Effect of Leu²⁷⁷ on Disproportionation and Hydrolysis Activity in Bacillus stearothermophilus NO2 Cyclodextrin Glucosyltransferase

ABSTRACT

The disproportionation activity of cyclodextrin glucosyltransferase (CGTase; EC 2.4.1.19) can be used to convert small molecules into glycosides, thereby enhancing their solubility and stability. However, CGTases also exhibit a competing hydrolysis activity. The +2 subsite of the substrate binding cleft plays an important role in both the disproportionation and hydrolysis activities, but almost all known mutations at this site decrease disproportionation activity. In this study, Leu²⁷⁷ of the CGTase from Bacillus stearothermophilus NO2, located near both the +2 subsite and the catalytic acid/base Glu²⁵³, was modified to assess the effect of side chain size at this position on disproportionation and hydrolysis activities. The best mutant, L277M, exhibited a reduced K_m for the acceptor substrate maltose (0.48 mM versus 0.945 mM) and an increased k_cat/K_m (1,175 s⁻¹ mM⁻¹ versus 686.1 s⁻¹ mM⁻¹), compared with those of the wild-type enzyme. The disproportionation-to-hydrolysis ratio of L277M was 2.4-fold greater than that of the wild type. Existing structural data were combined with a multiple-sequence alignment and Gly²⁸² mutations to examine the mechanism behind the effects of the Leu²⁷⁷mutations. The Gly²⁸² mutations were included to aid a molecular dynamics (MD) analysis and the comparison of crystal structures. They reveal that changes to a hydrophobic cluster near Glu²⁵³ and the hydrophobicity of the +2 subsite combine to produce the observed effects.

IMPORTANCE In this study, mutations that enhance the disproportionation to hydrolysis ratio of a CGTase have been discovered. For example, the disproportionation-to-hydrolysis ratio of the L277M mutant of Bacillus stearothermophilus NO2 CGTase was 2.4-fold greater than that of the wild type. The mechanism behind the effects of these mutations is explained. This paper opens up other avenues for future research into the disproportionation and hydrolysis activities of CGTases. Productive mutations are no longer limited to the acceptor subsite, since mutations that indirectly affect the acceptor subsite also enhance enzymatic activity.

INTRODUCTION

As can be seen from the CAZy database, the α-amylase family (glycoside hydrolase family 13 [GH13]) is large (1). It includes the α-amylases, pullulanases, isoamylases, cyclodextrin glucosyltransferases (CGTases), branching enzymes, α-glucosidases, oligo-1,6-glucosidases, amylopullulanases, and neopullulanases (2). These enzymes catalyze two reactions, hydrolysis and transglycosylation, albeit with different ratios (3). All of the enzymes in GH13 share similar catalytic region structures (3) and a common catalytic mechanism (Fig. 1) (2). The catalytic reaction can be divided into two steps. In the first step, a donor substrate reacts with a nucleophilic aspartic acid residue in the active site to produce a covalent enzyme-substrate intermediate. In the second step, an acceptor substrate reacts with the covalent intermediate to produce the product (3). If the acceptor substrate is water, the reaction is hydrolysis; if it is another sugar molecule, the reaction is transglycosylation (4). The ratio between these two activities is related to the conformation of the catalytic residues (5) and the active site’s acceptor affinity (6).

FIG 1.

Catalytic mechanism of hydrolysis and transglycosylation in GH13.

α-Amylase and CGTase are both members of GH13. They are very similar in sequence and are thought to have evolved from a common ancestor (7). For example, the α-amylase and CGTase from Bacillus share 87% identity (8). However, the α-amylases catalyze primarily hydrolysis and CGTases catalyze primarily transglucosylation (9). Interestingly, mutation of several residues can change the ratio between hydrolysis and transglucosylation activities. For example, the Bacillus licheniformis α-amylase mutant H235E (10) exhibits enhanced transglucosylation activity, from none to 47.3 s⁻¹, and the hydrolysis activity of the F183S/F259N mutant of Bacillus circulans 251 CGTase is enhanced 4.5-fold (6).

Unlike the almost exclusive hydrolysis activity of α-amylase, CGTase has three transglucosylation reactions: disproportionation, cyclization, and coupling (4). The disproportionation reaction is a transglucosylation that occurs between two different molecules. The cyclization reaction is an intramolecular transglucosylation, and the coupling reaction is the reverse of the cyclization reaction (6). In practical applications, the disproportionation reaction is often used to generate a variety of small-molecule glycosides with improved water solubility and enhanced stability, such as the glycosylation products (11) of l-ascorbic acid (12), hesperidin (13), and naringin (14). However, all known natural CGTases exhibit both disproportionation and hydrolysis activities, and the hydrolysis reaction reduces disproportionation product yield (4). To increase industrial production of these glycosides, the hydrolysis activity should be reduced as much as possible.

CGTases can be most efficiently modified to improve their disproportionation-to-hydrolysis ratio through the use of structure-based genetic engineering. Many crystal structures of CGTases and CGTase-substrate complexes have been published to date; for example, those of complexes formed by B. circulans 251 CGTase and maltononaose (PDB identifier [ID] 1CXK) (15) or maltotetraose (PDB ID 1CXH) (16). The CGTases have five structural domains, including a catalytic domain (domain A). Studies have identified nine substrate binding subsites known as subsites −7 to +2. Subsites +1 and +2 are acceptor subsites, while subsites −1 to −7 are donor subsites (17, 18). The catalytic residues Asp²²⁵, Glu²⁵³, and Asp³²⁴ are positioned between subsites −1 and +1 (19). Among GH13 family members, the catalytic acid/base Glu exhibits two distinct conformations: one oriented far away from −1 subsite and one oriented close to the −1 subsite (5). Relevant reports indicate that this Glu may affect the balance between disproportionation and hydrolysis activities (5, 20).

Since the main difference between disproportionation and hydrolysis is found in the second step of the catalytic mechanism, the key interactions affecting disproportionation and hydrolysis occur at the acceptor subsite (4). Consistent with this, the majority of mutants that affect disproportionation and hydrolysis activity are located at the +1 and +2 subsites (21). In particular, conserved hydrophobic residues Phe and Phe/Tyr at the +2 subsite seem to have the most significant effects (6). Changing either of these amino acid residues greatly reduces the disproportionation activity of the CGTase (20). For example, Bacillus sp. strain 1011 F259L and F283L CGTase mutants display disproportionation activities that are 12.5% and 4.03% that displayed by their wild-type parent, respectively. Their hydrolysis activities are 40.05% and 35.67% that displayed by their wild-type parent, respectively (22). Similarly, mutants of B. circulans 251 CGTase with changes at residues Phe¹⁸³ and Phe²⁵⁹ display disproportionation activities that are 17.56% to 57.35% that displayed by their wild-type parent and hydrolysis activities that are 2.75 to 18.84 times that displayed by their wild-type parent (6, 23). Versions of Thermoanaerobacterium thermosulfurigenes EM1 CGTase with mutations at Phe¹⁸⁴ and Phe²⁶⁰ display disproportionation activities that are 0.20% to 45.10%, and hydrolysis activities that are 31.48% to 327.78%, those displayed by their wild-type parent (24, 25). The disproportionation activity of the Y260R mutant of Paenibacillus macerans JFB05-01 CGTase is 138% of that displayed by its wild-type parent, while its hydrolysis activity is 226% that displayed by its wild-type parent (26). Based on these observations, optimization of the spatial structure of the +2 subsite may be the key to further increase disproportionation activity and decrease hydrolysis activity. It is generally believed that Phe¹⁷⁹ and Phe²⁵⁵ (Bacillus stearothermophilus NO2 CGTase numbering) provide a hydrophobic environment in the acceptor site that promotes the binding of sugar molecules and reduces the binding of water (6). The Y179A, Y255A, and Y179A/Y255A mutants showed significant decreases of the disproportionation-to-hydrolysis ratio (see Table S1 in the supplemental material). So the key to improving the disproportionation activity is to promote the binding of sugar molecules at the +2 subsite by enhancing the hydrophobic environment of the +2 subsite, and mutations at the +2 subsite should be avoided. Instead, other residues that might indirectly affect the spatial structure of this acceptor subsite were considered.

In this study, we selected the CGTase from B. stearothermophilus NO2 (4), which has high disproportionation activity. On the basis of its crystal structure and a multiple-sequence alignment, Leu²⁷⁷, which is near the +2 subsite, and Glu²⁵³ were chosen as targets for site-directed mutagenesis. We found that the disproportionation-to-hydrolysis ratio of the mutant L277M was 2.4-fold greater than that of the wild type. The mechanism through which the Leu²⁷⁷mutants influence CGTase activity was investigated using reaction kinetic parameters and molecular dynamics (MD) simulations. Due to the common catalytic mechanism of GH13 members, the results of this study may also provide the basis for further exploration of the key factors influencing hydrolysis and transglycosylation, allowing useful modification of α-amylase and other GH13 members.

RESULTS

Selection of mutation sites based on multiple-sequence alignment and three-dimensional structure.

To be selected for mutagenesis, the target residue had to be located near the +2 subsite (the Cα of the target residue must be within 10 Å of the Cα of Phe²⁵⁵), located near the catalytic acid/base Glu²⁵³ (the Cα of the target residue must be within 10 Å of the Cα of Glu²⁵³), and ≤30% conserved (the residue is not conserved, so enzyme activity is unlikely to change drastically due to mutation at the site) within the GH13 family (27). To find residues that have the potential to affect the disproportionation-to-hydrolysis ratio of CGTase and follow these three criteria, a part of CGTase sequence (domain A) from Bacillus stearothermophilus NO2 was submitted to the GREMLIN (Generative REgularized ModeLs of proteINs) server (http://gremlin.bakerlab.org). GREMLIN searched Pfam 27.0 and UniProt and retrieved the amino acid sequences to be aligned using the selection principles (coverage of 75% and identity of 90%). The output, a multiple-sequence alignment, is presented in Fig. 2C. Using the three-dimensional structure of CGTase (Fig. 2A and B) and the multiple-sequence alignment, the potential target sites were narrowed until Leu²⁷⁷ was finally selected. The multiple-sequence alignment revealed that several residues (Ile, Met, Phe, Tyr, Trp, and Thr) are found at position 277 (Fig. 2C). Most of these amino acids are hydrophobic (Ile, Met, Phe, Tyr, and Trp); only one is considered hydrophilic (Thr). The hypothesis was that mutating Leu²⁷⁷ to one of the other hydrophobic amino acids would increase disproportionation activity and/or reduce hydrolysis activity, thereby increasing the disproportionation-to-hydrolysis ratio. To test this hypothesis, Leu²⁷⁷ was converted into Ala, Thr, Ile, Met, and Phe.

FIG 2.

Selection of CGTase mutation sites. (A) Diagram of the X-ray crystal structure of CGTase from B. stearothermophilus NO2 (PDB ID 1CYG). Glu²⁵³ is shown in pink, and key amino acid residues Phe¹⁷⁹ and Phe²⁵⁵ in the +2 subsite are shown in blue. (B) A maltotetraose molecule is modeled into 1CYG. Leu²⁷⁷ is shown in yellow. The +1 to +2 subsites are shown in light blue, and the −1 to −2 subsites are shown in white. (C) Multiple-sequence alignment of GH13 members. The red box indicates the site of mutation.

Construction, expression, and purification of the wild type and Leu²⁷⁷ variants.

We produced the wild-type CGTase and its L277A, L277T, L277I, L277M, and L277F mutants in Escherichia coli BL21(DE3) and purified the desired proteins from the culture supernatant using immobilized-metal affinity chromatography (nickel column). Circular-dichroism results (Fig. S2) showed no significant differences in secondary structure among these CGTase variants. The purity and molecular weight of each protein were analyzed using SDS-PAGE (Fig. S3). The results show that each protein was essentially a single band with a molecular weight of about 66.2 kDa, indicating that the proteins were essentially pure. Purification data (Table 1) showed that the disproportionation specific activities of the wild type and the L277A, L277T, L277I, L277M, and L277F mutants were 496.8 U mg⁻¹, 157.1 U mg⁻¹, 148.9 U mg⁻¹, 346.2 U mg⁻¹, 359.4 U mg⁻¹, and 288.9 U mg⁻¹, respectively.

TABLE 1.

Summary of CGTase isolation and purification procedures^a

Enzyme	Purification step	Total activity (U)	Total protein (mg)	Sp act (U mg−1)	Purification (fold)	Yield (%)
Wild type	Crude enzyme	39,211.03	359.20	109.2	1.00	100.00
	Nickel column	4,120.72	8.29	496.8	4.55	10.51
L277A mutant	Crude enzyme	9,307.59	337.79	27.55	1.00	100.00
	Nickel column	167.83	1.07	157.1	5.70	1.80
L277T mutant	Crude enzyme	14,734.71	337.69	43.63	1.00	100.00
	Nickel column	946.98	6.36	148.9	3.41	6.43
L277I mutant	Crude enzyme	30,686.90	263.55	116.44	1.00	100.00
	Nickel column	6,001.66	17.34	346.2	2.97	19.56
L277M mutant	Crude enzyme	31,514.48	385.95	81.65	1.00	100.00
	Nickel column	7,879.45	21.92	359.4	4.40	25.00
L277F mutant	Crude enzyme	27,183.45	308.36	88.15	1.00	100.00
	Nickel column	3,468.14	12.01	288.9	3.28	12.76
G282A mutant	Crude enzyme	24,866.21	248.58	100.0	1.00	100.00
	Nickel column	12,339.86	32.50	379.7	3.80	49.63
G282A/L277M mutant	Crude enzyme	32,383.45	234.97	137.8	1.00	100.00
	Nickel column	13,082.48	26.83	487.6	3.54	40.40
G282A/L277F mutant	Crude enzyme	17,804.14	204.53	87.05	1.00	100.00
	Nickel column	9,881.93	26.02	379.7	4.36	55.50

^aEnzyme activity was measured using disproportionation activity at 50°C in 50 mM sodium phosphate buffer (pH 5.5).

Catalytic activities of the CGTase wild type and variants.

The disproportionation and hydrolysis activities of wild-type CGTase and L277A, L277T, L277I, L277M, and L277F mutants were measured, and the disproportionation-to-hydrolysis ratios were calculated (Table 2). Compared with that of the wild-type CGTase (given the value 100%), the disproportionation-to-hydrolysis ratios of the L277A, L277T, and L277I mutants (12.42%, 50.20%, and 67.75%, respectively) were lower, while those of the L277M and L277F mutants (2.4 times and 1.3 times, respectively) were higher.

TABLE 2.

Disproportion and hydrolysis activities^a

Enzyme	Disproportionation activity (U mg−1)	Hydrolysis activity (U mg−1)	Disproportionation/hydrolysis	Fold
Wild type	496.8 ± 26.49	8.21 ± 1.73	60.5 ± 9.51	1.00
L277A mutant	157.1 ± 3.94	21.02 ± 5.18	7.5 ± 1.65	0.12 ± 0.01
L277T mutant	148.9 ± 3.09	4.87 ± 0.4	30.58 ± 1.87	0.51 ± 0.05
L277I mutant	346.2 ± 8.48	8.44 ± 1.68	41.03 ± 7.17	0.68 ± 0.01
L277M mutant	359.4 ± 9.98	2.47 ± 0.68	145.5 ± 36.11	2.4 ± 0.22
L277F mutant	288.9 ± 1.43	3.67 ± 1.04	78.8 ± 22.02	1.3 ± 0.16
G282A mutant	379.7 ± 2.35	12.09 ± 2.38	31.4 ± 5.98	0.52 ± 0.02
G282A/L277M mutant	487.6 ± 47.83	2.83 ± 0.29	172.2 ± 0.77	2.85 ± 0.43
G282A/L277F mutant	379.7 ± 2.31	7.88 ± 0.92	48.2 ± 5.35	0.8 ± 0.04

^aData represent the means and SDs from three independent measurements. Enzyme activity was measured at 50°C in 50 mM sodium phosphate buffer (pH 5.5), and the concentrations of EPS, maltose, and soluble starch were 1.7 mM, 8.6 mM, and 5 g liter⁻¹, respectively.

Kinetic parameters of CGTase and Leu²⁷⁷ mutants.

To explore the mechanistic rationale behind these changes in disproportionation-to-hydrolysis ratios, the kinetic parameters of the disproportionation and hydrolysis reaction were determined (Table 3). The L277A mutant exhibited a Michaelis constant for maltose (K_{m Mal}) that was much higher than that of the wild type (2.24 mM versus 0.945 mM). The specificity constants (k_cat/K_{m Mal}) of the L277A and L277T mutants were much lower than that of the wild type (110.2 s⁻¹ mM⁻¹ and 239.6 s⁻¹ mM⁻¹, respectively, versus 686.1 s⁻¹ mM⁻¹). The K_{m Mal} (0.81 mM) of the L277I mutant was similar to that of the wild type, but its k_cat (478.4 s⁻¹) and k_cat/K_{m Mal} (594.4 s⁻¹ mM⁻¹) were lower than those of the wild type. Thus, the disproportionation of the L277I mutant was also lower than that of the wild type. The K_{m Mal} of the L277F mutant (0.49 mM), which had a disproportionation-to-hydrolysis ratio somewhat higher than that of the wild type, was much lower than that of the wild type. However, its k_cat/K_{m Mal} (773.8 s⁻¹ mM⁻¹) was only marginally higher than that of the wild type. Finally, the mutant with the best disproportionation-to-hydrolysis ratio, the L277M mutant, exhibited a K_{m Mal} (0.48 mM) much lower than that of the wild type and a k_cat/K_{m Mal} (1,175 s⁻¹ mM⁻¹) much greater than that of the wild type.

TABLE 3.

Kinetic parameters of disproportionation and hydrolysis^a

Enzyme	Maltose			Soluble starch
	kcat (s−1)	Km (mM)	kcat/Km (s−1 mM−1）	kcat (s−1)	Km (g liter−1)	kcat/Km (liters s−1 g−1）
Wild type	648.3 ± 6.23 A	0.945 ± 0.007 C	686.1	11.00 ± 0.02 C	0.34 ± 0.07 A	32.01
L277A mutant	247.3 ± 2.96 E	2.24 ± 0.06 A	110.2	27.18 ± 0.33 A	0.08 ± 0.005 B	341.9
L277T mutant	200.9 ± 0.39 E	0.84 ± 0.04 CD	239.6	6.18 ± 0.17 D	0.128 ± 0.003 B	48.30
L277I mutant	478.4 ± 7.67 C	0.81 ± 0.04 CD	594.3	11.64 ± 0.50 C	0.28 ± 0.016 A	41.95
L277M mutant	564.1 ± 19.60 B	0.48 ± 0.04 E	1,175	3.24 ± 0.12 G	0.16 ± 0.02 B	20.72
L277F mutant	375.3 ± 12.53 D	0.49 ± 0.05 E	773.8	4.91 ± 0.14 E	0.11 ± 0.004 B	43.65
G282A mutant	526.8 ± 26.32 BC	0.648 ± 0.05 DE	813.5	16.54 ± 0.62 B	0.325 ± 0.057 A	50.89
L277M/G282A mutant	621.3 ± 22.85 A	1.22 ± 0.24 B	507.6	4.04 ± 0.1 F	0.101 ± 0.01 B	40.00
L277F/G282A mutant	496.2 ± 47.02 C	0.58 ± 0.07 DE	860.0	10.87 ± 0.51 C	0.336 ± 0.05 A	32.35

^aStatistical analyses were performed using one-way analysis of variance (ANOVA) followed by SNK (Student-Newman-Keuls; P ≤ 0.05), and letters indicate significant differences among mutants. All data represent the means and SDs from three independent experiments. Enzyme activity was measured at 50°C in 50 mM sodium phosphate buffer (pH 5.5).

The kinetic parameters of the hydrolysis reaction were determined to more accurately determine the mechanism. The L277A and L277T mutants had Michaelis constants for soluble starch (K_{m SS}; 0.08 g liter⁻¹ and 0.128 g liter⁻¹) that were much lower than that of the wild type (0.34 g liter⁻¹). The specificity constants of the L277A and L277T mutants (k_cat/K_{m SS}; 341.9 liters s⁻¹ g⁻¹ and 48.30 liters s⁻¹ g⁻¹) were larger than that of the wild type (32.01 liters s⁻¹ g⁻¹). The K_{m SS} (0.28 g liter⁻¹) and k_cat/K_{m SS} (41.95 liters s⁻¹ g⁻¹) of L277I were essentially the same as those of the wild type. These results were consistent with the kinetic parameters for maltose. The K_{m SS} of the L277F (0.11 g liter⁻¹) was much lower than that of the wild type but its k_cat/K_{m Mal} (43.65 liters s⁻¹ g⁻¹) was higher than that of the wild type. Finally, the L277M mutant exhibited a K_{m SS} (0.16 g liter⁻¹) much lower than that of the wild type and a k_cat/K_{m Mal} (20.72 liters s⁻¹ g⁻¹) lower than that of the wild type. The reason the affinity of the L277M and L277F mutants for soluble starch was enhanced is that the hydrophobicity of the +2 subsite was enhanced.

MD simulations of the wild type and the L277M, G282A, and L277M/G282A mutants.

To further assess why the L277M mutant had the highest disproportionation-to-hydrolysis ratio of the Leu²⁷⁷ mutants, we performed MD simulations of the wild type and some mutants. The trajectories in VMD exhibited different distances between the important residues Phe¹⁷⁹ and Phe²⁵⁵. The Cα distance (28) between Phe¹⁷⁹ and Phe²⁵⁵ was calculated using cpptraj (Fig. 3A). In Fig. 3B, the average distance between residues Phe¹⁷⁹ and Phe²⁵⁵ of the L277M mutant over 80 ns (10.12 Å) was shorter than that of the wild type (10.35 Å). This shortening may reflect the fact that the radius of the sulfur atom in the methionine of the L277M mutant is larger than that of the carbon atom of leucine (C, 0.77 Å; S, 1.04 Å). This change might cause the methionine at position 277 to impinge upon Phe²⁷⁹ and Phe²⁵⁵, causing Phe²⁵⁵ to move closer to Phe¹⁷⁹. Furthermore, The MD simulations of the G282A and L277M/G282A mutants were analyzed. The average distance between Phe¹⁷⁹ and Phe²⁵⁵ of the G282A mutant (10.17 Å) was similar to that of the L277M mutant and shorter than that of the wild type. The average distance between Phe¹⁷⁹ and Phe²⁵⁵ of the L277M/G282A mutant (9.27 Å) was the shortest of the three mutants.

FIG 3.

MD simulations of wild-type CGTase and its L277M mutant. (A) Crystal structure of B. stearothermophilus NO2 CGTase (PDB ID 1CYG). Residues Glu²⁵³ is rendered in pink, Leu²⁷⁷ is in yellow, Phe¹⁷⁹ and Phe²⁵⁵ are in dark blue, and Gly²⁸² is in light blue. Other residues that may be affected by Gly²⁸² are rendered in green. (B) Distance between Phe¹⁷⁹ and Phe²⁵⁵ over 80 ns. The blue line represents the wild type, the red line L277M, the green line G282A, and the gold line L277M/G282A.

To verify the MD results, additional mutants altering Gly²⁸² were considered. Mutation of Gly²⁸² to an amino acid with a larger side chain might cause Phe²⁷⁹ and Phe²⁵⁵ to approach Phe¹⁷⁹. A comparison of the crystal structure of B. stearothermophilus NO2 CGTase (PDB ID 1CGY) with those of other CGTases (PDB IDs 1CXH, 1D7F, 3WMS, and 4JCL) revealed that Gly²⁸² is not conserved. The CGTases from B. circulans 251, Bacillus sp. 1011, P. macerans 602-1, and P. macerans IAM1243 all have Ala residues at this position. Interestingly, B. circulans 251 CGTase and Bacillus sp. 1011 CGTase have K_{m Mal} values of 0.83 mM (6) and 0.49 mM (22), respectively, which are lower than that of B. stearothermophilus NO2 CGTase (Table 3). This suggests that replacing Gly²⁸² with alanine might facilitate maltose binding. Therefore, B. stearothermophilus NO2 G282A, L277M/G282A, and L277F/G282A CGTase mutants were prepared.

Preparation of Gly²⁸² mutants.

The G282A, L277M/G282A, and L277F/G282A mutants were prepared using the same methods as employed to produce the mutants described above. The purification data are shown in Fig. S3 and Table 1. Circular-dichroism results suggested that the secondary structures of the Gly²⁸² mutants were similar to that of the wild type (Fig. S2). These enzymes were also pure and had specific activities similar to those of the other mutants. The kinetic parameters of these G282A mutants were also measured using maltose as the acceptor (Table 3). The K_{m Mal} of the G282A mutant (0.648 mM) was somewhat lower than that of wild-type CGTase (0.945 mM), while the K_{m Mal} of the L277M/G282A mutant (1.22 mM) was substantially greater than that of the L277M mutant (0.48 mM) and the K_{m Mal} of the L277F/G282A mutant (0.58 mM) was quite similar to that of the L277F mutant (0.49 mM). The specificity constants (k_cat/K_m) of the G282A, L277M/G282A, and L277F/G282A mutants followed a trend opposite to that of their K_{m Mal} values. Otherwise, the affinity of the G282A mutant for soluble starch was quite similar to that of the wild type. Thus, the effect of the G282A mutations was not significant.

DISCUSSION

According to previous reports, residues that affected the disproportionation and hydrolysis activities of CGTase were located mostly at the +2 subsite. However, mutations at these sites reduced the disproportionation-to-hydrolysis ratio. Examples include the Phe²⁵⁹ and Phe²⁸³ mutants of Bacillus sp. 1011 CGTase (22), the Phe¹⁸³ and Phe²⁵⁹ mutants of B. circulans strain 251 CGTase (6, 23), and the Phe¹⁸⁴ and Phe²⁶⁰ mutants of T. thermosulfurigenes strain EM1 CGTase (24, 25). Thus, the +2 subsite exerts substantial influence on the disproportionation and hydrolysis activities of CGTase, and modification of the +2 subsite may not improve the disproportionation-to-hydrolysis ratio. In this study, these sites were avoided in favor of Leu²⁷⁷ in B. stearothermophilus NO2 CGTase, which is located near both the +2 subsite and the catalytic acid/base Glu²⁵³. Two of the mutants produced, the L277M and L277F mutants, exhibited disproportionation-to-hydrolysis ratios 2.4-fold and 1.3-fold greater than that of the wild type, respectively. After examining the crystal structure of the wild-type CGTase, and considering the nature of the Leu²⁷⁷ and Gly²⁸² mutants, two factors seem to be responsible for these increases in the disproportionation-to-hydrolysis ratio (Fig. 4A): (i) the mutations affect the distribution of water molecules near Glu²⁵³, and (ii) the mutations affect the hydrophobicity of the +2 subsite.

FIG 4.

Distribution of hydrophobic amino acids and the water channel. (A) Distribution of hydrophobic residues near the acceptor site of B. stearothermophilus NO2 CGTase. (B) Water channel near the acceptor site of B. circulans 251 CGTase. (C) Multiple-sequence alignment of residues in the hydrophobic cluster. In panels A and B, the residues rendered using thick green sticks represent a complete hydrophobic cluster, including Leu²⁷⁷ (rendered in yellow), located near Glu²⁵³ (rendered in pink). Phe¹⁷⁹ and Phe²⁵⁵, which are near the +2 subsite, are rendered using thick dark blue sticks, while Gly²⁸² is rendered using light blue sticks. The red dotted spheres in panel B represent water molecules. In panel C, abbreviations of hydrophobic amino acid residues are printed in black.

The first factor is consistent with the crystal structure of this CGTase and a multiple-sequence alignment with other CGTases. According to previous reports, there is a water channel near the active site of the enzyme (29). Changing this water channel in members of GH13 could effectively change the disproportionation-to-hydrolysis ratio of the enzyme (30). The crystal structure of B. stearothermophilus NO2 CGTase (PDB ID 1CYG) shows that there are hydrophobic clusters near Glu²⁵³ (Fig. 4A, green residues) that may prevent water molecules from reaching Glu²⁵³. No water channels are seen near these hydrophobic clusters, perhaps because the crystal structure of NO2 CGTase does not contain a bound substrate. In contrast, a water channel is present in the structure of the CGTase from B. circulans 251 (PDB ID 1CXH), which shares high homology with the CGTase from B. stearothermophilus NO2 (Fig. 4B). The water channel is located above Leu²⁷⁷, suggesting that hydrophobic clusters can affect the water channel, thereby affecting the disproportionation-to-hydrolysis ratio of the enzyme. An alignment of the sequences of 46 CGTases present in the CAZy database (Fig. S1), including eight with available crystal structures and others with information about their enzymatic properties, strongly suggests that this hydrophobic cluster structure is widely present in CGTases (Fig. 4C). Leu²⁷⁷ is a member of this hydrophobic cluster, so Leu²⁷⁷ mutants may affect the hydrophobic cluster, resulting in a large change in the water channel. This may alter the ability of water molecules to interact with Glu²⁵³. Furthermore, we can explain the results in Table 3 very well. The L277A mutant disrupts the structure of the hydrophobic cluster, allowing water molecules to more effectively compete with acceptor sugar molecules, decreasing the affinity of CGTase for maltose molecules and increasing the affinity of CGTase for soluble starch. The L277T mutant has a maltose affinity similar to that of the wild type, but its affinity for soluble starch is enhanced. We presume that the reason for this phenomenon is that threonine is similar to leucine in structure but different in chemical character. The K_{m Mal} and K_{m SS} of the L277I mutant are similar to those of the wild type because the side chains of Leu and Ile are so similar in size. L277M and L277F enhance hydrophobic interactions, thereby hindering the ability of water molecules to contact Glu²⁵³. The effect of the hydrophobic cluster on the interaction between water molecules and Glu²⁵³ determines the disproportionation-to-hydrolysis ratio to some extent, with the disproportion-to-hydrolysis ratio increasing as fewer water molecules interact with Glu²⁵³.

The second factor is completely consistent with both the MD results obtained with wild-type CGTase and the L277M mutant and the properties of the Gly²⁸² mutants. The MD results showed that the distance between Phe¹⁷⁹ and Phe²⁵⁵ is smaller in the L277M, G282A, and L277M/G282A mutants than in wild-type CGTase (Fig. 3). Previous reports identified Phe¹⁷⁹ and Phe²⁵⁵ as key residues determining the disproportionation-to-hydrolysis ratio of CGTase (20). In CGTases from various sources, mutation of Phe¹⁷⁹ and Phe²⁵⁵ reduced disproportionation activity and increased hydrolysis activity (31). The rationale has been that Phe¹⁷⁹ and Phe²⁵⁵, which are located at the +2 subsite, participate in hydrophobic interactions with acceptor sugar molecules. It has been reported that Phe²⁸³ of Bacillus sp. 1011 CGTase, which corresponds to Phe²⁷⁹ in B. stearothermophilus NO2 CGTase, decreases the flexibility of a loop comprising residues 259 to 269, thereby maintaining the hydrophobic environment of the +2 subsite (28). Leu²⁷⁷ is located near Phe²⁷⁹, so although the L277M mutation did not directly change Phe¹⁷⁹ and Phe²⁵⁵, it may have altered the conformation of Phe²⁷⁹, reducing the distance between Phe¹⁷⁹ and Phe²⁵⁵. This may strengthen the hydrophobic interaction and increase the affinity of CGTase for the acceptor sugar molecule. However, when Phe¹⁷⁹ and Phe²⁵⁵ are too close, we presumed that steric hindrance of the acceptor site may be too large, leading to the opposite result. This hypothesis is also consistent with the results seen with the Gly²⁸² mutants (Fig. 3B and Table 3). The G282A mutation narrowed the distance between Phe¹⁷⁹ and Phe²⁵⁵ and increased hydrophobic interaction at the acceptor site, lowering the K_{m Mal} of the G282A mutant (26.3% lower than that of wild-type). The L277M/G282A mutation changed the distance between Phe¹⁷⁹ and Phe²⁵⁵ more significantly, decreasing the size of the acceptor site so much that its affinity for maltose was reduced, significantly increasing the K_{m Mal} of the L277M/G282A mutant, to about 1.9 times that of the L277M mutant. Meanwhile, the K_{m SS} was reduced to 3 times that of the L277M mutant. The K_{m Mal} and K_{m SS} values of the G282A and L277F/G282A mutants were almost the same, suggesting that the L277F mutation did not affect the distance between Phe¹⁷⁹ and Phe²⁵⁵. This result indicates that only the L277M mutation affected the +2 subsite. The L277F mutation had no effect on the +2 subsite. Although phenylalanine has a larger side chain than methionine, simulated mutation (using PyMOL) indicated that the influence of methionine is greater (Fig. 5). The cavity created by the hydrophobic cluster accommodated the side chain of phenylalanine because all its atoms are in the same plane. The cavity cannot, however, accommodate methionine because its side chain is not planar. Thus, L277M can impinge on Phe²⁷⁹ and cause Phe¹⁷⁹ to approach Phe²⁵⁵, but L277F cannot.

FIG 5.

Diagram of Leu²⁷⁷ mutation simulated using PyMOL. (A) Location of the relevant site. (B) Enlarged view of the red circle in panel A; Glu²⁵³ is rendered in pink, Leu277 in yellow, L277M in blue, and L277F in orange.

An interesting result was also obtained from a trajectory analysis of the MD data. Glu²⁵³ had two different conformations: the “contracted” conformation, in which the carboxyl group faces away from the hydrophobic cluster, and the “relaxed” conformation, in which the carboxyl group faces toward the hydrophobic cluster. When a transglucosylase interacts with an acceptor molecule, the catalytic acid/base residue typically adjusts to fit the catalytic conformation and facilitate the reaction (5). The MD data were reviewed to assess the angle of Glu²⁵³ in the wild type and the L277M mutant over 80 ns. The wild type maintained the relaxed conformation for the 80 ns, while the L277M mutant gradually changed from the relaxed conformation to the contracted conformation within the first 40 ns (Fig. 6). Combining experimental data (Tables 2 and 3), further analysis of the crystal structure and MD calculations indicated that the contracted conformation may be more suitable for disproportionation reactions. To verify the effect of the Glu²⁵³ conformation on the disproportionation and hydrolysis activities of CGTase, more mutants near Glu²⁵³ are needed.

FIG 6.

Angle of Glu²⁵³ in the wild type and the L277M mutant. (A) Contracted and relaxed conformations of Glu²⁵³. The red and yellow solid lines represent the angle measured (OE2-CA-O). (B) Change of the Glu²⁵³ angle over 80 ns. The solid blue line represents the wild type, the solid red line represents the L277M mutant, and the black dotted line is 90°.

In summary, mutations at a completely new site, Leu²⁷⁷, enhanced the disproportionation-to-hydrolysis ratio of a CGTase. The L277M mutation had the greatest effect, and the L277F mutation had the second-greatest effect. Further analysis of the crystal structure and MD calculations indicated that the L277M mutation affects both the hydrophobic cluster near Glu²⁵³ and the distance between Phe¹⁷⁹ and Phe²⁵⁵ at the +2 subsite. In contrast, L277F affects only the hydrophobic cluster. This study provides a new way of thinking about the molecular transformations of CGTase. The effect of Leu²⁷⁷ on the disproportionation and hydrolysis activities of CGTase discovered in this study suggests that other residues in this region may have similar effects. Thus, this study opens up other avenues for the future research on the disproportionation and hydrolysis activities of CGTase and other enzymes in the GH13 family. Productive mutations are no longer limited to the acceptor subsite, since mutations that indirectly affect the acceptor subsite can enhance enzymatic activity.

MATERIALS AND METHODS

Bacterial strains, plasmids, and materials.

Escherichia coli JM109 was used as the host for DNA recombination. E. coli BL21(DE3) was used as the host for CGTase production. Recombinant plasmid pET20b-cgt (see methods in the supplemental material) was used to produce B. stearothermophilus NO2 CGTase (PDB ID 1CYG) with a C-terminal His tag (32). PrimerSTAR HS DNA polymerase and DpnI were purchased from TaKaRa Biotechnology (Dalian, China). 4-Nitrophenyl-α-d-maltoheptaoside-4-6-O-ethylidene (EPS) was purchased from Yixin Biotechnology Co. Ltd. (Shanghai, China). 3,5-Dinitrosalicylic acid (DNS) and other chemicals were purchased from Sinopharm Chemical Reagent Co. Ltd. (Shanghai, China).

Construction of mutant CGTase production strains.

The L277A, L277T, L277I, L277M, L277F, and G282A mutants were constructed using plasmid pET-20b(+)/cgt as the template, and the L277M/G282A and L277F/G282A mutants were constructed using plasmids pET-20b(+)/L277M and pET-20b(+)/L277F as templates, respectively. The primers used for PCR (performed using an instrument from Thermal Cycler, Shanghai, China) are shown in Table 4. After treatment with DpnI, the PCR products were inserted into E. coli JM109 for amplification. Plasmids containing the desired mutations were confirmed with DNA sequencing and then inserted into E. coli BL21(DE3) for protein production.

TABLE 4.

Primers used for site-directed mutagenesis

Primer	Sequence (5′–3′ direction)a
L277A-For	TCTGGCATGAGCCTGGCGGATTTCCGTTTTGGT
L277A-Rev	ACCAAAACGGAAATCCGCCAGGCTCATGCCAGA
L277T-For	TCTGGCATGAGCCTGACGGATTTCCGTTTTGGT
L277T-Rev	ACCAAAACGGAAATCCGTCAGGCTCATGCCAGA
L277I-For	TCTGGCATGAGCCTGATCGATTTCCGTTTTGGT
L277I-Rev	ACCAAAACGGAAATCGATCAGGCTCATGCCAGA
L277M-For	TCTGGCATGAGCCTGATGGATTTCCGTTTTGGT
L277M-Rev	ACCAAAACGGAAATCCATCAGGCTCATGCCAGA
L277F-For	TCTGGCATGAGCCTGTTTGATTTCCGTTTTGGT
L277F-Rev	ACCAAAACGGAAATCAAACAGGCTCATGCCAGA
G282A-For	CTGGATTTCCGTTTTGCTCAGAAACTGCGTCAG
G282A-Rev	CTGACGCAGTTTCTGAGCAAAACGGAAATCCAG
L277M/G282A-For	ATGGATTTCCGTTTTGCTCAGAAACTGCGTCAG
L277M/G282A-Rev	CTGACGCAGTTTCTGAGCAAAACGGAAATCCAT
L277F/G282A-For	TTTGATTTCCGTTTTGCTCAGAAACTGCGTCAG
L277F/G282A-Rev	CTGACGCAGTTTCTGAGCAAAACGGAAATCAAA

^aThe underlined letters indicate the sites of mutation.

Production of wild-type and mutant CGTases.

Individual E. coli BL21(DE3) strains harboring a plasmid encoding either wild-type or a mutant CGTase were cultured in 10 ml of Luria-Bertani (LB) medium supplemented with 100 μg/ml of ampicillin at 37°C for 10 h with shaking (Shanghai Zhichu Instruments Co., Ltd., Shanghai, China) at 200 rpm. A 5-ml aliquot of each of culture was used to inoculate 100 ml of Terrific broth (TB) medium supplemented with 100 μg/ml of ampicillin, and the resulting culture was incubated at 37°C for 2 h with shaking at 200 rpm. At this point, the temperature was lowered to 25°C and the culture was shaken at 200 rpm for an additional 48 h. Finally, the culture was centrifuged (Beckman Coulter, Shanghai, China) at 11,300 × g for 20 min at 4°C and the supernatant was saved as the crude CGTase solution.

Purification of CGTase wild type and variants.

Since the wild-type and mutant CGTases contained His tags, they were all purified using immobilized-metal affinity chromatography. The crude CGTase solutions were loaded onto nickel-nitrilotriacetic acid (Ni-NTA) columns (GE, Shanghai, China) and then eluted in two steps. In the first step, impurities were eluted using aliquots of buffer A (25 mM Tris-HCl buffer, 500 mM NaCl [pH 7.4]) supplemented with 0, 15, 30, and 45 mM imidazole. In the second step, the CGTases were eluted using buffer B (buffer A supplemented with 300 mM imidazole). Fractions containing the purified enzyme were concentrated using ultrafiltration devices with a 30-kDa-cutoff filter (Millipore, Shanghai, China) and then placed in 50 mM sodium phosphate buffer (pH 5.5).

CGTase activity assays.

Disproportionation activity was measured as previously described (27, 32), with minor modifications. Reaction substrates EPS (pNPG₇, a maltoheptasaccharide blocked at the nonreducing end and containing a p-nitrophenyl group at its reducing end; 4 mM) and maltose (20 mM), used as donor and acceptor, respectively, were prepared using 50 mM sodium phosphate buffer (pH 5.5). Aliquots (300 μl) of each (final concentrations: EPS, 1.7 mM; maltose, 8.6 mM) were mixed and preheated (Senxin, Shanghai, China) at 50°C for 10 min. At this point, appropriately diluted CGTase (about 0.003 to 0.006 μg) was added to the substrate mixture and the reaction mixture was incubated at 50°C and pH 5.5 for 10 min. Finally, the reaction was terminated by placing the tube in boiling water for 10 min. After cooling, the disproportionation products, pNPG_n (n < 7), were hydrolyzed using 100 μl of α-glucosidase (EC 3.2.1.20; Sigma) to release p-nitrophenol at 60°C and pH 5.5 for 1 h. Sodium carbonate solution (1 M, 200 μl) was added to stop the reaction. The p-nitrophenol released during the assay turned the mixture yellow, and the absorbance was measured at 405 nm. One unit of disproportionation activity was defined as the amount of enzyme required to convert 1 μmol of EPS per minute.

Hydrolysis activity was measured as previously described (32), with minor modifications. The substrate consisted of 10 g liter⁻¹ of soluble starch prepared using 50 mM sodium phosphate buffer (pH 5.5). To perform the assay, 1 ml of substrate (final concentration: 5 g liter⁻¹) and 0.9 ml of buffer were added to a tube and then preheated at 50°C for 10 min. At this point, 100 μl of appropriately diluted CGTase (about 3 to 6 μg) was added and the assay mixture was incubated at 50°C and pH 5.5 for 10 min. Finally, 3 ml of DNS reagent (which reacted with the reducing sugar, maltose, that was released during the assay) was added to stop the reaction, and the absorbance of the mixture was measured at 540 nm. One unit of hydrolysis activity was defined as the amount of enzyme required to produce 1 μmol of maltose per minute. All of the enzyme activity data presented represent the means of three independent measurements.

Kinetic parameters.

Kinetic parameters for the disproportionation activities of CGTase variants were measured using several series of assays conducted for 10 min at 50°C in 50 mM sodium phosphate buffer (pH 5.5) using EPS (4 mM), maltose (0.6, 0.8, 1, 2, 4, 6, 8, 10, 20, 40, and 60 mM), and soluble starch (0.6, 0.8, 1, 2, 4, 6, 8, 10, and 20 g liter⁻¹) as substrates. The concentration of one substrate was kept constant while studying the effect of the other substrate. Experimental rate data were fitted to a model based upon the ping-pong mechanism using R software version 3.6.1. The ping-pong mechanism model has been previously described (9). All of the kinetic parameters presented represent the means of three independent measurements.

MD analysis.

Three-dimensional structural models of the CGTase variants were obtained using the three-dimensional structure of wild-type B. stearothermophilus NO2 CGTase (PDB ID 1CYG) as the template and PyMOL software (33). Amber18 (34) was used to perform classical molecular dynamics (MD) analysis simulations, using ff14SB (protein residues) (35), and TIP3P (water molecules) (36) to simulate 80 ns at 300K. The trajectories (the coordinate and velocity of each atom at different times) of the CGTase variants were analyzed using the cpptraj script of AMBER and VMD.

Multiple-sequence alignment of CGTases.

The amino acid sequences of appropriate members of GH13 subfamily 2 (GH13_2) were obtained from the Carbohydrate-Active enZYmes (CAZy) database (http://www.cazy.org/GH13.html) (see Fig. S1). Multiple-sequence alignment was performed using GREMLIN (http://gremlin.bakerlab.org/submit.php) (37) and MAFFT v7.037b, and the resulting alignments were plotted using ESPript 3.0 (http://espript.ibcp.fr/ESPript/cgi-bin/ESPript.cgi) (38) and WebLogo 3 (http://weblogo.threeplusone.com/create.cgi) (39).