Extracellular expression of Saccharomyces cerevisiae’s L-asparaginase II in Pichia pastoris results in novel enzyme with better parameters

Abstract L-asparaginase (ASNase) is an efficient inhibitor of tumor development, used in chemotherapy sessions against acute lymphoblastic leukemia (ALL) tumor cells; its use results in 80% complete remission of the disease in treated patients. Saccharomyces cerevisiae’s L-asparaginase II (ScASNaseII) has a high potential to substitute bacteria ASNase in patients that developed hypersensitivity, but the endogenous production of it results in hypermannosylated immunogenic enzyme. Here we describe the genetic process to acquire the ScASNaseII expressed in the extracellular medium. Our strategy involved a fusion of mature sequence of protein codified by ASP3 (amino acids 26–362) with the secretion signal sequence of Pichia pastoris acid phosphatase enzyme; in addition, this DNA construction was integrated in P. pastoris Glycoswitch® strain genome, which has the cellular machinery to express and secrete high quantity of enzymes with humanized glycosylation. Our data show that the DNA construction and strain employed can express extracellular asparaginase with specific activity of 218.2 IU mg−1. The resultant enzyme is 40% more stable than commercially available Escherichia coli’s ASNase (EcASNaseII) when incubated with human serum. In addition, ScASNaseII presents 50% lower cross-reaction with anti-ASNase antibody produced against EcASNaseII when compared with ASNase from Dickeya chrysanthemi.


Introduction
L-asparaginase (ASNase -EC 3.5.1.1) catalyzes the deamination of L-asparagine (Asn) to L-aspartate and ammonia. Besides that, as secondary activity hydrolyzes L-glutamine (Gln) in glutamic acid and ammonia (GLNase activity). [1] ASNase is widely known for its use in therapeutics against acute lymphoblastic leukemia (ALL), which is a blood cancer that prevails mostly in children. [2] These neoplastic cells are defective in the synthesis of the amino acid Asn due to low or absent expression of asparagine synthetase gene (ASNS), thus becoming dependent on serum circulating Asn. [3] The administration of ASNase causes a depletion of Asn in the bloodstream, resulting in amino acid starvation that blocks protein synthesis in cancer cells, ultimately leading to failure of cell function and apoptosis. ASNase has become a hallmark in ALL treatment, as upon its introduction in clinical protocols the remission rate increased from 20% to 80%. [4] Bacteria Escherichia coli and Dickeya chrysanthemi (former known as Erwinia chrysanthemi) are the microorganisms approved by Foods and Drugs Administration (FDA) and European Medicine Agency (EMA) as sources of commercial ASNase. [5] Only the PEGylated (ASNase covalently conjugated with polyethylene glycol) form of E. coli ASNase is also available at the market. PEGylated form of D. chrysanthemi ASNase was at phase two of clinical trial; however, the study was interrupted by severe immunogenic reactions. [6,7] The main reason for this reaction was postulated as preexisting PEG sensitization. [8,9] Although the importance of using L-asparaginase in the treatment of ALL is unquestionable, the use of bacterial enzyme shows a high incidence of adverse effects, such as the formation of antibody anti-asparaginase that may lead to hypersensitivity, neurotoxicity and hepatotoxicity (caused by hydrolysis of L-glutamine, releasing high doses of ammonium, among others). Some patients present antibody crossreaction between the native form of E. coli ASNase (EcASNaseII) and its PEGylated form; in those cases, ErwASNase (ASNase from D. chrysanthemi) is used, since it does not present high antibody cross-reaction, serving as a second or third line of treatment. [10] The use of ErwASNase is not the first option due to its low human serum stability [10] and high glutaminase activity, [11] in addition to the high cost of this medicine. [12] Other ASNases that may overcome these drawbacks, especially the antibody cross-reaction, have been identified in eukaryotic sources, such as the ASNase from yeast Saccharomyces cerevisiae. Nevertheless, there are still some problems with this yeast-derived biologic drug, such as posttranslational hypermannosylation added into proteins produced by S. cerevisiae that can trigger the human immune response. [13] Hence, some Pichia pastoris strains, called Glycoswitch V R , are engineered to avoid the hypermannosylation process, providing an alternative to produce biologics in yeast. [14][15][16] The S. cerevisiae ASNase II (ScASNaseII) was first studied by Dunlop in 1978. [17] At first glance it was not so promising enzyme, as it was characterized with a specific activity of 48 IU mg À1 (half of EcASNaseII) and a K m of 250-300 mM for Asn (one order of magnitude higher than EcASNaseII). Later, further study [18] indicated higher specific activity of 204 UI mg À1 for ScASNaseII expressed in the periplasm of a recombinant P. pastoris, but the kinetics parameters were not determined. Also, expressing ScASNaseII in prokaryotic system is not an option, as it tends to remain insoluble [5] and with residual activity, [19] suggesting a major role of glycosylation in the enzyme's folding process.
The expression of heterologous proteins in the periplasm of microorganism faces big challenges when producing at large scale, such as loss in yield and biological activity, proteolytic degradation, toxicity of the extraction agents and obligatory monitoring of their removal, difficulty of industrial implementation and mixing of the periplasmic proteins with cytoplasmic proteins. [20] Therefore, the present study focuses on obtaining the recombinant ScASNaseII extracellularly using P. pastoris Glycoswitch expression system with humanized type of glycosylation, and on analyzing the effect of this type of expression system on the activity, oligomerization, physicalchemical properties and kinetic parameters. In addition, enzyme cross-reaction with anti-ASNase antibody generated against EcASNaseII and cytotoxicity against ALL cell line were evaluated. Altogether, our intention is to improve the expression system (extracellular production) and obtain a novel asparaginase with better parameters (potent and less immunogenic enzyme) that can be used as an alternative option in anti-ALL therapy.

Strain construction
The reagents applied were at the analytic purity level, as stated by the suppliers (e.g., Sigma-Aldrich, San Luis, Missouri, United States of America (USA); Synth, Diadema, São Paulo, Brazil; Thermo Fischer Scientific, Waltham, Massachusetts (MA), USA, etc.). The ASP3 gene was synthetized by GenScript, New Jersey, USA with optimized codons for expression in P. pastoris (GeneBank accession number MN913566) and amplified by polymerase chain reaction. The cloning method was adapted from the circular polymerase extension cloning technique. [21][22][23] The oligonucleotides (Table 1) were obtained at Exxtend Biotecnologia São Paulo, Brazil and were designed to amplify the ASP3 gene starting at the 26 amino acid. The gene was cloned into pJAG-s1 yeast expression vector (Biogrammatics V R Inc., Carlsbad, California, USA), however the signaling peptide in this vector was replaced by PHO1 signaling sequence, [24] which was obtained by PCR from genomic DNA of P. pastoris GS115 strain. The expression vector was transformed into E. coli DH5a for cloning and then integrated into genome host P. pastoris Glycoswitch pep4-sub2-(Biogrammatics V R Inc.). The clones resistant to geneticin were selected at a concentration of 5 mg mL À1 in YPD plate (10 g L À1 yeast extract, 20 g L À1 peptone, 20 g L À1 glucose, 20 g L À1 agar). A control strain was generated with the empty pJAG-s1 vector.

Yeast protein expression
Transformed P. pastoris were grown in BMGY (10 g L À1 yeast extract, 20 g L À1 peptone, 3.4 g L À1 yeast nitrogen base, 10 g L À1 ammonium sulfate, 100 mM pH 6.0 potassium phosphate, 0.75% glycerol) until an OD 600 of 20, then the medium was replaced with BMMY (same composition of BMGY, but with methanol instead of glycerol, ranging from 0.5% to 1.5%) and the induction was held for 5 days, maintaining the initial concentration of methanol every 12 h. The medium broth was collected and analyzed for its asparaginase activity by the L-aspartyl-b-hydroxamic acid (AHA) method indicated for no-purified enzymes. [2] The samples were centrifuged at 8000g and the supernatant was incubated with 0.1 M Asn (Sigma-Aldrich) and 0.1 M hydroxylamine pH 7.0 at 37 C for 30 minutes. Then ferric reagent (10% FeCl 3 , 5% trichloroacetic acid, in 0.66 M HCl) was added to react with the AHA and form a colorimetric substance that absorbs at 500 nm. The absorbance was interpolated to a standard curve made of AHA. GA GTC TAC GGT GGT TAA TCA AGA GGA TGT CAG pJAG-s1-forward ASP3 gene (italic) and pJAG-s1 (underlined) 6 C CAG ACT TAG AAT AGG AGA AAACAT CGT TTC GAA TAA TTA GTTG pJAG-s1-reverse PHO1 (italic) and pJAG-s1 (underlined)

Purification of recombinant ScASNaseII
Two cycles of tangential filtration using a 10 kDa ultrafiltration membrane (Merck-Millipore, Burlington, MA, USA) were applied for buffer exchange; the concentrate was then loaded into HiTrap TM DEAE FF (GE V R Healthcare Life Sciences (GE V R ), Boston, MA, USA) previously equilibrated with 50 mM Tris-HCl pH 8.8 and eluted with steps of NaCl concentration ranging from 100 mM to 1 M in the same buffer. The fractions with asparaginase activity were concentrated and then loaded into Superdex TM 200 Increase 10/300 GL (GE V R ) previously equilibrated with 50 mM Tris-HCl pH 7.4 and eluted in the same buffer. The chromatography steps were applied with the FPLC € AKTA and € AKTA Start (GE V R ) systems.
All fractions were analyzed by SDS-PAGE according to the method of Laemelli. [25] The protein concentration was measured using the Bradford method [26] and interpolated in a bovine serum albumin (BSA, Sigma-Aldrich) standard curve.

Dynamic light scattering (DLS)
For determination of molecular mass, ScASNaseII was filtered with a 0.22 mm polyvinylidene fluoride (PVDF) sterile filter, then 40 mg were applied to a quartz cuvette for the measurement. The equipment was calibrated with the same buffer. Measurements were done each 5 seconds at a total of 20 measurements in DynaPro NanoStar (Wyatt, Santa Barbara, California, USA) and the data was analyzed by DYNAMICS 7.8.0.26 software.

Enzymatic activity assay
The ASNase enzymatic activity of pure ScASNaseII was measured with Nessler's reagent (Merck-Millipore). [27] A range from 25 to 150 nM of enzyme was incubated for 10 minutes at 37 C with 20 mM Asn in 50 mM Tris-HCl pH 7.4. The reaction was diluted (10x) in water with trichloroacetic acid (TCA) 1.5 M followed by the addition of Nessler's reagent. The colorimetric solution was measured in SpectraMax Microplate Reader (Molecular Devices, San Jose, California, USA) at a wavelength of 440 nm and the values were interpolated in a standard curve of ammonium sulfate. One unit (UI) of enzyme activity is defined as 1 mmol of ammonia release per minute at 37 C.
AHA was also applied to evaluate the specific activity for hydrolysis of Asn and to compare it with the specific activity of hydrolysis of L-Gln. For the hydrolysis of Asn, a range of 25-150 nM of enzyme was incubated for 10 minutes at 37 C with 20 mM Asn and 0.1 M Hydroxylamine pH 7.0 in 50 mM Tris-HCl pH 7.4. Then ferric reagent was added and the absorbance was measured at 500 nm and interpolated in a standard curve of AHA. For L-Gln hydrolysis, the reaction was the same as for Asn, except a range of 250 nM to 1.5 mM of enzyme was incubated and 20 mM L-Gln was applied instead of Asn. One International Unit (U) of L-ASNase was defined as the amount of enzyme capable of hydrolyzing 1 mmol of amino acid L-asparagine into aspartate/glutamate per minute at 37 C.
All enzymatic analyses were done in conjunction of a blank sample and a negative control. Blank sample was settled as all the reaction without Asn, and negative control was settled as all the reaction without the enzyme.

Optimum pH
To determine the optimum pH of ScASNaseII activity, the enzyme was assayed with different buffers (100 mM): citrate pH 3.0, 4.0 and 5.0; potassium phosphate pH 6.0, 7.0 and 8.0; tris-HCl pH 9.0; sodium bicarbonate pH 10 and 11. In addition to the buffers, 500 mM of NaCl was added to eliminate the interference of different ionic strength. The reaction was measured using Nessler's reagent as presented in section 2.5.

Effect of human serum on enzyme stability
ScASNaseII was incubated at 37 C with 10% human serum (Sigma-Aldrich V R ) for up 5 days, and the enzymatic specific activity was measured using Nessler's reagent. The results were compared with a control and with the E. coli enzyme commercially available in the same conditions.

Glycosylation analyses
ScASNaseII were incubated with 1 unit of PNGase F (Sigma-Aldrich V R ) in 50 mM ammonium bicarbonate pH 7.8 at a final volume of 20 mL at 37 C, overnight. A control was assembled without the addition of PNGase F. The solution was analyzed by Native-PAGE and Blue Native-PAGE (12% acrylamide, 6 mL final volume), adapted from Kurien and Scofield, 2012. [28] The specific activity of glycosylated and deglycosylated enzymes was revealed by the Nessler's reagent.

Enzyme kinetics
The kinetics parameters of ScASNaseII was measured by the method described and adapted from Rigouin et al. [29] A fixed concentration of 50 nM ScASNaseII was incubated at 37 C with 100 mM Tris-HCl pH 7.4, 0.4 mM a-ketoglutarate, 0.13 mM b-nicotinamide adenine dinucleotide (b-NADH), 0.5 U malic dehydrogenase (MDH) and glutamic oxaloacetic transaminase (GOT). Asn concentration varied in a range of 0 to 2.5 mM. The oxidation of b-NADH is continuously measured at a wavelength of 340 nm and the extinction coefficient used was 6.3 Â 10 3 Mol À1 cm À1 at 340 nm.

Enzyme-linked immunosorbent assay (ELISA)
As ALL treatment using asparaginase is distinguished by several potential side effects (including immune reactions which may prevent treatment depending on the toxicity level) was evaluated the immunogenic reactions between EcASNaseII's antibody and ScASNaseII. Enzyme cross-reaction with anti-ASNase antibody generated against EcASNaseII was evaluated by ELISA technique. ELISA plate (Bio-rad) was coated overnight at 4 C with 10 ng of EcASNaseII as control and with up to 1000 ng of ScASNaseII or ErwASNase. The plate was washed with washing solution (WS, PBS 0.05% Tween-20) and the wells were blocked with BSA 1% in PBS for 1h30m at 37 C. The supernatant was discarded and washed with WS and the primary antibody (anti-EcASNaseII produced in rabbit-commercially obtained from Rheabiotech) was applied at a dilution of 1:10,000 (diluted in PBS, fetal bovine serum 10% (FBS, Vitrocell Embriolife, Campinas, Brazil)) and the plate was incubated at 37 C for 1h30m. The plate was washed with WS and the secondary antibody (anti-rabbit conjugated with alkaline phosphatase -KPL) was applied to the wells at a dilution of 1:3000 (diluted in PBS, FBS 10%); the plate was incubated at 37 C for 1h30m. The remaining solution was discarded, and the wells washed with WS and then with PBS. The substrate solution (1% o-phenylenediamine, 0.003% H 2 O 2 in 50 mM citrate-phosphate buffer pH 5.5) was added and incubated at 37 C for 15 min and stopped with 2 M H 2 SO 4 . The absorbance was measured at 492 nm at SpectraMax Microplate Reader.

Cytotoxicity against ALL cell line
The acute lymphoblastic leukemia human cell line MOLT-4 was obtained from the Banco de C elulas do Rio de Janeiro (RJ/Brazil). Cells were incubated at 37 C in a 5% CO 2 incubator with RPMI 1640 medium with 10% FBS, 2.5 g L À1 of glucose, 10 mM HEPES and 1 mM sodium pyruvate. The cells were centrifuged, resuspended in the same fresh medium, and 5 Â 10 3 cells were incubated in a 96-well microplate with up to 1 IU mL À1 of ScASNaseII at a final volume of 150 mL. As controls, cells were incubated with the same enzyme's dilution buffer (50 mM Tris-HCl pH 7.4) or with only culture medium. The incubation was held for 72 h and then thiazolyl blue was added and followed the standard protocol. [30] Statistical analysis and imaging software All analyses were done in triplicate, and the results were shown as mean ± standard deviation (SD). Statistical analysis and all graphs were generated by Prism 5 from GraphPad Software.

Extracellular expression and purification
Pichia pastoris Glycoswitch pep4-sub2-strain was transformed with pJAG-PHO-ASP3. The original signaling sequence from the pJAG-s1 expression vector (Biogrammatics) was replaced with the secretion signaling sequence of acid phosphatase PHO1 from P. pastoris. [24] This final expression vector encodes the ASP3 gene (starting at the 26 amino acid-mature ScASNaseII protein sequence) fused with PHO1. Transformed P. pastoris pep4-sub2-clones resistant to 5 mg mL À1 of G418 were selected and tested for ScASNaseII extracellular production. Clones with ScASNaseII expression were tested with different methanol concentration in the function of time (Figure 1). Compared to the control strain (transformed with empty expression vector), the expression of ScASNaseII in the extracellular medium is very high, achieving 30 units per milligram of total protein. An International Unit of ScASNaseII (or any L-asparaginase) is defined as the amount of enzyme capable of hydrolyzing 1 mmol of asparagine into aspartate and ammonia per minute at 37 C. The clone 29 was chosen, as it has higher specific activity and yields compared with the other clones after the purification step.
It was necessary to concentrate the proteins secreted in extracellular medium and to exchange the buffer to purify the protein. Thus, a tangential filtration was applied with a cutoff of 10 kDa. The concentrate was applied to an anionic exchange chromatography (Figure 2a); the red rectangle represents the single peak with asparaginase activity. Although this step was sufficient for purifying our protein (data not shown), the solution remained with a brownish color, therefore, an exclusion chromatography was applied (Figure 2b) and the protein eluted pure as shown by the SDS-PAGE (Figure 2c, uncut version in Supporting Information Figure  S1). The theoretical molecular weight (MW) of each monomer of ScASNaseII without glycosylation is 36 kDa; here we found two bands close to 39 and 40 kDa, suggesting addition of glycans of around 3-4 kDa in total. All fractions with ASNase activity were pooled and the biochemical parameters measured.

Biochemical and kinetic parameters
The DLS data (Supporting Information, Figure S2) showed a signal with mean hydrodynamic radius of 5.006 ± 0.049 nm, which translates to a MW of 146 ± 3 kDa, suggesting a tetrameric structure of ScASNaseII.
The specific activity of purified ScASNaseII for Asn substrate measured by the Nessler's reagent at pH 7.4 was 77 IU mg À1 (Figure 3a). We also measured ScASNaseII activity by the AHA method, and it was found 218 IU mg À1 ( Figure  3b). The glutaminase (GLNase) activity was measured by the AHA method ( Figure 3c) and represents 3% of ASNase activity (measured by the same method), which is very close to the percentage of EcASNaseII's GLNase activity. [31] The optimum pH was 6.0, same as EcASNaseII, [32] and the activity measured was 82 IU mg À1 at this pH (Figure 3d).
For comparison, the Leuginase V R (commercial formulated EcASNaseII) and EcASNaseII (recombinant ASNase from E. coli obtained and purified at our laboratory) activities were also measured at physiological pH using the Nessler's reagent (Figure 3e,f), and they were 178 IU mg À1 and 80 IU mg À1 , respectively.
The stability in human serum of ScASNaseII was measured and compared to Leuginase V R (Figure 4a,b). The ScASNaseII maintained 98% ± 2.36 of its activity in human serum and 90% ± 2.54 in 50 mM Tris-HCl pH 7.4 (control), for 120 h at 37 C. This is higher than the stability of Leuginase V R , which maintained only 60% ± 3.15 in human serum and 40% ± 4.25 in 50 mM Tris-HCl pH 7.4 (control).
The presence of glycosylation in this genetic expression system did not interfere with the specific activity of ScASNaseII ( Figure 5). The Zymogram and Blue Native-PAGE revealed a relative homogeneous glycosylation pattern (Figure 5a,b, uncut version in Supporting Information Figure S3a and S3b), as the protein band is not smeared. For comparison, Leuginase V R was also applied to the gel ( Figure 5). The brownish intensity of the band represents the intensity of ASNase activity.
The kinetic parameters were determined by coupled assay with NADH oxidation. A fixed concentration of ScASNaseII was incubated with different concentrations of Asn. The initial velocity of the hydrolysis of Asn was plotted in function of Asn concentration (Figure 6a). The kinetics parameters are described in Table 2; secreted ScASNaseII presents Michaelian behavior, with K m of 120.5 ± 6.7 mM for Asn, i.e., lower affinity for Asn than EcASNaseII. [27] The turnover (k cat ) of ScASNaseII was 47 ± 0.6 s À1 , and the catalytic efficiency was 3.8 Â 10 5 M À1 s À1 .

ELISA
The cross-reaction between EcASNaseII's antibody and ScASNaseII was measured by ELISA technique. The anti-EcASNaseII recognizes the enzyme EcASNaseII resulting in a mean absorbance of 0.9 ± 0.011. Meanwhile, the reaction of this antibody to ScASNaseII presented mean absorbance of 0.040 ± 0.006 at the same protein concentration. Even incubating one hundred-fold higher ScASNaseII protein amount, the maximum cross-reaction reaches approximately 1/3 of the value found to EcASNaseII protein. For comparison, ErwASNase, a second-line choice of ALL treatment, presented a mean absorbance of a maximum of 0.620 ± 0.002 at 1000 ng of protein incubated (Figure 7) and at almost all protein amount tested, ScASNaseII reached less than 50% of ErwASNase's absorbance.

Cytotoxicity against MOLT-4
The cytotoxic effect of ScASNaseII was measured by incubating MOLT-4 ALL cell line. In 96-well microplates, 5 Â 10 3 cells were added in 150 mL supplemented culture medium (RPMI-1640 þ 10% FBS) per well and ASNases were added to a final concentration from 0.00 to 1.0 U mL À1 . A control study was run in parallel with EcASNaseII, obtained at our laboratory. The data from ScASNaseII was compared with the EcASNaseII to evaluate the antileukemic activity. After 72 h of incubation, the cell viability was measured by MTT assay (Figure 8). ScASNaseII killed 50% of viable cells at a concentration of 0.9 IU mL À1 , as EcASNaseII killed 50% of viable cells at a concentration of 0.03 IU mL À1 .

Discussion
The enzyme ASNase has been used as first line treatment of ALL because of its ability to deplete Asn from the bloodstream, thus leading to leukemic cell apoptosis. Many sources of ASNase have been found and are being studied as a potential biopharmaceutical to treat ALL; [33,34] but, in addition to biochemical features, the complexity of obtainment process is crucial to industrial viability.
As recombinant protein, there are innumerous expression systems for reaching certain protein form, and slight variations in this expression system change to whole procedure and final product. The combination of the secretion signaling sequence of PHO1 from P. pastoris and the glycoengineered strain of P. pastoris knockout for the genes pep4-and sub2-Glycoswitch allowed us to achieve the secretion of ScASNaseII to the extracellular medium and the resultant protein presented high specific activity and better kinetic properties. Previous studies [17,18] obtained ScASNaseII in the periplasm of S. cerevisiae and P. pastoris, respectively, which its extraction and purification is heavily impaired at industrial large scale. [20] Our work obtained the enzyme in the extracellular medium, which is easier to purify and less costly to obtain for industrial application. ScASNaseII has a theoretical MW of approximately 144 kDa and is present in tetrameric form, which is in conformity with the enzyme obtained in this work. The DLS data (Supporting Information Figure S2) suggest a MW of 146 ± 11 kDa. This is the same MW of EcASNaseII from the literature [32] and this data is corroborated with the Native-PAGE electrophoretic mobility of Leuginase V R (Figure 5b). Interestingly, the current programs that analyze DLS' data cannot perceive slightly changes in  The effect of different pH on the specific activity of ScASNaseII for hydrolysis of Asn: buffers citrate pH 4 and 5; potassium phosphate pH 6, 7 and 8; Tris-HCl pH 9; sodium bicarbonate pH 10 and 11. The enzyme only presented measurable activity at pH 5 through 10. In addition to the buffers, 500 mM of NaCl were added to eliminate the difference in electrostatic effect. (e) Specific activity of Leuginase V R for hydrolysis of Asn measured by the Nessler's reagent method. (f) Specific activity of EcASNaseII for hydrolysis of Asn measured by the Nessler's reagent method. All points in the graph represents mean ± SD (n ¼ 3). Figure 4. Stability of ScASNaseII versus Leuginase V R in (a) 10% human serum and (b) 50 mM Tris-HCl pH 7.4 (control) at 37 C for 120 h. Y-axis is percentage of activity; 100% represents the value found to the first data (0 h) and subsequent data calculated as a percentage of this first record. Specific activity analyzed by the Nessler's reagent method. Bars represents mean ± SD (n ¼ 3). Statistical analysis measured by two-way ANOVA, Ã p < 0.001. hydrodynamics radius that human-like glycosylation does, [35] thus this result is not taking into account the glycosylation profile of ScASNaseII. Each ScASNaseII's monomer has a theoretical value of 36 kDa and it has three potential glycosylation sites. Reducing SDS-PAGE analysis (Figure 2c) reveals two bands representing ScASNaseII, each of them with approximately 39 and 40 kDa, respectively. This difference from the theoretical MW (36 kDa) is due to the presence of glycan structure with an addition of 3-4 kDa on each monomer and each glycan structure has approximately 1 kDa. This glycosylation profile is very different from what was obtained previously in the literature. This enzyme was already obtained in other expression systems with a heterogeneous ramified high-mannose glycosylation profile with more than 200 fragments of mannose, and around 10 kDa of glycan structure added, respectively. [17,18] Our enzyme has the same ASNase activity measured by the AHA method of previously reported study. [18] This method is indirect and has the advantage of not being disturbed by already present ammonia in the solution.   Table 2. Kinetics parameters of hydrolysis of Asn of ScASNaseII and EcASNaseII [27] .
EcASNaseII [27] 23 ± 5 0.032 ± 0.00061 134 ± 2 5.8 Â 10 6 ScASNaseII 120.5 ± 6.7 0.051 ± 0.00066 46.73 ± 0.59 3.8 Â 10 5 However, hydroxylamine is unstable above neutrality, [36] thus this method is not suitable to measure optimum pH in wide range. Therefore, we decided to measure activity by the Nessler's reagent as it is a direct method, indicated by the FDA to measure ASNase activity. [37] Our data shows an optimum pH of 6.0 (same as EcASNaseII [31] ), measured by the Nessler's reagent, rather than the optimum pH found at the literature of 7.2 and 9, measured by the AHA method. [18] The activity of ScASNaseII was the same as EcASNaseII obtained at our laboratory, but lower than Leuginase V R 's activity measured by Nessler's reagent. Although the comparison to a commercial enzyme is necessary, the formulation of Leuginase V R has additives that enhance their activity and stability. This was already shown to ErwASNase [38] and depending on the additive, it could double its activity. Also, ScASNaseII enzyme maintained 98% ± 2.36 of its activity through 120 h of incubation with human serum, compared to 60% ± 3.15 of Leuginase V R . This is a huge advantage for ScASNaseII, as Leuginase V R has additives that help to stabilize the drug inside the formulation [38] and the present preparation does not.
The glutaminase activity from ScASNaseII was 3% of its ASNase activity, the same as EcASNaseII. [31] The therapeutic enzyme EcASNaseII usually has low affinity to Gln. [33,39] However, the Gln concentration in the bloodstream is very high, close to millimolar range. [40] Thus, glutaminase activity is an important factor that must be considered, since it is one of the culprits of hyperammonemia. [41] However, the glutaminase activity is vital to the cytotoxicity of ASNases for ALL cells that present medium to high ASNS expression. [11] In comparison, ErwASNase's GLNase activity is approximately 10% of its ASNase activity, and it was already found higher amounts of ammonia and glutamate and lower glutamine content in patient's serum [42] than that treated with EcASNaseII, which may result in more frequent adverse effects. [41] The glycosylation has no effect on the ScASNaseII activity ( Figure 5c); this means that the P. pastoris Glycoswitch used as expression system in this work produced a glycosylation profile that did not interfere in the active site of ScASNaseII. This is different from reported in the literature, [18] in which a higher ASNase activity was observed after the treatment with PNGase F (capable to make protein deglycosylation). This can be explained by the fact that P. pastoris strain without genetic modification on the glycosylation machinery was applied in the literature, and the glycosylation profile of normal P. pastoris adds high mannose content that can be responsible for the decreasing of the enzyme activity. [43,44] The kinetic profile of ScASNaseII was only characterized by Dunlop et al., [17] which described ScASNaseII as possessing a Michaelis-Menten behavior with K m of 250-300 mM. However, they obtained this result after a series of purification steps of a periplasmic enzyme with high heterogeneous mannose content and still did not obtain a pure enzyme. Our results are the first report of the kinetics of pure recombinant ScASNaseII. Our data reveals a K m for Asn of 120.5 mM (Figure 6a and Table 2). The difference may be because our enzyme is pure and has a homogeneous glycosylation profile as shown by the SDS and Native-PAGE (Figures 2c and 5b) results.
The affinity for Asn is intrinsically related to the cytotoxicity effect of ASNase in vivo. [45] K m (affinity for Asn) of micromolar range is required, as the concentration of Asn in the bloodstream is ranging from 1 to 32 mM. [40] We detected lower affinity of ScASNaseII for Asn when compared with EcASNaseII previously obtained at our laboratory (Table 2). Although the affinity for Asn is an important factor, other parameters may overcome this drawback, as occurs in Oncaspar V R (commercial PEGylated EcASNaseII). The pegylation process reduces 70% of its specific activity but it drastically increases its half-life (from approximately 1 day to several days), [46,47] thus compensating its lower activity.
The cross-reaction between EcASNaseII's antibody and ScASNaseII is an important factor to study for a potential biopharmaceutical intended as second-line treatment of ALL, as the antibodies against EcASNaseII may react with the second enzyme administrated (e.g., another asparaginase Figure 7. ELISA to evaluate cross reaction between EcASNaseII's antibody with ScASNaseII and ErwASNase. The absorbance signal of ScASNaseII was considerably lower than EcASNaseII, even with 10 times more protein, and at every amount of protein, ScASNaseII is approximately 50% less immunogenic than ErwASNase. The bars in the graph represent mean ± SD, n ¼ 3. Figure 8. Viability of MOLT-4 after treatment with ScASNaseII. The IC 50 of ScASNaseII was at 0.9 IU mL À1 . The bars in the graph represent mean ± SD (n ¼ 3). Statistical analysis measured by one-way ANOVA, p ¼ 0.0147. that will replace that produced by E. coli). Here we have analyzed cross-reaction between EcASNaseII's antibody and ScASNaseII or ErwASNase (Figure 7). It was necessary 100fold more ScASNaseII to reach the same binding signal (antibody recognition) that was found using EcASNaseII (1000 ng of ScASNaseII versus 10 ng of EcASNaseII, p < 0.001). This is expected, as ScASNaseII is a different enzyme and in case of severe immune reactions caused by EcASNaseII, this new asparaginase could be used in the treatment. Moreover, ScASNaseII still did not react as much as ErwASNase. In fact, ScASNaseII is around 50% less recognized by antibodies against EcASNaseII compared to ErwASNase. Overall, this means as a substitute for EcASNaseII for second-line treatment of ALL, ScASNaseII is a better choice in terms of serum stability and antibody cross-reaction.
The cytotoxicity of ScASNaseII was previously measured against K562, an acute myeloid leukemia (AML) cell line. However, they obtained an IC 50 of 5 IU mL À1 , [18] which is inviable to apply as an antileukemic drug in vivo, due to high amount of protein required to achieve therapeutic effect. Our findings demonstrates a IC 50 of 0.9 IU mL À1 against MOLT-4, [48,49] a ALL cell line, suggesting a more suitable application of ScASNaseII as an antileukemic drug against ALL than AML.

Conclusion
This the first report of ScASNaseII being secreted into the medium broth due to our genetic modification strategy; this ensure a faster and cheaper way of purifying the enzyme, as it was needed only one chromatography step to purify it from contaminant proteins. Also, we have shown that this strategy produced a better enzyme in terms of biochemical characteristics and kinetic properties than reported earlier and it was considerably more stable in human serum than EcASNaseII. ScASNaseII is more capable of being applied as an anti-ALL drug than anti-AML and it is a better choice of a second-line treatment of ALL than ErwASNase in terms of antibody cross-reaction. Also, its activity could be further enhanced by adding osmolytes as excipients in the formulation, as the drugs available at the market. Here we showed that genetic expression system manipulation can modify biopharmaceutical final properties.

Ethics statement
The article does not contain any studies with human participants or animals performed by any of the authors.