Recombinant expression of an l‐amino acid oxidase from the fungus Hebeloma cylindrosporum in Pichia pastoris including fermentation

Abstract l‐amino acid oxidases (LAAOs) are flavoenzymes that catalyze the oxidative deamination of l‐amino acids to the corresponding α‐keto acids, ammonia, and hydrogen peroxide. Here, we show the overexpression, purification, and the characterization of LAAO4 from the fungus Hebeloma cylindrosporum in the yeast Pichia pastoris with a 9His‐tag and compare this with the recently characterized 6His‐hcLAAO4 expressed in E. coli. The expression of the enzyme with an ER‐signal sequence in P. pastoris resulted in a glycosylated, secreted protein. The enzymatic activity without activation was higher after expression in P. pastoris compared to E. coli. Due to treatment with acidic pH, a striking increase of activity could be detected for both expression systems resulting in similar specific activities after acid activation. Regarding the substrate spectrum, temperature stability, K m, and v max values, hcLAAO4 showed very few differences when produced in these two expression systems. A higher yield of hcLAAO4 could be obtained by fermentation.

. Only d-amino acid oxidases (DAAOs) can be produced in a sufficient amount and quality for a biotechnological application (Pollegioni & Molla, 2011).
Recently, we were able to express two fungal LAAOs in E. coli, a gram-negative bacterium. The LAAO1 from the fungus Rhizoctonia solani (rsLAAO1) was expressed as a fusion protein with maltose-binding protein (Hahn, Neumeister, et al., 2017) and the LAAO4 from the fungus Hebeloma cylindrosporum (hcLAAO4) with an N-terminal 6His-tag (Bloess et al., 2019). rsLAAO1 showed low activity and has to be activated by low amounts of SDS (Hahn, Hertle, et al., 2017). Because of the possible interference of SDS in process development, we identified another LAAO with a broader substrate spectrum. hcLAAO4 can be activated by brief exposure to acid pH and stays in the active conformation (Bloess et al., 2019). Here, we show the expression of hcLAAO4 in Pichia pastoris as secreted protein and describe differences of the enzyme expressed in E. coli and in P. pastoris with regard to glycosylation, acid activation, substrate spectrum, kinetic parameters, and temperature stability. Moreover, we use fermentation in bioreactors to produce a higher cell density and achieving a higher yield of recombinantly expressed enzyme.

| Amplification and cloning of 9His-hcLAAO4
The synthetic gene of the Hebeloma cylindrosporum LAAO4 (hcLAAO4) without the putative ER signal sequence (GenBank accession number MH751433, Bloess et al., 2019) was amplified with a forward primer (AAGAATTCCATCATCATCACCATCACCACCACC ATATGGCTGACACTACTTCCGTTC) to introduce an EcoRI recognition site (underlined sequence) followed by a 9His-tag (bold) and a codon for a methionine residue upstream of the hcLAAO4 coding sequence. The reverse Primer (AAGGCGGCCGCTTAAACGGAAAC) contained the stop-Codon (bold) and the NotI site (underlined sequence). The PCR product was cloned via an intermediate step into pGEM-TEasy (Promega), digested with EcoRI and NotI, and ligated into the EcoRI and NotI digested pPIC9K (MPI), which encodes the prepro-sequence of α-factor, yielding pSBL11. The four N-glycosylation sites predicted by NetNGlyc were changed into alanine residues one after the other via overlap PCR with the primers listed in Table A1 resulting in pMCH24.

| Genome integration and selection of clones
For genome integration, pSBL11 or pMCH24 were linearized at the SalI site and transformed into the spheroplasts of P. pastoris SMD1163 yeast strain (Life Technologies) for integration into HIS4 locus. After electroporation, 1 M ice-cold sorbitol was added immediately and the cells were spread on MD-His agar plates (0.17% yeast nitrogen base, 0.5% biotin, 2% glucose, amino acids, and 1.5% Agar) and were incubated at 30°C for 2 days, according to Wu & Letchworth 2004. The clones were spread onto YEPD agar plates with increasing concentration of G-418 (Geneticin, Calbiochem) (1, 2.5, 5, 7.5, and 10 mg/ml) and cultured at 30°C for several days. The selection was done according to Scorer, Clare, McCombie, Romanos, & Sreekrishna, 1994. Clones were screened for the highest 9His-hcLAAO4 expression after induction with methanol.
Cells were centrifuged and diluted into 500 ml BMMY medium (2% peptone, 1% yeast extract, 100 mM potassium phosphate pH 6.0, 1.34% yeast nitrogen base, 0.4 µg/ml biotin, and 1% methanol) for 72 hr at 15°C. Methanol was added every 24 hr to a final concentration (v/v) of 1% to keep the methanol concentration up. Cells were pelleted (4°C, 4600 g, 10 min), and the supernatant consisting of the medium with secreted 9His-hcLAAO4 was stored at 4°C and used for purification of the enzyme. Cells were spheroplasted and 0.5 ml ruptured with a Precellys24 bead homogenizer with four ceramic beads (1.4 mm) by two repeats of twice 20 s at 6500 rpm interrupted by a 10-second break to analyze for 9His-hcLAAO4 remaining within the cells. A soluble fraction was separated from larger particles by centrifugation (4°C, 18,000 g, 5 min).
9His-hcLAAO4 was expressed in a 7 L-NLF (Bioengineering AG) bioreactor with an initial volume of 3 L BSM-medium (26.7 ml/L ophosphoric acid (86%), 0.93 g/L CaSO 4 , 18.2 g/L K 2 SO 4 , 14.9 g/L MgSO 4 ·7H 2 O, 4.13 g/L KOH, 40 g/L glycerol (99%), 4.35 ml/L PTM trace salts (6 g/L CuSO 4 , 0.08 g/L NaI, 3 g/L MnSO 4 ·H 2 O, 0.2 g/L MoO 3 ·H 2 O, 0.02 g/L H 3 BO 3 , 0.5 g/L CoCl 2 , 20 g/L ZnCl 2 , 65 g/L FeSO 4 ·7H 2 O, 0.2 g/L biotin, 5 ml/L sulfuric acid)). The pH was kept at 6.0 by automated addition of 25% (w/w) ammonia solution and 10% (w/w) phosphoric acid, respectively. To avoid foaming Pluronic ® , PE 8100 was added automatically. Fermentations were performed at 30°C with 0.2 bar overpressure and an aeration rate of 5 NL/min. The initial stirrer frequency was set to 200 min −1 and speed was increased in steps of 2% when the relative dissolved oxygen saturation (rDOS) fell below 30% (batch-phase). A first supplementary feeding started automatically after 5 hr when rDOS raised from 30% to 60% for the first time and stopped each time when rDOS felt below 60% again. When 1 L of 500 g/L glycerol (99%) with 1.2% (v/v) PTM trace salts was added during the first feeding phase the cultivation temperature was lowered to 15°C and a second feeding phase with a solution of 1 L methanol with 1.2% (v/v) PTM trace salts was started after 47 hr of cultivation. A methanol concentration of 0.3% (v/v) in the bioreactor was maintained by feed control via a methanol sensor (Alcosens; Heinrich Frings GmbH & Co. KG). The expression was performed for about 4-5 days. In total, 735 g (928 ml) of methanol solution was added.
To avoid foaming Pluronic ® , PE 8100 was added automatically.
Fermentations were performed with 0.2 bar overpressure and an aeration rate of 2 NL/min. The initial stirrer frequency was set to 200 min −1 and speed was increased in steps of 2% when rDOS felt below 30%. In the case of fed-batch procedures, feeding of a solution of 600 g/L glycerol, 90 g/L yeast extract, and 1 g/L MgSO 4 ·H 2 O was started when rDOS raised from 30% to 60% for the first time and stopped each time when rDOS felt below 60% again. Several growth periods were tested ranging from 28 to 70 hr at 30°C or 37°C. Expression was induced with 0.2 mM IPTG after cooling down the fermentation broth to 11°C for 17-25 hr. After fermentation, cells were disrupted by the French press and the enzyme purified according to Bloess et al., 2019.

| Purification of the secreted 9His-hcLAAO4
Purification of the 9His-hcLAAO4 fusion protein was carried out at 4°C. After the removal of the cells by centrifugation, the supernatant with the secreted 9His-hcLAAO4 in the medium was loaded onto Ni 2+ -NTA resin. The flow-through was caught, mixed, and reloaded onto Ni 2+ -NTA resin via a peristaltic pump for 24 hr and a flow velocity of 1 ml/min. The column was washed with 40 column volumes (cv) of His-washing buffer (50 mM Na 2 HPO 4 pH 7.0, 300 mM NaCl, and 20 mM imidazole). After washing, 9His-hcLAAO4 was eluted with His buffer (50 mM Na 2 HPO 4 pH 7.0, 300 mM NaCl) containing different concentrations of imidazole (50, 100, 250, and 500 mM).

| Enzymatic assay
Determination of the enzymatic activity of 9His-hcLAAO4 or 6His-hcLAAO4 by measuring the initial rate of H 2 O 2 production was done by coupled peroxidase/o-dianisidine assay with 10 mM l-amino acid as described (Hahn, Neumeister, et al., 2017). The standard assay mixture contained 10 mM l-glutamine, 50 mM TEA/HCl buffer (pH 7.0), 0.2 mg/ml of o-dianisidine, 5 U/ml peroxidase and hcLAAO4 in limiting amounts (0.75 µg in a 200 µl assay). Reactions were carried out in 96-well plates at 30°C in a Tecan Spark microplate reader at 436 nm. One unit was defined as the amount of enzyme that catalyzes the conversion of 1 μmol l-amino acid per minute. Initial velocities of H 2 O 2 production were determined with 16 different l-amino acid concentrations between 0.02 and 20 mM for the untreated and acid-activated 9His-hcLAAO4. K m and v max values were calculated from Hanes-Woolf plots. To determine the pH optimum, the enzyme was treated with buffers at different pH-values (Sörensen glycine-HCl buffers pH 2-2.5 (Sörensen, 1909), 200 mM Na 2 HPO 4 , 100 mM citric acid buffers pH 3-6.5 (McIlvaine, 1921), 50 mM TEA buffers pH 7-9 and 50 mM glycine pH 9.5-11) before the assay. Activation was tested by preincubation for 10 min at the indicated pH and activity determined at pH 7.0. For determining the temperature stability, the enzymes were incubated at the respective temperature in the TEA/HCl buffer for 1 hr.
The digestion was done overnight at 37°C. As control samples, PNGaseF was replaced with water and these samples were also incubated at 37°C or 4°C overnight. Samples were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE, Laemmli, 1970) and stained with Coomassie Brilliant Blue R-250.

| Expression and purification of 9His-hcLAAO4
Recently, we described the successful expression of 6His- The gene was cloned into the P. pastoris expression vector pPIC9K which encodes the prepro-sequence of the α-mating factor of S. cerevisiae as an N-terminal ER import signal sequence. pSBL11 was integrated into the HIS4 locus of P. pastoris SMD1163 cells. Clones were selected using G-418 as the resistance against G-418 is proportional to the integrated expression cassettes (Scorer et al., 1994). The optimal parameters for expression of 9His-hcLAAO4 were identified as a temperature of 15°C, a pH-value of 6.0, and a methanol concentration of 1% for 72 hr. About 25% of 9His-hcLAAO4 was secreted into the cell medium and could be found in the supernatant (Figure 1a,b, fraction S) whereas about 75% were found in the cell pellet fraction (Figure 1a,b, fraction P). After the rupture of the cells, low-speed centrifugation was used to separate soluble proteins and small organelles (fraction SF) from larger particles (fraction ISF). 9His-hcLAAO4 was detected in similar amounts in both fractions. The intracellular enzyme had higher mobility compared to the secreted form. These data indicate that 9His-hcLAAO4 obtained posttranslational modifications before secretion. Ni 2+ -NTA resin was used to purify the enzyme from the culture medium, the supernatant obtained after removing the cells by centrifugation (Figure 1c, fraction S). Only low amounts of 9His-hcLAAO4 did not bind to the resin and could therefore not be purified (Figure 1c, fraction F). In the washing fraction, as well as the first elution fractions, no protein could be eluted ( Figure 1c, fractions W1, 1-5). 9His-hcLAAO4 of high purity was obtained with 500 mM imidazole (Figure 1c, fractions 6-11). The purified enzyme was subjected to size exclusion chromatography ( Figure 1d,e). It eluted as a single peak corresponding to a molecular mass of 216 kDa. The FAD loading of the enzyme expressed in P. pastoris was 80%-100% in different expressions ( Figure A4), as determined spectroscopically (see (Hahn, Neumeister, et al., 2017)) indicating that the enzyme is folded properly for cofactor binding. In total, we obtained 4 mg of the enzyme with a total activity of about 90 U for the untreated enzyme from a 500 ml culture.

F I G U R E 1
Purification of recombinant 9His-hcLAAO4. The construct for the 9His-hcLAAO4 with prepro-⍺-factor ER signal sequence was integrated into the genome of P. pastoris SMD1163. 9His-hcLAAO4 fusion protein was expressed at 15°C, pH 6 and 1% methanol (final concentration (v/v), added every 24 hr) for 72 hr. (a) The expression was analyzed by Western blotting. A secreted enzyme could be found in supernatant (S). The pellet (P) contained the cells including cytosolic proteins. The pellet was homogenized and separated into a soluble fraction (SF) and an insoluble fraction (ISF). (b) The fractions were quantified using Image J (n = 6; error bars represent standard deviations). (c) The culture medium (S) was loaded onto Ni 2+ -NTA resin. After collection of the flow-through (F) and wash fractions (W1), proteins were eluted using increasing concentrations of imidazole (1-11). Fractions were separated by SDS-PAGE, and proteins were stained with Coomassie. (d) Untreated 9His-hcLAAO4 was separated by size exclusion chromatography. It eluted at a retention volume of 11.5 ml corresponding to a molecular mass of 216 kDa. Protein was measured as absorbance at 280 nm (black) and FAD as absorbance at 450 nm (red). (e) Fractions were analyzed by SDS-PAGE. The column load was diluted 1:10

| Activity and pH optimum of 9His-hcLAAO4
The activity of 9His-hcLAAO4 was assayed using 10 mM l-glutamine as a substrate. The initial production rate of H 2 O 2 was measured using a coupled peroxidase/o-dianisidine assay. To investigate the pH optimum of the 9His-hcLAAO4, the purified enzyme was incubated with buffers at pH 2-11 ( Figure 2). For the untreated enzyme, the activity remains on the same level in the range of pH 4-9 (Figure 2a).
The specific activities of 9His-hcLAAO4 in untreated conditions varied between 10 and 20 U/mg while only 2 U/mg were observed for 6His-hcLAAO4 expressed in E. coli (Bloess et al., 2019). As the E. coli expressed enzyme was activated by incubation for 10 min at pH 3.0, this procedure was also tested for 9His-hcLAAO4 expressed in P.
pastoris. With acid treatment, the specific activity could be increased twofold up to 20-40 U/mg. This specific activity is similar to the one obtained for the activated E. coli expressed enzyme. A pH optimum between pH 6-9 could be detected for the acid-activated (pH 3) enzyme ( Figure 2b). A complete loss of activity for a pH under 3 could be determined for both conformations. With more basic pH-values, a decrease in activity was observed.

| Glycosylation of 9His-hcLAAO4
Expressed in E. coli, the 6His-hcLAAO4 fusion protein has a mobility of about 65 kDa in SDS-gels (Bloess et al., 2019). This is in agreement with the estimated molecular weight of about 69.6 kDa (determined via ExPASy ProtParam). Due to an apparent molecular weight slightly above 75 kDa, 9His-hcLAAO4 may be glycosylated when expressed in P. pastoris. It is known that snake venom LAAOs and a secreted LAAO from the fungus Trichoderma harzianum are glycosylated (Pollegioni et al., 2013;Yang et al., 2011). The sequence of hcLAAO4 contains four putative N-glycosylation sites predicted by the NetNGlyc program at amino acid residues N54, N164, N193, and N331 ( Figure A1). Deglycosylation of purified 9His-hcLAAO4 was performed by adding PNGaseF. After incubation, two lower molecular weight bands could be obtained compared to the control is partially deglycosylated. Furthermore, without activation, higher specific activity was obtained when the enzyme was expressed in P.
pastoris compared to E. coli. Therefore, we wondered whether glycosylation increases enzymatic activity. The activity was determined after treatment with and without PNGaseF for the untreated and the acid-activated enzyme (Figure 3c). Neither the untreated enzyme nor the activated (pH 3) enzyme showed a loss of activity after deglycosylation.

| Enzymatic properties of 9His-hcLAAO4
Next, we wanted to test, whether the expression system affected the substrate scope. The best substrates for untreated 9His-hcLAAO4 were l-glutamine followed by l-leucine, other hydrophobic l-amino acids, l-glutamic acid, and basic l-amino acids at a substrate concentration of 10 mM (Table 1). Amino acids with a branched β-position were hardly accepted as substrates (l-valine) or not converted (l-tert-leucine and l-phenlyglycine). Glycine and all tested d-amino acids were not accepted. Interestingly, amino acids F I G U R E 2 pH optimum of 9His-hcLAAO4. Enzymatic activity was assayed with peroxidase and o-dianisidine in a glycine-HCl buffer (circles), citric acid buffer (squares), TEA buffer (triangles), and glycine buffer (diamonds) at the indicated pH with l-glutamine as substrate. (a) The untreated 9His-hcLAAO4 showed a broad maximum between pH 4 and 9 (10-20 U/mg). (b) 9His-hcLAAO4 was preincubated for 10 min at pH 3 before dilution into the indicated pH. The pH optimum was between pH 6 and 9 for the acid-activated enzyme with a twofold increase in the specific activity. Data are means of three independent experiments; error bars represent standard deviations like dl-homophenylalanine and 4-chloro-dl-phenylalanine were converted with 44 and 34% relative activity compared to l-glutamine (Table 1). Even methyl and ethyl esters were substrates for the 9His-hcLAAO4 expressed in P. pastoris. There was no big difference in substrate scope for untreated and acid-activated 9His-hcLAAO4.
Kinetic properties were determined for untreated and acid-activated 9His-hcLAAO4 by measuring initial velocities at different substrate concentrations and calculations from Hanes-Woolf plots (Table 2). l-glutamine, l-leucine, l-phenylalanine, and l-methionine were tested as well as the methyl esters of l-leucine and l-phenylalanine. K m -values were below 1 mM for these free amino acids and barely changed after acid activation. These results indicated that activation did not change the substrate affinity of the enzyme.
Only the two tested methyl esters showed a lower affinity (K m -value 1.85 mM for l-phenylalanine methyl ester and 3.0 mM for l-leucine methyl ester) for the untreated enzyme. The v max values increased about threefold due to acid activation.
To determine the effect of the expression system and acid activation on the stability of hcLAAO4, we incubated the enzymes for an hour at various temperatures. Activities were assayed by coupled peroxidase/o-dianisidine assay (Figure 4). 9His-hcLAAO4 expressed in P.
pastoris was stable up to 70°C independent of acid treatment. Their stability was similar to acid-activated 6His-hcLAAO4 expressed in E. coli.
Interestingly, untreated 6His-hcLAAO4 showed an increase in activity with a maximum at 50°C compared to 30°C. However, the specific activity was sixfold lower at 50°C compared to the acid-activated enzyme.
At 80°C, a total loss of activity for all enzymes could be detected.

| High cell density fermentation of E. coli and P. pastoris in bioreactors
To obtain higher amounts of recombinant 9His-hcLAAO4, we grew P. pastoris to high cell density in a fermenter culture. P. pastoris is F I G U R E 3 Glycosylation of 9His-hcLAAO4 expressed in P. pastoris. (a) Purified 9His-hcLAAO4 was treated with PNGaseF overnight at 37°C to remove the glycan structure from the fusion protein or incubated at 37°C without additions. To exclude that the temperature had an impact, the enzyme was stored at 4°C. The separation was performed by SDS-PAGE, and proteins were stained with Coomassie. (b) Untreated 9His-hcLAAO4 (P. pastoris), deglycosylated 9His-hcLAAO4 (P. pastoris +PNGaseF), 9His-hcLAAO4 without any predicted glycosylation sites (△N-Glyc.), and untreated 6His-hcLAAO4 (E. coli) were separated by Coomassie-stained SDS-PAGE. (c) Deglycosylation did not reduce the activity of the 9His-hcLAAO4. 9His-hcLAAO4 was incubated as indicated above. Enzymatic activity was assayed with peroxidase and o-dianisidine in TEA buffer pH 7 for untreated 9His-hcLAAO4 and acid-activated fusion protein with l-glutamine as substrate. Data are means of three independent experiments; error bars represent standard deviations. The data were normalized to the 4°C control of the untreated enzyme known for its high cell density fermentation to produce recombinant proteins (Siegel & Brierley, 1989). P. pastoris was grown to a cell density of about 500 OD600 (up to 112 g/L dry cell weight concentration) before induction of 9His-hcLAAO4 expression by methanol for 4-5 days in a bioreactor. The time course of 9His-hcLAAO4 activity as well as the cell density was followed during P. pastoris bioreactor cultivation ( Figure 5). Whereas the cell density remained the same during fermentation (about 450-550), the activity in the supernatant increased steadily. The activity in the supernatant reached 9200 U/L compared to 180 U/L in shaking flask cultures (Table 3).
We also fermented E. coli Arctic Express (DE3) to increase the yield of recombinant 6His-hcLAAO4. Various cultivation conditions for cell growth before induction were tested, whereby the best result was performed after a 47 hr growth phase in batch procedures at 30°C and an induction time of 25 hr. While cell densities were higher after fermentation compared to shaking flask cultures, the activity per gram dry mass was lower (Table 3). Compared to P. pastoris, induction periods for the expression in E. coli were significantly shorter (110 hr vs. 25 hr).

| D ISCUSS I ON
Even though H. cylindrosporum is a fungus (Basidiomycota), we were able to express hcLAAO4 without its ER signal sequence in the cytosol of E. coli (Bloess et al., 2019). The yeast Pichia pastoris (Ascomycota) has some major advantages over bacterial expression systems such as posttranslational modifications in ER and Golgi followed by secretion into the medium (Cereghino & Cregg, 2000). lamino acid oxidases from snake venom and flounders have already been expressed in P. pastoris as secreted proteins (Kasai et al., 2015;Kommoju, Macheroux, & Ghisla, 2007). Here, we describe that by adding the prepro-sequence of the α-factor from S. cerevisiae, recombinantly expressed 9His-hcLAAO4 was secreted into P. pastoris Note: Standard substrate concentration: 10 mM. a 2.5 mM was used because of low solubility.
Glycosylation was described for snake venom LAAOs (Geyer et al., 2001;Hayes & Wellner, 1969) and a secreted LAAO from the fungus Trichoderma harzianum ETS 323 belonging to Ascomycota (Yang et al., 2011). Due to this glycosylation of a secreted fungal LAAO, the glycosylation of 9His-hcLAAO4 in the yeast P. pastoris and its native signal sequence for ER import we assume that this enzyme is also glycosylated in H. cylindrosporum.
Enzymes expressed in P. pastoris and E. coli could both be activated by brief exposure to acidic pH and reached similar specific activities after acid activation. The kinetic parameters K m and v max were also similar. However, the expression of the enzyme in P. pastoris resulted in a higher specific activity without activation.
9His-hcLAAO4 expressed in P. pastoris showed v max values of 9 U/ mg for l-glutamine, whereas the activity of the enzyme expressed in E. coli displayed 2 U/mg for this substrate (Bloess et al., 2019).
Deglycosylation of untreated 9His-hcLAAO4 expressed in P. pastoris did not reduce activity. These data argue against a direct influence of N-glycosylation on activity. However, glycosylation may induce a conformation of intermediate activity during ER import, which is stable even after the removal of the N-glycans. Alternatively, there may be an additional posttranslational modification, which is not evident by the mobility shift in SDS-PAGE.
The enzyme is stable during a preincubation for one hour at a temperature of up to 70°C independent of expression in E. coli or P. pastoris or treatment with acidic pH. Untreated 6His-hcLAAO4 expressed in E. coli was even more active after preincubation at temperatures between 40°C and 70°C compared to a preincubation at 30°C. These data indicate that this enzyme can be transformed into a more active conformation by heat in addition to freezing, acid treatment, and SDS addition as reported before (Bloess et al.,  (open circles (untreated) and open squares (acid-activated)) and 9His-hcLAAO4 expressed in P. pastoris (filled triangles (untreated) and crosses (acid-activated)). Enzymes were incubated for 1 hr at the indicated temperature and assayed immediately with peroxidase and o-dianisidine in TEA buffer pH 7 at 30°C with l-glutamine as substrate. Enzyme expressed in E. coli or P. pastoris was stable up to 70°C. Untreated enzyme expressed in E. coli was activated by incubation between 40°C and 70°C. Data are means of three (E. coli enzymes) or six (P. pastoris enzymes) independent experiments; error bars represent standard deviations F I G U R E 5 Temporal progression of cell density (OD600) and activity of expressed 9His-hcLAAO4 in a bioreactor after induction. Expression was carried out at 15°C and pH 6 for 108 hr in a bioreactor. Final methanol concentration was regulated with the pO 2 value. Cell density was measured by determining the absorbance of the culture (open triangles). Samples were taken and centrifuged, and enzymatic activities were assayed in the untreated supernatants (circles) and pH 3 treated supernatants (squares) with peroxidase, o-dianisidine, and l-glutamine at 30°C 2019). Heat activation has also been observed before. Tyrosine aminotransferase from chick liver can undergo repeated cycles of reversible heat activation and cold inactivation (Shioji et al., 1978).
The substrate scope of 9His-hcLAAO4 expressed in P. pastoris is nearly the same as for 6His-hcLAAO4 expressed in E. coli (Bloess et al., 2019) indicating that glycosylation is without influence.
In comparison with shaking flask cultures (4 g/L), fermentation of E. coli in bioreactors leads to a higher cell dry mass per L (16-39 g/L) with a higher activity per L (2800-10,000 U/L compared to 2000 U/L). In contrast, the activity per g dry mass was lower (175-260 U/g and 455 U/g). However, results in fermentation were variable regarding enzyme yields and activities. Various cultivation conditions for the growth of the E. coli cells were tested, whereby batch procedures at 30°C gave the best results. This is likely to be due to the better plasmid stability under these conditions. But the results in fermentation were variable regarding enzyme yields and activities even with the same cultivation conditions. Also, realizing higher dry cell mass concentrations by high cell density fermentation was limited due to decreasing plasmid stability over time. Moreover, biomass yields regarding the medium-dry mass are rather low (0.14 for fed-batch fermentation), indicating cell stress, probably caused by the depletion of l-amino acids and the production of toxic H 2 O 2 by hcLAAO4. Because of shorter cultivation times, the productivity of hcLAAO4 was higher for E. coli compared to P. pastoris.
Fermentation of 9His-hcLAAO4 in P. pastoris yielded a 50-fold higher activity per liter (9200 U/L) in the medium compared to expression in a shaking flask (180 U/L). One reason is the 10-fold higher cell density, which can be obtained by fermentation in a bioreactor. The others are the controlled feeding and the prolonged cultivation time of 4.5 days possible in fermentation compared to 3 days in the shaking flasks. Besides, hcLAAO4 was secreted into the extracellular environment protecting the cells from the depletion of l-amino acids and the production of toxic H 2 O 2 as well as facilitating down-stream processing. Thus, P. pastoris is an excellent expression system to produce recombinant hcLAAO4 in large quantities for industrial use.

ACK N OWLED G M ENTS
We thank Thomas Schäffer, Sena Yilmaz, Seda Yilmaz, and Annamaria Latus for excellent technical support. Funding was supplied by Bielefeld University. Open access funding enabled and organized by Projekt DEAL.

CO N FLI C T O F I NTE R E S T
None declared.

E TH I C S S TATEM ENT
Work with recombinant DNA has been performed according to national requirements (Dt-55.3.5-5/94-Bi, Anlage Nr. 412). The absorption spectrum between 350 and 500 nm is typical for flavoproteins. Cofactor occupancy was calculated from the ratio of the absorbance at 280 and 450 nm. The loading was 80%-100% in different expressions. Spectra were determined from three different experiments as described (Hahn, Neumeister, et al., 2017)