Glycosylation influences activity, stability and immobilization of the feruloyl esterase 1a from Myceliophthora thermophila

Heterologous protein production is widely used in industrial biotechnology. However, using non-native production hosts can lead to enzymes with altered post-translational modifications, such as glycosylation. We have investigated how production in a non-native host affects the physicochemical properties and enzymatic activity of a feruloyl esterase from Myceliophthora thermophila, MtFae1a. The enzyme was produced in two microorganisms that introduce glycosylation (M. thermophila and Pichia pastoris) and in Escherichia coli (non-glycosylated). Mass spectrometric analysis confirmed the presence of glycosylation and revealed differences in the lengths of glycan chains between the enzymes produced in M. thermophila and P. pastoris. The melting temperature and the optimal temperature for activity of the non-glycosylated enzyme were considerably lower than those of the glycosylated enzymes. The three MtFae1a versions also exhibited differences in specific activity and specificity. The catalytic efficiency of the glycosylated enzymes were more than 10 times higher than that of the non-glycosylated one. In biotechnology, immobilization is often used to allow reusing enzyme and was investigated on mesoporous silica particles. We found the binding kinetics and immobilization yield differed between the enzyme versions. The largest differences were observed when comparing enzymes with and without glycosylation, but significant variations were also observed between the two differently glycosylated enzymes. We conclude that the biotechnological value of an enzyme can be optimized for a specific application by carefully selecting the production host. Electronic supplementary material The online version of this article (10.1186/s13568-019-0852-z) contains supplementary material, which is available to authorized users.


Figure S1
Examples of MS/MS spectra for an N-glycopeptide Glycopeptide SPNQTCAQGLQKTAQEWGDFVRNAY, (a) from M-Fae, carrying a non-phosphorylated high-mannose glycan: HexNAc(2)Hex(13), (b) from P-Fae, carrying a phosphorylated high-mannose glycan: HexNAc(2)Hex(13)P. Peptide identification was confirmed by the detected series of y-fragment ions along the peptide chain. Hex: hexose. Values are given as a percentage of all observed glycopeptides containing the Asn179 (NQT) glycosylation site.
Results are presented as averages ± one standard deviation. NO: Not Observed. Hex: hexose. Due to differences in ionization efficiency and retention time between high-mannose-and mannose-phosphate-containing glycopeptides, mannose-phosphate-containing glycopeptides were excluded from the calculation of the glycoform distribution. The relative glycoform abundances were calculated as an average of the observed intensity for three chymotriptic peptides: "SPNQTCAQGLQKTAQEWGDFVRNAY", "SPNQTCAQGLQKTAQEWGDFVRNAYAGY", and "GCAAGAESATPFSPNQTCAQGLQKTAQEWGDFVRNAY". Values are given as a percentage of all observed glycopeptides containing the Asn117 (NYT) glycosylation site.
Results are presented as averages ± one standard deviation. NO: Not Observed. Hex: hexose. Due to differences in ionization efficiency and retention time between high-mannose-and mannose-phosphate-containing glycopeptides, mannose-phosphate-containing glycopeptides were excluded from the calculation of the glycoform distribution. The relative glycoform abundances were calculated as an average of the observed intensity for two chymotriptic peptides: "DIQNPDTLTHGQGGDALGIVSMVNYTLDKHSGDSSRVY", "DIQNPDTLTHGQGGDALGIVSMVNY". The chymotrypsin cleavage efficiency seemed to be strongly impaired by the presence of large high-mannose structures close to its cleaving site. This could explain some of the large standard deviations observed.

Figure S2
Enzyme kinetics for the three MtFae1a versions (a) M-Fae, (b) P-Fae, (c) E-Fae. V0 (initial velocities) was plotted depending on the initial substrate concentration (MpCA). Activities were measured at 35°C in a continuous assay. The data were first fitted to the Michaelis-Menten equation, Equation (1), allowing the determination of Km and Vmax. If the data did not fit Equation (1), substrate inhibition was taken into account. This was done by fitting the data to the substrate inhibition equation (Equation (2)) which returned the parameter Ksi in addition to Km, and Vmax. kcat was then calculated by dividing Vmax by the initial molar concentration of the enzyme.
V is the measured initial reaction rate and [S] is the substrate concentration. (a) and (b) were fitted using the Michaelis-Menten substrate inhibition equation (Equation (2)), (c) was fitted using the Michaelis-Menten equation (Equation (1)). Results are presented as the mean values of triplicate experiments, and error bars represent one standard deviation.

Figure S3 Data points and fitting curves used during non-linear regression for calculations of the melting temperature (a) M-Fae. (b) P-Fae. (c)
E-Fae. Black: data points, blue: 95% confidence band, red: 95% prediction band. AU: Arbitrary Units. The data, from the relevant part of the curve, were fitted by non-linear regression to a 4parameter sigmoid curve (Equation (3)) using SigmaPlot software.
Where, y0 is the minimum fluorescence intensity, a is the difference between the maximum and minimum fluorescence intensities measured, b is the slope of the curve around Tm, and x0 = Tm. Results are presented as the mean values of triplicate experiments, and error bars represent one standard deviation.

Figure S4 Visualization of the amino acids forming the catalytic triad and of the two glycosylated asparagine residues on a homology model (a)
The overall structure showing the proposed dimer biological unit (grey and copper colors). (b) The location of glycosylated Asn117 (green) is predicted to be at the dimer interface and symmetry axis. (c) Rotation of the overall structure (90° with respect to a). (d) The glycosylated Asn179 (green) is predicted to be close to the catalytic residues Ser136, His275 and Asp219 (magenta), dotted lines represent distance (in Angstroms) between residues. The homology model was deposited by Topakas et al. (Topakas et al. 2012) and recovered from the Swiss Model repository (SMR ID: G2QND5). Visualization and images were obtained using the PyMol software (Schrödinger). .9 ± 0.7 1.5 ± 0.7 3.3 ± 4.7 90.2 ± 6.1 6.5 ± 1.4 Values given are the percentage of all observed peptides/glycopeptides carrying the potential glycosylation site.
Results are averages ± one standard deviation. NG: Non-Glycosylated, High-man: high-mannose structure, man-P: mannose-phosphate containing structures, NO: Not Observed. Due to different ionization efficiency and different retention times among non-glycosylated as well as high-man-and man-P-containing glycopeptides, only estimates of the relative distribution between these groups were possible for each site.   7.0 0.8 ± 0.6 Immobilization was performed in phosphate citrate at 20°C for 24 h. Activity data were obtained in sodium phosphate buffer (pH 7.0) at 35°C, using the standard stopped assay. Results given are the average of three experiments ± one standard deviation. Phosphate citrate buffer at pH 6.5-7.5 (open squares), sodium phosphate buffer at pH 7.0-8.0 (filled triangles), and bicine buffer at pH 8.0-8.5 (open circles). pH profiles were determined after immobilization in sodium phosphate buffer at pH 6.0 by measuring the activity at 35°C, using the standard stopped assay. The activities obtained are presented as a percentage of the highest activity observed. Results are the average of triplicate experiments, and error bars represent one standard deviation.