Codon Optimisation Is Key for Pernisine Expression in Escherichia coli

Background Pernisine is an extracellular serine protease from the hyperthermophilic Archaeon Aeropyrum pernix K1. Low yields from the natural host and expression problems in heterologous hosts have limited the potential applications of pernisine in industry. Methodology/ Principal Findings The challenges of pernisine overexpression in Escherichia coli were overcome by codon preference optimisation and de-novo DNA synthesis. The following forms of the pernisine gene were cloned into the pMCSGx series of vectors and expressed in E. coli cells: wild-type (pernisinewt), codon-optimised (pernisineco), and codon-optimised with a S355A mutation of a predicted active site (pernisineS355Aco). The fusion-tagged pernisines were purified using fast protein liquid chromatography equipped with Ni2+ chelate and gel filtration chromatography columns. The identities of the resultant proteins were confirmed with N-terminal sequencing, tandem mass spectrometry analysis, and immunodetection. Pernisinewt was not expressed in E. coli at detectable levels, while pernisineco and pernisineS355Aco were expressed and purified as 55-kDa proforms with yields of around 10 mg per litre E. coli culture. After heat activation of purified pernisine, the proteolytic activity of the mature pernisineco was confirmed using zymography, at a molecular weight of 36 kDa, while the mutant pernisineS355Aco remained inactive. Enzymatic performances of pernisine evaluated under different temperatures and pHs demonstrate that the optimal enzymatic activity of the recombinant pernisine is ca. 100°C and pH 7.0, respectively. Conclusions/ Significance These data demonstrate that codon optimisation is crucial for pernisine overexpression in E. coli, and that the proposed catalytic Ser355 has an important role in pernisine activity, but not in its activation process. Pernisine is activated by autoproteolytical cleavage of its N-terminal proregion. We have also confirmed that the recombinant pernisine retains the characteristics of native pernisine, as a calcium modulated thermostable serine protease.


Introduction
Assembly of pMCSGx expression vectors (x = 7,9,10) Aeropyrum pernix was cultivated as previously described [16], and its genomic DNA (gDNA) was isolated using gDNA isolation kits (Sigma). This A. pernix gDNA was used as the template for the wild-type pernisine (pernisine wt ). The pernisine, pernisine co and pernisine S355Aco genes were obtained using polymerase chain reaction (PCR), and cloned according to the relevant instruction manuals [17]. Briefly, the PCR products of these pernisine genes were amplified using sense and antisense primers: wild-type (5`-TACTTCCAATCCAATGCCGCAGCAGGATC GGCGGCTGGGGCTAG-3`, 5`-TTATCCACTTCCAATGTTAGCTTGAGACGGCAGTC TGCAC-3`) and codon-optimised (5`-TACTTCCAATCCAATGCCGCAGCAGGTAC GAAAATCGCCGCTATCGC-3`, 5`-TTATCCACTTCCAATGTTAACTGGAGACAGCC GTTTGGACAG-3`). The treatment of the PCR products with T4 DNA polymerase in the presence of dCTP generated 15 nucleotides with long single-strand overhangs. Conversely, the treatment of the previously linearised pMCSGx vectors with the restriction enzyme SspI followed by T4 DNA polymerase in the presence of dGTP created the complementary overhangs. In the next step, the plasmid and the PCR products were linked in the annealing process. The ligation-independent cloned constructs were transformed into competent DH5α cells, which were grown in Luria-Bertani (LB) medium to produce larger quantities of the vectors. Three different constructs with different tags for each gene were constructed.
A single colony was cultivated overnight at 37°C in 25 ml LB medium supplemented with the appropriate antibiotic, under constant agitation at 240 rpm. The next day, 475 ml fresh LB medium containing the appropriate antibiotic was added to 25 ml of the overnight culture. When the cells reached an optical density at 600 nm (OD 600 ) of 0.6 to 0.8, expression was induced with 1 mM isopropyl β-D-1-thiogalactopyranoside. The culture growth times after this induction ranged from 1 h to 4 h, as optimised initially by the detection of pernisine on dot blots. The cells were centrifuged (6,000x g, 20 min, 4°C) and resuspended in 25 ml lysis buffer (30 mM Tris-HCl, 0.3 M NaCl, 1 mg ml -1 lysozyme, pH 7.5). The cells were then lysed by sonication (amplitude 40%; 10 s on, 10 s off; 120 s; VCX 750 by Sonics), and centrifuged (19,000x g, 20 min, 4°C). The supernatants were used for analysis and purification of pernisine. The pellets were resuspended in 4 M urea and subjected to SDS-PAGE, for determination of the insoluble pernisine fraction.
N-terminal His 6 -tagged pernisine was purified using Ni 2+ -Sepharose 6 FF columns (GE Healthcare), followed by size exclusion chromatography using a HiLoad 16/60 Superdex 200 preparative grade column (GE Healthcare). The unbound samples were washed out with 20 column volumes of binding buffer (20 mM Na 2 HPO 4 , 0.5 M NaCl, 20 mM imidazole, pH 7.4), and the bound samples were eluted with the same buffer containing 500 mM imidazole. The eluted proteins were applied directly onto the size exclusion column, which was equilibrated with 30 mM Tris HCl, 0.3 M NaCl, pH 7.4. The pernisine fractions were collected, dialysed (SPECTRA/POR, MWCO 8-10 kDa) in 10 mM HEPES, pH 8.0, for 4 h, and lyophilised (Christ alpha 1-2LD Plus, Germany). All of the purification procedures were performed at 4°C.

Activation of recombinant pernisine
Recombinant pernisine (1 mg ml -1 ) was dissolved in activation buffer (10 mM HEPES, 1 mM CaCl 2, pH 8.0) and heat activated in 100 μl aliquots in PCR tubes, at 90°C for 1 h; this activated recombinant pernisine was then used for proteolytic assays, unless otherwise specified. Preliminary tests for the activation were performed, with the recombinant pernisine (1 mg ml -1 ) incubated in activation buffer at different temperatures (60, 80, 90°C) for specific times (0, 15, 30, 60, 120 min). Immediately after these incubations, azocasein assays were carried out to determine the proteolytic activities of the samples (not shown). The specified time when the proteolytic activity at each temperature reached maximum was considered as full conversion of the proform of recombinant pernisine to the mature form of pernisine. The optimal activation conditions chosen were 1 h at 90°C.

SDS-PAGE, Western blotting and dot blots
Protein samples (10 μg) were analysed by SDS-PAGE using 12% polyacrylamide gels, and visualised using Coommassie brilliant blue staining. Cell lysates were normalised to cell density. For the Western blotting, the proteins were electrotransferred to nitrocellulose membranes. The membranes were blocked with 5% (w/v) non-fat dry milk in Tris-buffered saline with 0.05% (v/v) Tween 20 (TBST) at room temperature for 1 h. The His 6 -tagged pernisine was detected using rabbit polyclonal anti-histidine antibodies diluted 1:1,000, with incubation at room temperature for 1 h. The bound antibodies were detected with horseradish-peroxidaseconjugated goat anti-rabbit IgG antibodies (dilution, 1:2,000). Visualisation was performed using the ELC detection reagent, according to the manufacturer instructions (GE Healthcare). The dot blot analysis was carried out using anti-His 5 antibodies (Qiagen) diluted 1:1,000, according to manufacturer instructions. These dot blots for the different expression times for pernisine after the induction were quantified using the ImageJ software.

N-terminal sequencing
After separation by SDS-PAGE, the proteins were transferred onto PVDF membranes (Bio-Rad). These membranes were rinsed with Milli-Q water, stained with Ponceau S (Sigma) for 2-3 min, and destained with several changes of Milli-Q water. Purified His 6 -tag pernisine was cleaved with tobacco etch virus (TEV) protease, and subjected to SDS-PAGE. The excised tagless pernisine band was subjected to N-terminal sequencing, by automatic degradation in a Procise 492A protein sequencer (PE Applied Biosystems) at the Jozef Stefan Institute, Slovenia.

Tandem mass spectrometry
Marked bands I and II shown in Fig 1A were cut out of the SDS-PAGE. The reagents were prepared in 100 mM ammonium bicarbonate buffer. The protein samples were reduced using 10 mM dithiothreitol at 56°C for 30 min, and alkylated with 55 mM iodacetamide at room temperature for 20 min. The gel pieces were transferred to glass tubes (300 μL) and 20-30 μL 3 M HCl was added. These tubes were placed inside tubes containing 700 μL water and microwaved for 10 min. Afterwards, the supernatant was removed from the glass tubes and desalted directly on Oasis HLB Elution Plate (Waters), according to the manufacturer instructions. The samples were eluted (50 μL) and dried in a SpeedVac centrifuge (Eppendorf). The dried samples were dissolved in 10 μL reconstitution buffer (water: acetonitrile [96:4, v/v] in 0.1% formic acid), and analysed at the European Molecular Biology Laboratory (EMBL), Germany.
The peptides were separated using a nanoAcquity ultra-performance liquid chromatography system (Waters) fitted with a trapping column (nanoAcquity Symmetry C 18 ) and an analytical column (nanoAcquity BEH C 18 ). Solvent A was 0.1% formic acid in water, and solvent B was 0.1% formic acid in acetonitrile. The samples (8 μL) were loaded onto the trapping column with 5 μL min -1 solvent A. The peptides were eluted via the analytical column with a flow rate of 0.3 μL min -1 . During the elution step, solvent B increased as a linear gradient from 3% to 10% over the first 5 min, and then to 40% over the next 10 min. The peptides were introduced into the mass spectrometer (Orbitrap Velos Pro; Thermo Scientific) using a Pico-Tip Emitter tip (New Objective), with a spray voltage of 2.2 kV applied. The capillary temperature was set to 300°C. Full scan mass spectra with a mass range of 300-1700 m/z were acquired in profile mode in the Fourier transform with a resolution of 30,000. The most intense ions (up to 15) from the full scan mass spectra were selected for sequencing in the linear trap quadropole mass spectrometer. A normalised collision energy of 40% was used, and fragmentation was performed after accumulation of 3 ×10 4 ions, or after a filling time of 100 ms for each precursor ion (as whichever occurred first). The tandem mass spectrometry (MS/MS) data were acquired in centroid mode. Only multiply charged precursor ions (2+, 3+, 4+) were selected for the MS/MS.
The MaxQuant software was used to filter the data and to create the. mgf files that were needed for searching in MASCOT (Matrix Science). The data were searched against a speciesspecific (Aeropyrum pernix K1) Uniprot database. The data were searched with the following modifications: Carbamidomethyl (C) (Fixed) and Oxidation (M) (Variable). The mass error tolerance for the full scan mass spectra was set at 20 ppm, and for the MS/MS spectra, at 0.5 Da. A search with no enzyme was used. The termini were postulated based on peptide ladders of increasing aa length, with all either starting or ending at the same residue (for the N-and C-termini, respectively). Representative gels of the purified pernisines, following electrophoresis on standard 12% SDS-PAGE (A) and on 12% SDS-PAGE with casein as substrate (B) for the zymography activity (4 h at 80°C). Staining was with Coomassie blue dye. Lanes 0, protein MW markers (indicated left); lanes 1, recombinant pernisine co ; lanes 2, recombinant pernisine S355Aco ; lanes 3 and 4, heat-activated pernisine co and pernisine S355Aco . Selected protein bands of pernisine co that were analysed by MS/MS are marked as I and II, (see also S3 and S4 Figs). *protein load of pernisine S355Aco is three times higher than pernisine co . doi:10.1371/journal.pone.0123288.g001

Protease activity
The protein concentrations were determined spectrophotometrically using the extinction coefficient of 1% 280nm = 60,850 M -1 cm -1 for pernisine with its proregion, and 1% 280nm = 57,870 M -1 cm -1 for the mature pernisine. Alternatively, they were determined by the Bradford method [18], using BioRad Protein Assay (BioRad) with bovine serum albumin as standard.
To determine the qualitative proteolytic activity of the recombinant pernisine, zymography procedures with standard SDS-PAGE were used, as described previously [2]. Briefly, the samples were applied in duplicates onto 12% SDS-PAGE gels without and with 0.1% (w/v) casein (Sigma Aldrich) and electrophoresed (125 V, 70 min). The gels with added casein were transferred into 2.5% Triton X-100 for 1 h, washed twice with buffer (50 mM Tris-HCl, 1 mM CaCl 2 , pH 8.0), and incubated in the same buffer at 80°C for 4 h. The proteolytic activity was visualised as clear bands on the gels, against a blue background, using Coomassie brilliant blue staining. The SDS-PAGE gels without casein were stained immediately after electrophoresis.
To characterise the recombinant pernisine, azocasein assays were used, as described previously [2], with addition of the pernisine activation step. The samples were assayed as triplicates and the standard errors calculated. Initially, the optimum proteolytic activities of the recombinant pernisine in the presence of different CaCl 2 concentrations (0-32 mM) were examined. Then, the effects of ionic strength on the pernisine activity were investigated, as different NaCl concentrations (0-500 mM). To define the optimum pernisine activity, standard azocasein assays were conducted at different temperatures from 40°C to 120°C, and at different pHs from pH 2 to pH 12. The buffers used were: pH 2 to pH 4, 50 mM glycine-HCl; pH 6 to pH 8, 50 mM HEPES; and pH 9 to pH 13, 50 mM glycine-NaOH. The pH at each incubated temperature was calculated according to the dpH/dT correction coefficient [19]. Then, a three-dimensional graph of the temperature and pH dependence against the pernisine relative activity was plotted using the OriginPro 8 programme. In the same way, the thermostability of the pernisine was evaluated using standard azocasein assays at different temperatures (40, 80, 110, 120°C) and incubation times (0.1, 1, 2, 4 h) in 50 mM Tris-HCl, pH 8.0 with 1 mM CaCl 2 .

Results and Discussion
Design and cloning of the wild-type and synthetic pernisine genes The efficiency of pernisine overexpression in the BL21(DE3) E. coli cells was compared between the wild-type and codon-optimised pernisine sequences. The synthetic pernisine gene was designed using the GeneOptimiser algorithm (Genscript) and synthesised by Genscript. Moreover, the predicted catalytic serine at site 355 was mutated into alanine (S355A) to analyse the enzymatic activity that then remained. The pernisine gene consists of 1293 bp (European Molecular Biology Laboratory: BAA79718.2), and it was amplified using specified primers and the gDNA of A. pernix K1 (i.e., pernisine wt ) or the synthetic codon-optimised genes (i.e., pernisine co , pernisine S355Aco ) as templates. This codon optimisation replaced the rare codons in the pernisine wt gene with more frequent codons, as given in the S1 Table, while the aa sequence remained unchanged. The GC nucleotide content remained the same, at 57%. Altogether, 25.3% of the nucleotides in the DNA sequence were changed (S1 Fig). With the de-novo synthesis, the mRNA stability was improved, and the unfavourable mRNA structures and rare codons were reduced. Heterologous expression of rare codon-containing genes is likely to exhaust the endogenous pools of the analogous tRNAs and lead to growth inhibition, premature termination of transcription and/or translation, decreased mRNA stability, and increased frameshifts, deletions and misincorporations. Similar techniques of improved overexpression have been shown for some other proteins [8,10,20,21].
High-throughput cloning of the pernisine wt and pernisine co sequences in the pMCSGx series of vectors incorporates the N-terminal tags, followed by a TEV cleavage site (His 6 -TEV, His 6maltose binding protein [MBP]-TEV, His 6 -glutatione S-transferase [GST]-TEV). Agarose electrophoresis revealed that the lengths of the amplified pernisine wt and pernisine co corresponded to the expected ca. 1,300 bp. Linearised pMCSGx (x = 7, 9, 10) vectors were seen at the expected lengths of ca. 5300, ca. 6000 and ca. 6300 bp (data not shown).
Specific removal of the tags is an option when there is a TEV cleavage site between the tags and the pernisine [17]. The His 6 tag was used for simplified purification and detection of pernisine and MBP, with the GST tag to improve solubility [22].

Overexpression and purification of recombinant pernisine
Various expression strains of E. coli transformed with the pMCSGx constructs containing the wt or codon-optimised pernisine sequences were tested. The recombinant pernisine was purified using affinity chromatography and gel-exclusion chromatography, as presented schematically in Fig 2A. The chromatogram for the purified pernisine ( Fig 2B) and the SDS-PAGE of selected pernisine fractions showed the purified pernisine at around 55 kDa (Fig 2C, red line).
The constructs containing the pernisine wt gene were not successfully overexpressed in any of the tested expression cells (i.e., E. coli BL21(DE3), BL21(DE3)pLysE and BL21(DE3)pMagic, BL21-CodonPlus(DE3)RIL), as none of the expressed protein was seen at significantly higher levels compared to the cell lysates before and after induction, using SDS-PAGE (S2 Fig). The BL21(DE3) cell growth curve obtained by measuring OD 600 did not show any significant deviation compared to the control BL21(DE3) cells (transformed with empty pMCSG7; data not shown), which indicated that the recombinant pernisine is non-toxic for E. coli.
While E. coli is the most used host for heterologous gene expression, sometimes codon use indicates that it is not an optimal host for expression of the recombinant proteins because of the significantly divergent codon bias between the two organisms, especially for the first 10 codons at the beginning of the translation [23]. As a consequence, for heterologous gene expression, the presence of non-optimal codons in the DNA sequence expressed can result in inefficient translation, and sometimes in aborted translation. Using the BL21(DE3)pMagic and BL21-CodonPlus(DE3)RIL E. coli strain carrying a plasmid for rare tRNA aminoacyls (e.g., Arg [AGG], Arg [AGA], Ile [AUA], Leu [CUA]), we replaced the rarest codons for pernisine, as given in S1 Table. However, supplying these extra tRNAs did not resolve the problem of producing pernisine at detectable levels using SDS-PAGE.
Catara and co-workers overexpressed recombinant pernisine wt that lacked the signal sequence in E. coli, but could not detect it distinctively in crude extracts using SDS-PAGE [1]. They observed the recombinant pernisine indirectly, through its degradation products.
We compared the expression of pernisine wt and pernisine co and pernisine S355Aco in the pMCSGx constructs using the BL21(DE3) E. coli strain. Initially, small-scale expression was carried out to evaluate the time of expression for pernisine production. First, dot blots of the total cells lysates at prolonged growth times after induction (1, 2, 3, 4 h) showed that 3 h after induction was optimal for pernisine co overexpression ( Fig 2D) and again no His 6 -tagged protein using the pernisine wt sequence was detected. Later, large-scale expression was performed as described in the Materials and methods. Analysis of the cell lysates by Western blotting revealed overexpressed pernisine co at around 55 kDa, 100 kDa and 80 kDa (Fig 2F, lanes 2, 4, 6), which represented pernisine with the His 6 tag, fusion with MBP, and fusion with the GST tag, respectively. Also here pernisine wt could not be detected ( Fig 2F, lanes 1, 3, 5). The protein lysates of pernisine co and the mutant pernisine S355Aco showed these overexpressed protein bands, as marked with black arrows in Fig 2E (lanes 1-6), at the molecular weights corresponding with the Western blotting. From the SDS-PAGE analysis of the pellet, we estimated that around 20% of the recombinant pernisine was insoluble (data not shown). Addition of the MBP or GST tags did not significantly improve the pernisine co solubility or overexpression in E. coli. TEV cleavage efficiency was 88% in the case without fusion partners. Whereas fusion partners (GST and MBP tag) resulted in about two times lower efficiency (data not shown). The apparent molecular weights of the pernisine co fused with the tags were higher than the theoretical molecular weights (S2 Table). The reason for this is most likely the physical nature of the recombinant pernisine itself. The shape and charge of proteins have effects on their mobility under SDS-PAGE. Altered SDS stoichiometry can result in electrophoretic anomalies, as shown for highly hydrophobic proteins or their parts [24], as can incomplete denaturation of the pernisine before electrophoresis. The final yields of the lyophilised pernisine co and pernisine S355Aco were around 10 mg per litre of culture. Indeed, the use of such codon-optimised genes is becoming more attractive, and recently more examples of improved overexpression of such proteins have been reported [25,26,27]. Although there have been further studies carried out, to date, improvements to heterologous protein expression using codon-optimisation or by supplying extra tRNAs remain more or less empirical [28].

Identification and maturation of recombinant pernisine
Recombinant pernisine was detected by immunodetection (Fig 2D and 2E) and by N-terminal sequencing and MS/MS analysis. Immunodetection showed distinctively the codon-optimised His 6 -tagged pernisine (Fig 2D and 2E). N-terminal sequencing showed that the purified recombinant pernisine cleaved with TEV protease starts with S-N-A-A-A. Those five aa represent the remaining residues from the TEV cleavage site of the tagged pernisine. The recombinant pernisine at an apparent molecular weight of 55 kDa represents its preform, and includes the His 6 tag (Fig 1A, lanes 1, 2, mark I). During temperature maturation in the presence of CaCl 2 , the pernisine underwent autoproteolytic cleavage of its N-terminal proregion, and the resulting mature pernisine was seen at around 36 kDa (Fig 1A, lanes 1, 2, mark II). The theoretical mass of the putative signal sequence plus the proregion (1-92 aa) of the recombinant pernisine was around 9.3 kDa. The addition of a fusion tag modifies this by ca. 2 kDa. The apparent molecular mass of the recombinant pernisine was around 8 kDa above the theoretical  (1, 2, 3, 4 h) of pernisine wt and pernisine co for total cell lysates. Proteins were transferred (dot blot) onto nitrocellulose membranes and His 6 -tagged pernisine was detected with anti-His 5 -tag antibodies. Quantification was done using the ImageJ software (right panel). (E) SDS-PAGE analysis of pernisine co and pernisine S355Aco for total cell lysates of BL21(DE3) E. coli containing the pMCSGx series of vectors. 1, 4, Pernisine co/wt -pMCSG7; 2, 5, Pernisine co/wt -pMCSG9; 3, 6, Pernisine co/wt -pMCSG10. (F) Immunodetection of pernisine wt and pernisine co for cell lysates containing the pMCSGx series of vectors. Proteins were transferred onto nitrocellulose membranes and His 6 -tagged pernisine was detected using anti-His 6 -tag antibodies. (G, H) Azocasein assays of the purified pernisine co , showing effects of CaCl 2 (G) and NaCl (H). Relative proteolytic activities of activated pernisine co are shown according to the CaCl 2 concentrations, and to the NaCl concentrations in the presence (grey, dot-dash line) and absence (black line) of 1 mM CaCl 2 . Non-activated pernisine co in the absence of 1 mM CaCl 2 is also shown (black, dot line). Abbreviations: co-codon-optimised, wtwild-type.
doi:10.1371/journal.pone.0123288.g002 molecular mass. MS/MS investigation of the N-and C-terminals of the selected matured and non-matured pernisine indicated that the cleavage site of the proregion appears to be between Gln92 and Ala93 (S3 and S4 Figs). Indeed, a comparison of S3 and S4 Figs shows higher abundance of the peptides identified from the pernisine N-terminal to aa 92. The abundance of the peptides from aa 93 to the pernisine C-terminal is more or less the same across S3 and S4 Figs These data are in agreement with the SDS-PAGE analysis and the prediction of the native pernisine proregion defined from its alignment with Tk-subtilisin [2]. The purified pernisine was dissolved in 10 mM HEPES, pH 8.0, with 1 mM CaCl 2 , and it was activated for 1 h at 90°C, as determined as the optimised conditions in the azocasein assay. The mutation of Ser355 into Ala (S355A) did not affect the process of pernisine maturation, as seen in Fig 1A, comparing lanes 1 and 2, but have resulted in a complete activity inhibition of pernisine, comparing lanes 1 and 2 or after heat activation lanes 3 and 4 ( Fig 1B). That supports the thesis that Ser355 is a nucleophile involved in a catalytic triade. Thermally induced maturation is already known for other proteases [29], and Ca 2+ is important for enzyme stability at higher temperatures [1,2,30]. Binding sites for Ca 2+ are one of the general adaptations of thermostable enzymes. Such bound Ca 2+ increases the thermostability of the subtilases or protects them from autolysis [30].

Biochemical characterisation of recombinant pernisine
The enzymatic activity of the activated pernisine was determined qualitatively using zymography and quantitatively using the azocasein assay, unless otherwise indicated. Before the enzymatic assays were carried out, the pernisine was heat activated in 10 mM HEPES, pH 8.0, with 1 mM CaCl 2 at 90°C for 1 h, as seen for Fig 1B, comparing lanes 1 and 3. Only wild-type of recombinant pernisine was used in further studies.
The effects of CaCl 2 and/or NaCl on proteolytic activity. Fig 2G illustrates the relative activities of recombinant pernisine evaluated using the azocasein assay (92°C, 20 min, pH 8.0), according to increasing CaCl 2 concentrations from 0 mM to 32 mM. The maximum enhanced relative activity of the recombinant pernisine was observed at 1 mM CaCl 2 . Further increases in the CaCl 2 concentration led to a gradual decrease in the pernisine relative activity. The effects of increasing NaCl concentrations from 0 to 500 mM were also investigated for the relative activity of the heat-activated recombinant pernisine (90°C, 1 h, pH 8.0) both in the absence and presence of 1 mM CaCl 2 . Relative activity of recombinant pernisine is enhanced in the range from 100 to 300 mM NaCl. It is likely that higher ionic strength induces favourable surface-surface electrostatic interactions that are especially important for thermostability of proteins [31]. Pernisine without heat activation was also tested as a negative control (Fig 2H). These data indicate that this pernisine heat activation is Ca 2+ dependent, but is not affected by increased ionic strength with NaCl.
Effects of pH and temperature on enzymatic activity. The enzyme activities of recombinant pernisine at different temperatures (40, 80, 110, 120°C) at pH 7.6 (±0.6) for prolonged incubation times (12, 60, 120, 240 min) were also investigated. This revealed that the pernisine enzymatic activity was retained to at least 80°C (Fig 3). Also, after a 4-h incubation at 120°C, pernisine still retained 40% of its activity.
To investigate the full variability of this pernisine activity, the pH values were corrected according to the temperature change factors (dpH dT -1 ). As illustrated in Fig 4, the three-dimensional representation of the dependence of the pernisine relative activity on temperature and pH shows that recombinant pernisine shows more than 90% relative activity in the pH range from 4.5 to 9.1 and in the temperature range from 90°C to 110°C, with its maximum activity at pH 7.0 and 100°C.
Thus, this recombinant pernisine retains equivalent enzymatic performance under these temperature and pH conditions as the native pernisine, which has shown optimum activity at pH 6.8 and 105°C [2].
Effect of inhibitors and denaturing agents on the enzymatic activity. Different protease inhibitors and denaturing agents were also tested for their effects on the enzymatic activity of recombinant pernisine, as evaluated using azocasein assays. Inhibitors of metallo, serine and cysteine proteases were included in the repertoire, as specified in Table 1. As expected, the greatest inhibitory effects on this pernisine activity were seen for PMSF, EDTA and EGTA, which confirms that this recombinant pernisine is properly folded serine protease that is modulated by Ca 2+ . EDTA and EGTA are chelators of Ca 2+ , which, as expected, resulted in inhibition of the recombinant pernisine activity. Iodoacetamide did not have any relevant effect on this enzymatic activity. Table 2 gives the effects of the various reductants, denaturants and detergent on the enzymatic activity of recombinant pernisine. The effects of the two reductants ( Table 2, DTT, 2-mercaptoethanol) on the recombinant pernisine activity at 1 mM and 5 mM each were similar, with residual pernisine activities of 58% and 48%, and 51% and 38%, respectively. In contrast, the presence of the denaturants ( Table 2, guanidine-HCl, urea) resulted in increased pernisine enzymatic activity above the control. With the addition of 0.1% and 3% SDS, the recombinant pernisine activity showed 9% and 90% inhibition, respectively. The enzymatic activities of recombinant pernisine in the presence of these reductants, denaturants and detergent showed less inhibition compared to those of the native pernisine [2]. The reason for this might arise from the different pathway in the pernisine maturation process. The native pernisine was matured in vivo and was seen as a 36-kDa band as well as a ca. 23-kDa band, which might have resulted from its further processing [2]. In contrast, the recombinant pernisine was matured in vitro and was just seen as a 36-kDa band. Indeed, there remains the possibility that the maturation process of the recombinant pernisine is not fully complete or optimise. Thus, this recombinant pernisine retains an equivalent enzymatic performances against specified inhibitors and better enzymatic performances against denaturing agents as the native pernisine.

Conclusions
In the present study, we have shown that codon optimisation is a key step for the successful expression of pernisine in E. coli from a distant host like Archaea. With codon optimisation using the Genscript algorithm we replaced the codons that are rare for the host with more frequent ones, and we minimised any unfavourable mRNA structures during the translation. This resulted in increased expression levels (up to 10 mg L -1 ), making this recombinant pernisine a potential product for industry. Furthermore, we have shown that mutation of the pernisine aa sequence at the catalytic site (S355A) leads to a complete loss of pernisine activity, as expected. This recombinant pernisine has an N-terminal proregion that is autocleaved during maturation in the presence of CaCl 2 .