. Kim, J. H. Lee, N. H. Won, J. H. Kim, I. S. Park, D. K. Yoon, B. H. Min, Int. J. Cancer 2007, 120, 897 – 905; f) L. Stelter, M. J. Evans, A. A. Jungbluth, V. A. Longo, P. Zanzonico, G. Ritter, J. S. Bomalaski, L. Old, S. M. Larson, Mol. Imaging 2013, 12, 67 – 73. [8] P. A. Ascierto, S. Scala, G. Castello, A. Daponte, E. Simeone, A. Ottaiano, G. Beneduce, V. De Rosa, F. Izzo, M. T. Melucci, C. M. Ensor, A. W. Prestayko, F. W. Holtsberg, J. S. Bomalaski, M. A. Clark, N. Savaraj, L. G. Fenn, T. F. Logan, J. Clin. Oncol. 2005, 23, 7660 – 7668. [9] a) L. G. Feun, A. Marini, G. Walker, G. Elgart, F. Moffat, S. E. Rodgers, C. J. Wu, M. You, M. Wangpaichitr, M. T. Kuo, W. Sisson, A. A. Jungbluth, J. Bomalaski, N. Savaraj, Br. J. Cancer 2012, 106, 1481 – 1485; b) L. Feun, M. You, C. J. Wu, M. T. Kuo, M. Wangpaichitr, S. Spector, N. Savaraj, Curr. Pharm. Des. 2008, 14, 1049 – 1057; c) M. P. Kelly, A. A. Jungbluth, B. W. Wu, J. Bomalaski, L. J. Old, G. Ritter, Br. J. Cancer 2012, 106, 324 – 332. [10] Y. Ni, U. Schwaneberg, Z. H. Sun, Cancer Lett. 2008, 261, 1 – 11. [11] L. J. Shen, W. C. Shen, Curr. Opin. Mol. Ther. 2006, 8, 240 – 248. [12] L. Zhu, R. Verma, D. Roccatano, Y. Ni, Z. H. Sun, U. Schwaneberg, ChemBioChem 2010, 11, 2294 – 2301. [13] K. Hirokawa, A. Ichiyanagi, N. Kajiyama, Appl. Microbiol. Biotechnol. 2008, 78, 775 – 781. [14] H. H. Liao, Enzyme Microb. Technol. 1993, 15, 286 – 292. [15] F. W. Holtsberg, C. M. Ensor, M. R. Steiner, J. S. Bomalaski, M. A. Clark, J. Controlled Release 2002, 80, 259 – 271. [16] a) J. Nazor, S. Dannenmann, R. O. Adjei, Y. B. Fordjour, I. T. Ghampson, M. Blanusa, D. Roccatano, U. Schwaneberg, Protein Eng. Des. Sel. 2008, 21, 29 – 35; b) A. Jakoblinnert, J. Wachtmeister, L. Schukur, A. V. Shivange, M. Bocola, M. B. Ansorge-Schumacher, U. Schwaneberg, Protein Eng. Des. Sel. 2013, 26, 291 – 298. [17] W. M. Patrick, A. E. Firth, J. M. Blackburn, Protein Eng. 2003, 16, 451 – 457. [18] J. Sambrook, E. F. Fritsch, T. Maniatis, Molecular Cloning: A Laboratory Manual, Vol. 3, 2nd ed., 1989, Appendix B.12. [19] W. K. Vaughin, R. A. Neal, A. J. Anderson, Comput. Biol. Med. 1976, 6, 1 – 7.

ChemBioChem 2014, 15, 276 – 283

282

CHEMBIOCHEM FULL PAPERS [20] Y. Ni, Y. M. Liu, U. Schwaneberg, L. L. Zhu, N. Li, L. F. Li, Z. H. Sun, Appl. Microbiol. Biotechnol. 2011, 90, 193 – 201. [21] L. Zhu, K. L. Tee, D. Roccatano, B. Sonmez, Y. Ni, Z. H. Sun, U. Schwaneberg, ChemBioChem 2010, 11, 691 – 697. [22] G. A. Kotzia, N. E. Labrou, FEBS J. 2009, 276, 1750 – 1761. [23] a) D. Q. Gao, D. L. Narasimhan, J. Macdonald, R. Brim, M. C. Ko, D. W. Landry, J. H. Woods, R. K. Sunahara, C. G. Zhan, Mol. Pharmacol. 2009, 75, 318 – 323; b) D. Narasimhan, G. T. Collins, M. R. Nance, J. Nichols, E. Edwald, J. Chan, M. C. Ko, J. H. Woods, J. J. G. Tesmer, R. K. Sunahara, Mol. Pharmacol. 2011, 80, 1056 – 1065. [24] R. Martinez, F. Jakob, R. Tu, P. Siegert, K. H. Maurer, U. Schwaneberg, Biotechnol. Bioeng. 2013, 110, 711 – 720. [25] A. Galkin, L. Kulakova, E. Sarikaya, K. Lim, A. Howard, O. Herzberg, J. Biol. Chem. 2004, 279, 14001 – 14008. [26] T. Ohkuri, A. Yamagishi, Protein Eng. 2003, 16, 615 – 621. [27] K. Miyazaki, M. Takenouchi, BioTechniques 2002, 33, 1033 – 1038. [28] W. Wang, B. A. Malcolm, BioTechniques 1999, 26, 680 – 682. [29] C. Louis-Jeune, M. A. Andrade-Navarro, C. Perez-Iratxeta, Proteins Struct. Funct. Bioinf. 2012, 80, 374 – 381.

 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.chembiochem.org [30] a) S. F. Altschul, T. L. Madden, A. A. Schffer, J. Zhang, Z. Zhang, W. Miller, D. J. Lipman, Nucleic Acids Res. 1997, 25, 3389 – 3402; b) D. T. Jones, J. Mol. Biol. 1999, 292, 195 – 202; c) E. Krieger, K. Joo, J. Lee, J. Lee, S. Raman, J. Thompson, M. Tyka, D. Baker, K. Karplus, Proteins Struct. Funct. Bioinf. 2009, 77, 114 – 122; d) U. Muckstein, I. L. Hofacker, P. F. Stadler, Bioinformatics 2002, 18, S153 – S160; e) J. Qiu, R. Elber, Proteins Struct. Funct. Bioinf. 2006, 62, 881 – 891. [31] Y. Duan, C. Wu, S. Chowdhury, M. C. Lee, G. M. Xiong, W. Zhang, R. Yang, P. Cieplak, R. Luo, T. Lee, J. Caldwell, J. M. Wang, P. Kollman, J. Comput. Chem. 2003, 24, 1999 – 2012. [32] J. Schymkowitz, J. Borg, F. Stricher, R. Nys, F. Rousseau, L. Serrano, Nucleic Acids Res. 2005, 33, W382 – W388. [33] a) R. Guerois, J. E. Nielsen, L. Serrano, J. Mol. Biol. 2002, 320, 369 – 387; b) J. Van Durme, J. Delgado, F. Stricher, L. Serrano, J. Schymkowitz, F. Rousseau, Bioinformatics 2011, 27, 1711 – 1712.

Received: July 3, 2013 Published online on December 20, 2013

ChemBioChem 2014, 15, 276 – 283

283

Protein engineering of the antitumor enzyme PpADI for improved thermal resistance.

Arginine deiminase (ADI, EC 3.5.3.6) is a potential antitumor drug for the treatment of arginine-auxotrophic tumors such as hepatocellular carcinomas ...
779KB Sizes 0 Downloads 0 Views