Identification of potential residues to be substituted
There is no universal procedure for the thermostabilization of an enzyme. However, in this research, we follow a comprehensive strategy that combines different principles contributing to thermostability. For this, the B factor profile of the crystalline structures 4hox, 4how and 4hoz (Xu et al., 2013) of the sucrose isomerase PalI NX-5 was analyzed. Thus, 19 amino acid residues with B values between 2.5 and 3.5 times the average of the molecule (average B factor: 16.5) were identified (Supplementary Table 1). Using the ROSETTA DESIGN software, these residues were subjected to individual substitutions. It has been observed that the prevalence of hydrogen bond networks in mesophilic enzymes is lower than in thermophilic ones (Tompa et al., 2016). In addition, the hydrogen bond is the most referenced physicochemical interaction to explain the increase in thermostability (Querol et al., 1996). For this reason, the appearance of potential interactions involving the mutant residues, or their neighborhood, was analyzed in silico. Only in the substitutions Leu202Glu, Lys174Gln, and Asp195Glu new interactions or reinforcement of the existing ones were observed.
As shown in Fig. 1a, when Glu occupied position 202, a new interaction between Glu202 and Lys248 is predicted. A restructuring of the hydrogen bond network is also observed in the vicinity of Asp211, which now interacts with Ala190 instead of Tyr204. The readjustment observed in the hydrogen bonds network could imply an increase in the thermal stability of the molecule since, in addition to the appearance of a new interaction (Glu202–Lys248), the network now encompasses more distant residues belonging to elements different from the side-chain. When comparing the neighborhood of the Lys174 and Gln174 residues (Fig. 1b), no changes were observed in the interaction distances with the His177 residue. However, a new interaction between Asp175 and Gln174 was predicted. For the Asp195Glu substitution (Fig. 1c), a repositioning of the Asp194-Ser197 hydrogen bond was observed, settling between Asp 194 and Lys 196. Associated with this restructuring, a reduction in the interaction distance was predicted, going from 2.5 Å to 2.1 Å. Precisely, Zheng et al. (2020) reported that for the occurrence of hydrogen bonds, the distance between the proton acceptor and the hydrogen atom must be less than the sum of the Van der Waals atomic radii of the two atoms in question; in the case of the oxygen-hydrogen combination, this distance is 2.5 Å (Pauling, 1960). Thus, the reduction in the distance of the hydrogen bond would imply a strengthening of this interaction.
PLACE FIG. 1 HERE.
Two elements that are interesting up to this point are: first that the substitutions for which interactions were observed do not correspond to the residues that presented the highest values of B factor, such as Asp175 and His177, confirming that the thermal stability is achieved through cooperative optimization of several factors rather than predominant interaction. Second, of the three substitutions where new interactions were observed in the neighborhood, one corresponds to Gln, a residue on which the protein statistics of Khan et al. (2019) indicate a poor prevalence in thermophilic molecules, while Warren et al. (1995) reported that the presence of Gln in helices of thermostable proteins is greater than in mesophilic proteins. Given this ambiguity, the practical evaluation will be conclusive.
The conservation of the enzyme-substrate interaction was verified in silico in the mutants Leu202Glu, Lys174Gln, Asp195Glu, and their combinations. In molecular docking analysis at the catalytic site, glucose was used instead of sucrose. It was taken into account that Véronèse et al. (1999) showed that glucose is a competitive inhibitor of sucrose; and that Véronèse et al. (1998) previously deduced the glycosyl binding of sucrose to the active site, assuming that free glucose binds to the enzyme at the same site and in the same way. The highest affinities (Viewed as |Energy < 0|) enzyme-glucose were predicted for the L202E, L202E/K174Q, K174Q/D195E, and L202E/D195E variants (Table 1). These variants presented interaction energies that were 6% higher than the native enzyme; the highest increase observed among the 1440 configurations analyzed. Based on the in silico data shown in Table 1, the mutants selected for recombinant expression and subsequent evaluation were the L202E, L202E/K174Q, and K174Q variants. Figure 2 shows the predicted interactions between D-glucose and the three variants L202E, L202E/K174Q, and K174Q, including the native enzyme. We were able to identify interactions between glucose and residues R325, D241, E295, D369, and H368. Zhang et al. (2003) have previously reported the catalytic function of these residues. These results suggest the potential conservation of the enzyme-substrate interaction.
Table 1
Interaction energies at the catalytic site of sucrose isomerase PalI NX-5 variants and the D-glucose molecule. Default docking parameters were used, and point charges were initially assigned according to the AutoDock semi-empirical force field.
Variant | Affinity (kcal/mol) | Condition |
Native | -6.3 | Seed: 11, state: 1 |
L202E | -6.7 | Seed: 1, state: 1 |
K174Q | -6.5 | Seed: 7, state: 1 |
D195E | -6.3 | Seed: 11, state: 2 |
L202E/K174Q | -6.7 | Seed: 17, state: 1 |
K174Q/D195E | -6.7 | Seed: 10, state: 1 |
L202E/D195E | -6.7 | Seed: 10, state: 1 |
K174Q/L202E/D195E | -6.3 | Seed: 1, state: 1 |
PLACE Table 1 HERE.
PLACE FIG. 2 HERE.
Potential N-glycosylation of sucrose isomerase variants
Through Prosite (https://prosite.expasy.org/) (De Castro et al., 2006), we evaluated the potential presence of post-translational modifications that influence thermostability. Three potential sites to be N-glycosylated were identified in the sucrose isomerase PalI NX-5 variants. The first of these sites is found at position 144NHT146, including the catalytic residue H145, just inside the catalytic pocket. Taking into account that in the "endoplasmic reticulum – Golgi apparatus" pathway, the previously assembled glycans are added to the growing polypeptide chain (Colley et al., 2015), it would be expected that the glycosylation of residue N144 leads to obtaining of a glycoform with a non-functional tertiary structure. However, if the recombinant expression of sucrose isomerase PalI NX-5 were mediated by the α factor secretion signal from Saccharomyces cerevisiae, the translocation to the endoplasmic reticulum would be post-translational (Barrero et al., 2018), so the protein it would fold in the cytosol, and residue N144 would not be available to be glycosylated. The remaining two potential N-glycosylation sites (530NNS532 and 573NNS575) are located on the surface of the C-terminal domain of the molecule. Fortuitously, the position of these consensus sequences for glycosylation is very close to hot spots of high B value. Specifically, the potential site 530NNS532 is located in ten residues from the 541GAE543 hot spot, while the site 573NNS575 is within the high B factor segment from residue E553 to residue R587.
Reports such as those by Colley et al. (2015), Katla et al. (2019), Benoit et al. (2006) and (Zou et al., 2013) agree that the presence of glycosylation in a protein constitutes, in a general sense, a contribution to the stability of the molecule, whether thermal, against pH, presence of proteases or stress physiological. Specifically, Benoit et al. (2006) isolated a feruloyl esterase A from Aspergillus niger (naturally glycosylated), which is more thermostable than its non-glycosylated counterpart expressed in E. coli. On the other hand, Zou et al. (2013) observed that the structural and functional stability of β-glucuronidase from Penicillium purpurogenum was improved after its glycosylated expression in P. pastoris GS115 compared to its wild-type counterpart. However, as evidenced in the works of Han et al. (2020) y Hu et al. (2019), the effect of this post-translational modification is very complex and not always predictable or beneficial.
It has been observed that the stabilization caused by glycosylation is closely associated with the entropy and with the positions of the glycosylation sites in the protein (Helenius et al., 2004; Shental-Bechor et al., 2008). Glycans bound to flexible regions, in general, would confine the conformational space and stimulate protein entropy reduction, thus improving conformational stability at high temperatures (Adney et al., 2009; Dotsenko et al., 2016). Taking this context into account, it is possible to hypothesize that, since residues N530 and N573 are found in loop segments with high B factor, the addition of glycans to these positions will have a positive effect on stability compared to what we would expect for the wild-type variant of E. rhapontici NX-5 (not glycosylated), and even complement the substitutions made at positions K174Q and L202E.
Site-directed mutagenesis and expression vectors
We started from the amino acid sequences of the four variants of sucrose isomerase PalI NX-5 (native, L202E, K174Q, and L202E/K174Q) previously analyzed. Using the Sequence Manipulation Suite software (Stothard, 2000), the reverse translation of each of these sequences was carried out, for which the usage codons of P. pastoris were used. The variants of the palI NX-5 gene were inserted into the genomic integration vector pPICZα A (INVITROGEN®), thus obtaining the constructions pSINX5n, pSINX5-L202E, pSINX5-K174Q and pSINX5-L202E-K174Q (Supplementary Fig. 1). These constructions were used for the transformation of P. pastoris X33 and consequently enabling the glycosylation of the variants of the sucrose isomerase PalI NX-5.
Expression analysis of sucrose isomerase PalI NX-5 variants
P. pastoris X33 colonies transformed with the pSINX5n, pSINX5-L202E, pSINX5-K174Q, and pSINX5-L202E-K174Q constructs were cultured in a protein induction medium (see “Materials and methods”), the supernatant was collected, dialyzed and purified using a high-affinity Ni-Charged Resin (GenScript®). Aliquots of purified sucrose isomerase PalI NX-5 variants were analyzed by SDS-PAGE. After Coomassie blue staining, the presence of bands at approximately 70 kDa was confirmed (Fig. 3a, lanes 3, 4, 5, and 6), close to the molecular weight predicted by Protein Calculator v3.4 (http://protcalc.sourceforge.net/). Interestingly, at least two bands were observed in the purified sucrose isomerase PalI NX-5 (Fig. 3a, lanes 3, 4, 5, and 6). One hypothesis for this behavior could be the capacity of P. pastoris to glycosylate secreted proteins. Thus, the bands observed for each of the purified PalI NX-5 variants could correspond to different levels of glycosylation reached during cell cultures.
PLACE FIG. 3 HERE.
To qualitatively verify the presence of glycoproteins in the purified samples, the Glycoprotein Detection Kit (Sigma-Aldrich®) was used. In Fig. 3b, the protein bands with glycans attached to their surface can be seen stained in purple. When comparing the migration profiles of the purified ones (Lanes 3, 4, 5, and 6) shown in Fig. 3a and 3b, it is verified that the observed bands correspond to glycoproteins. Thus, at least two main glycoforms of the sucrose isomerase PalI NX-5 variants would be obtained, this being the first report of expression and characterization of glycosylated sucrose isomerases.
Influence of temperature on the activity and stability of recombinant sucrose isomerase PalI NX-5 variants
The influence of temperature on the isomerase activity of PalI NX-5 variants was studied in the range of 15 to 68°C (Fig. 4a). The native variant of PalI NX-5 showed its maximum activity at 30°C, coinciding with its counterparts: wild-type (from E. rhapontici NX-5) (Ren et al., 2011) and recombinant (from E. coli BL21(DE3)) (Li et al., 2011). Thus, the glycosylated expression of the native variant apparently had no influence on the optimum temperature observed. In this sense, Hua et al. (2014), Han et al. (2014b) and Petrescu et al. (2004) have documented that N-glycosylation of recombinant enzymes expressed in yeast may not have significant effects on activity. Likewise, Han et al. (2014a) reported that some N-glycosylation sites are not essential for folding or cannot affect enzymatic activity due to their distance from the active site of the recombinant enzyme; thus, the presence (or absence) of the glycans has a minimal effect on enzyme activity.
PLACE FIG. 4 HERE.
The K174Q, L202E, and K174Q/L202E mutants showed an increase in the optimal temperature of 5°C (Optimal temperature: 35°C). Furthermore, its relative activity at 40°C was 96% of the observed optimum, while that of the native variant was only 87% of its optimum. Thus, the simple effect of the mutations made on the PalI NX-5 isoform, or its combination with the glycosylation of the molecule, resulted in the improvement of this indicator parameter of thermostability. The improvements achieved in the activity profile versus temperature (Fig. 4a) are not only important in the context of the PalI NX-5 isoform since most sucrose isomerases have a temperature optimum between 30 and 35°C (Mu et al., 2014; Zhan et al., 2020; Zhang et al., 2019; Zhang et al., 2018).
The thermostability of the PalI NX-5 variants was evaluated by incubating the enzymes in 50 mM sodium phosphate/citric acid buffer (pH 6.0), at a temperature of 40°C. Their residual activities were then tested at different incubation times. As shown in Fig. 4b, the thermostabilities of the mutants were higher than that shown by the native enzyme, with the K174Q/L202E variant showing the best performance. At 40°C and pH 6.0, the half-life of the native variant was 10.1 minutes. In comparison, the half-lives of the K174Q, L202E, and K174Q/L202E variants were 22.3 min, 17.5 min, and 29.2 min, respectively, which are 2.21, 1.73, and 2.89 times higher than those of the native variant. Both the increase observed in the optimal temperature of the mutants (Fig. 4a) and half-life times (Fig. 4b) confirm that the comprehensive strategy outlined in this research to achieve thermostabilization of sucrose isomerase PalI NX-5 was successful.
Two central factors that influence the thermal stability of a protein are the increase of hydrogen bonds (Bi et al., 2020; Masakari et al., 2020; You et al., 2019) and salt bridges (Bian et al., 2015). For this reason, we analyzed the possible appearance of these interactions and observed that associated with the K174Q and L202E substitutions was a reinforcement of the bonds network. Probably, the optimization of the bonds network and its combination with the rigidity provided by N-glycosylation are the keys to the improvement achieved in the thermostability of sucrose isomerase PalI NX-5. Furthermore, with the L202E substitution, we removed an apolar residue, replacing it with a polar/charged one. Khan et al. (2019), when analyzing the relative abundance of amino acids in proteins, observed that charged residues (Charged residues > Glu > Basic residues, abundance) had the highest abundance in thermophilic proteins than in mesophilic ones. In this sense, Chakravorty et al. (2017) state that charged amino acids are involved in the appearance of ionic interactions and the formation of salt bridges, hence their positive contribution to thermostability. Also, Wang et al. (2020) confirmed the role of salt bridges in enhancing the thermal stability of the r27RCL enzyme from Rhizopus chinensis; precisely, the appearance of a new salt bridge between residues Glu292 and His171 of the mutant m31 was one of the factors that conditioned the improvement of the parameters indicating thermostability. Likewise, Vicente et al. (2020), after implementing the Ser264Lys substitution, refer to the appearance of a new salt bridge (with the Asp205 residue) as a stabilizing interaction in a laccase variant with high redox potential from white rot fungi.
Influence of pH on the activity of recombinant sucrose isomerase PalI NX-5 variants
The study of the influence of pH on the activity of recombinant sucrose isomerase PalI NX-5 variants was carried out for pH between 4 and 9 (Fig. 4c). The native PalI NX-5 variant showed its maximum activity at pH 6.0, which was consistent with its wild-type counterpart expressed in E. rhapontici NX-5 (Ren et al., 2011). In general, the activity profiles vs. pH of native and wild-type PalI NX-5 were found to be very similar. These results indicate that glycosylation of the molecule did not significantly affect the activity profile against pH, which was also observed by Han et al. (2020) during the improvement of the thermal stability of the endoglucanase CTendo45 from Chaetomium thermophilum.
The mutants and the native variant showed an optimal pH value of 6.0. In addition, the substitutions made did not substantially affect the activity profile against pH. Analogous behavior was observed by Duan et al. (2016) during the thermostabilization of sucrose isomerase PalI AS9. Similarly, Wang et al. (2020) did not observe changes in the profile of enzymatic activity against pH after making the substitutions S142A, D217V, Q239F, and S250Y, in the lipase r27RCL from Rhizopus chinensis. However, Wang et al. (2020) found differences when analyzing the stability of the r27RCL lipase variant against pH. Also Khan et al. (2019), during the thermostabilization of the mesophilic lipase of Bacillus subtilis, reported changes in the enzymatic activity profile against pH.
Few research addresses mutagenesis directed at changing the optimal pH of enzymes. In general, the proposed strategies focus on modifying the surface charges of the molecule (Yang et al., 1993), changing the values of the acid dissociation constants of the catalytic residues (Dey et al., 2018; Li et al., 2019), and cavity filling (Nielsen et al., 2000). In particular, Shi et al. (2022) replaced amino acids near catalytic residues to modify the optimal pH of the β-galactosidase from Aspergillus oryzae. For this, it was taken into account that in enzymes with acid-base catalysis, such as sucrose isomerases (Mu et al., 2014), the residues of the active center are required to be in a catalytically adequate protonation state. Thus, a possible explanation of the behavior observed in Fig. 4c is the fact that the substitutions made (K174Q and L202E) are far from the catalytic residues (Glu277, proton donor, and Asp223, nucleophile) without affecting their protonation state.
Figure 4c shows that only at pH higher than 8 or lower than 5 considerable activity losses occur. This behavior, together with the observed improvement in thermostability, gives the PalI NX-5 mutants great flexibility in industrial applications. Be it the case in the context of a sugar mill, where operating fluctuations demand flexible biocatalysts, or in mitigating the risk of microbial contamination by adjusting the operating pH of an enzymatic reactor.
Evaluation of the kinetic parameters of the recombinant sucrose isomerase PalI NX-5 variants
The kinetic parameters of the recombinant PalI NX-5 sucrose isomerase variants were measured at 30°C. As shown in Table 2, compared to native PalI NX-5, the Km values of K174Q, L202E, and K174Q/L202E mutants decreased by 5.2%, 7.9%, and 9.4%, respectively. This implies that, concerning the native variant, the mutants present an increase in the affinity for the substrate. The PalI NX-5 variants analyzed in this investigation showed Km values in the same order of magnitude as their counterparts: wild-type (Ren et al., 2011) and recombinant (Li et al., 2011). Furthermore, regarding the report by Ren et al. (2011), the native variant and the mutants K174Q, L202E, and K174Q/L202E increased their specific activity by 14.4%, 25.3%, 20.3%, and 20.4%, respectively. Finally, the mutants showed improved catalytic constants (kcat) and catalytic efficiencies (kcat/Km). Remarkably, the catalytic efficiency of the mutants increased by up to 16%.
Table 2
Evaluation of the kinetic parameters of sucrose isomerase PalI NX-5 variants.
Variant | Specific activity (U/mg) | Km (mM) | kcat (s− 1) | kcat /Km (s− 1·mM− 1) | References |
native PalI NX-5 | 483.8 ± 6.9 | 255.1 ± 9.6 | 564.5 ± 1.0 | 2.21 ± 0.13 | This investigation |
PalI NX-5 K174Q | 529.9 ± 8.4 | 241.9 ± 6.8 | 618.2 ± 0.9 | 2.55 ± 0.15 |
PalI NX-5 L202E | 509.1 ± 9.9 | 234.9 ± 5.1 | 594.0 ± 0.8 | 2.52 ± 0.08 |
PalI NX-5 K174Q/L202E | 509.3 ± 8.7 | 231.2 ± 7.8 | 594.3 ± 1.1 | 2.57 ± 0.21 |
wild-type PalI NX-5 | 423 | 222 | NR | NR | (Ren et al., 2011) |
Recombinant PalI NX-5 | NR | 257 | NR | NR | (Li et al., 2011) |
PLACE Table 2 HERE.
The catalytic performance of the mutants was improved. Ren et al. (2020), Duan et al. (2016), Pang et al. (2020), Teng et al. (2019), and Ban et al. (2020) have shown that enzyme activity can be increased when thermostability is improved. Taking this into account, works such as that of Liu et al. (2021) have focused their attention on increasing enzyme activity using optimization strategies for thermal stability. In particular, Duan et al. (2016) justified the increase in activity of the sucrose isomerase PalI AS9 by stating that the substitutions made (E175N and K576D) were far from the catalytic center and the isomerization region. Thus, the mutations could have caused these regions to be more compact, producing a positive effect on the kinetic parameters. The hypothesis that would be applicable in this research since both the substitutions made in the PalI NX-5 (K174Q and L202E), and the potential sites for N-glycosylation (530NNS532 and 573NNS575) are located on the periphery of the molecule, away from the active site. As can be observed in Table 2 the designed mutants present an optimized catalysis with respect to both the native PalI NX-5 variant and the wild-type expressed in E. rhapontici NX-5. These results demonstrate that the outlined comprehensive approach can improve thermostability without sacrificing catalytic activity.
The characteristics of the K174Q, L202E, and K174Q/L202E mutants, together with an adequate immobilization and reaction strategy, would allow the reuse of the biocatalyst and, therefore, the reduction of the costs associated with the operation (Choi et al., 2015; Cicerone et al., 2015; Kazlauskas, 2018), thus influencing the stigma of costs when working with enzymes (Wu et al., 2020). In addition, an increase in the operating temperature would lead to a decrease in the viscosity of the reaction mixture, higher solubility of sucrose, and an increase in the mass transfer rate without drastically affecting the activity of sucrose isomerase during the production of isomaltulose. Another positive consequence would be reducing the risk of microbial contamination (Singh et al., 2015; Turner et al., 2007), elements that will make the operation of an enzyme reactor more effective and efficient.