Thermoadaptation of α-Galactosidase AgaB1 in Thermus thermophilus

Bacterial strains, growth conditions, and transformation procedure. Thermus thermophilus OF1053GD (8) (ΔagaT Δkan Gal⁺), a mutant of strain TH125 (10), was used as a host strain for the cloning and the selection procedure. The strain was grown under strong aeration in mineral medium 162 (4) with 0.25% tryptone and 0.25% yeast extract at pH 7.5 (T162). Growth was carried out at 65°C under nonselective conditions and at 60°C when cultures contained kanamycin (20 μg ml⁻¹) for plasmid selection. Growth on a single carbon source (melibiose, 0.1, 0.2, and 0.4%) was tested on agar plates containing the minimal medium 162 (4) with a slight modification as previously described (8). The method of Koyama et al. (14) was used for the transformation of T. thermophilus, with a slight modification as previously described (7). Transformants were selected on T162 agar plates containing 20 μg of kanamycin ml⁻¹ and incubated at 60°C for 48 h. All plasmids constructed, except those which could replicate only in Thermus bacteria, were brought into Escherichia coli JM109 [supE44 Δ(lac-proAB) hsdR17 recA1 endA1 gyrA96 thi-1 relA1] (F′traD36 proAB lacI^qZ ΔM15) (31) by transformation (2). Transformants were either selected for ampicillin resistance on agar plates with 100 μg of ampicillin per ml or for kanamycin resistance on plates with 25 μg of kanamycin per ml.

DNA manipulation and general plasmid construction.Recombinant DNA techniques, i.e., plasmid preparation, subcloning, and agarose gel electrophoresis, were performed by conventional protocols (25). Structures of the plasmids were confirmed by restriction mapping. Further, the inserts of pOF287 and pOF288 were confirmed by sequencing. Sequencing reactions of double-stranded DNA were carried out according to the dideoxy chain termination method with universal and internal primers (26). The plasmids used in this study are listed in Table 1.

Construction of Thermus replication plasmids containing agaA and agaB hybrid genes.Hybrid genes, based on agaA and agaB, of B. stearothermophilus KVE39 were cloned into a Thermus autonomously replicating plasmid, downstream of the E. coli tac promoter (1) and a ribosome binding site from T. thermophilus (8). The construction of plasmid pOF2811, carrying an AgaA type encoding gene, agaA1, which served as a positive control, was as follows. Plasmid pOF1253 contains agaA1 downstream of the Thermus ribosome binding site. It was constructed by replacing the kan gene of pOF1155 (8), which contains the marker between flanking sequences of the native α-galactosidase gene agaT (7), with agaA1. Due to an NdeI restriction site in agaA, a three-fragment ligation was performed with the following: the agaA gene following amplification (8) and NdeI-HindIII digestion; agaB, following amplification (8) and HindIII-BglII digestion; and the pOF1155 following NdeI-BamHI digestion. The resulting chimeric gene encoded an enzyme with the characteristic phenotypes of AgaA. The construction of pOF10726 is described elsewhere (8). The plasmid contains a stable plasmid mutant of pOF5714 (8) in a pUC19 derivative. A PstI fragment from pOF10726 containing repA, the minimal replication unit (3) from the Thermus plasmid pTSP1 (16), was cloned in pIC20R (19) to produce pOF184. agaA1 along with the Thermus putative Shine-Dalgarno (SD) sequence from pOF1253 was amplified by PCR with primers S1065 and S952 (8). The EcoRI restriction site in the S1065 primer sequence was disrupted by EcoRI digestion and a Klenow modification. This was done in order to be able to use other EcoRI restriction sites in later steps. Subsequently, the gene was blunt end ligated downstream of the tac promoter in pBTac1 (1) to produce pOF182. Further, agaA1 along with the tac promoter and the putative Thermus SD sequence was amplified from pOF182 by PCR with primers S1318 and S952 (8) and cloned between the HindIII and BglII sites of pOF184. This was carried out in two steps. First, the 3′ region of agaA1 (a HindIII-BglII fragment) was ligated between the HindIII and BglII sites of pOF184 to produce pOF185. Then, the 5′ region of agaA1, along with the upstream P_tac sequence, was ligated into the HindIII site of pOF185 to produce pOF187. The kan gene (21), along with the Thermus slpA promoter (16) from pOF1056 (8), was ligated into the MunI site of pOF187 to produce pOF287. The pIC region of pOF287 was deleted by EcoRI digestion, and the remaining plasmid was self-ligated to produce pOF2811 before transformation of T. thermophilus. The corresponding AgaB type plasmid, pOF2812, was generated in the same way except the plasmid pOF762 was the initial source of agaB1. To construct pOF762, a BsmI-MunI fragment of pOF1253 was replaced by a corresponding agaB BsmI-MunI fragment following the amplification of agaB from pCG3 (11), using the primers S951 and S952 (8) and a BsmI-MunI digestion and a ligation. All plasmids involved in the construction of pOF2811 and pOF2812 are listed in Table 1. Figure 2 shows the last step in the construction of pOF2812 containing agaB1, which was used to transform T. thermophilus OF1053GD for subsequent thermoadaptation experiments.

• Open in new tab
• Download powerpoint

FIG. 2.

The last step in the construction of the T. thermophilus plasmid pOF2812, containing agaB1. The progenitor plasmid pOF288 (shuttle vector) was established in E. coli as described in Materials and Methods. Deletion of the pIC region in pOF288 to produce pOF2812 before transformation of T. thermophilus was supposed to improve plasmid stability in strain OF1053GD, according to previous results (8).

Thermoadaptation in aqueous culture, N-methyl-N′-nitro-N-nitrosoguanidine (MNNG) plasmid mutagenesis, and recloning into E. coli.Cells from a late-exponential-phase culture of T. thermophilus OF1053GD carrying pO2812 (with agaB1) were washed in 0.9% NaCl and suspended in 162 minimal medium containing 0.4% melibiose. The cells were incubated at 67°C under aeration. Following incubation for 6 days, the cells were diluted (three times) in a fresh medium. Following incubation for 13 days, the culture was diluted and plated on a minimal melibiose agar medium.

Plasmid mutagenesis was performed by adding MNNG to an exponential culture (optical density at 600 nm, 0.4) of T. thermophilus OF1053GD harboring pOF2812, in T162 kanamycin medium. Several concentrations of MNNG were tested (0.05 to 0.250 mg ml⁻¹). Following growth overnight at 60°C, the cells were washed in 0.9% NaCl and plated on 162 minimal melibiose agar medium. Plasmids from strains, capable of growth, were isolated by alkali lysis (25) and reintroduced into a plasmid-free strain of OF1053GD by transformation for the confirmation of the plasmid-dependent phenotype. Further, the mutant genes were amplified by PCR, with primer S462, CGG AAT TCA GAG AAT GTC AGT TGC ATA C, containing an EcoRI linker and S952, GAA GAT CTC AAT TGT CTT ATT GTT GAA CAG, containing a BglII linker (8). The amplified fragments were digested with EcoRI and BglII and ligated between the EcoRI and BamHI sites in pBTac1 (1) for the expression and characterization of mutant enzymes and in pUC18 (31) for sequencing.

Enzyme assays.Preparation of crude extracts, estimation of protein concentrations, and procedure of enzyme assays have been described previously (7). α-Galactosidase activity was determined by measuring the rate of para-nitrophenyl-α-d-galactoside (pNP-α-galactoside) hydrolysis (4 mg ml⁻¹ in 0.1 M potassium buffer [pH 6.5]) at the desired temperature. One unit of activity is defined as the amount of enzyme required to hydrolyze 1 μmol of pNP-α-galactoside min⁻¹ under a given assay condition (9). Further, the rate of d-raffinose or melibiose hydrolysis, in 0.1 M potassium buffer, pH 6.5, was determined by assessing the amount of d-galactose released using high performance liquid chromatography equipment (7). Colonies, which displayed α-galactosidase activity, were detected by histochemical staining as previously described (8).

Determination of temperature optimum and stability.Production of enzymes for characterization was achieved with E. coli JM109 by using the pBTac1 expression vector. The presence of equal amounts of enzymes in crude extracts was verified on a sodium dodecyl sulfate-polyacrylamide gel (15) (data not shown). The temperature optimum was determined by performing pNP-α-galactoside assays at a temperature range of 25 to 75°C. Thermal stability was determined by estimating the T[1/2] value (the temperature at which 50% enzyme activity is measured after 30 min of incubation) as follows. The α-galactosidase variants were partially purified by thermal precipitation of E. coli crude extracts (1 mg of protein ml⁻¹) at 60°C for 20 min. The enzymes were then incubated in 0.1 M potassium phosphate buffer (pH 6.5) for 30 min at various temperatures over a range from 25 to 80°C. Following incubation the enzyme was cooled, and the remaining activity was determined at the same temperature for all enzymes (37°C).

Nucleotide sequence accession numbers.The GenBank accession numbers of the nucleotide sequences of agaA and agaB are AY013286 and AY013287 , respectively.

Establishing a selection system.Our intention was to subject AgaB1 to thermoadaptation for selection of enzyme variants, which supported growth of the host strain on minimal melibiose medium at 67°C. For the establishment of a selection system, we intended to use AgaA1 as a positive control. In our previous work (8), we inserted the α-galactosidase genes agaA and agaB into a Thermus autonomously replicating vector downstream of the strong slpA promoter (16). Cells carrying these plasmids displayed a high-level production of α-galactosidases, verified by sodium dodecyl sulfate gel electrophoresis and activity tests (8). Although cells harboring the agaA plasmid (pOF5714) grew at 67°C on agar plates containing minimal melibiose medium, a noncontinuous growth in minimal melibiose aqueous medium was observed (data not shown). This was interpreted as an effect of the high-level α-galactosidase gene expression. In our previous work, we observed that a Gal⁺ phenotype of T. thermophilus was essential for growth on melibiose (8). As the host strain OF1053GD was capable of growing only slowly on a minimal agar medium containing 0.3% galactose, it is possible that the galactose metabolism could not fully compensate for the large amount of galactose produced from the melibiose hydrolysis in the cell, due to the high production of AgaA. Therefore, the α-galactosidase production was confined by constructing new plasmids with α-galactosidase genes inserted downstream of a Thermus ribosome binding site and the E. coli tac promoter, which was known to show low activities in T. thermophilus (20). Concomitantly, the kanamycin selection marker (21) was inserted downstream of the slpA promoter as in the shuttle vector pMY1 (16). To facilitate the construction of the plasmid modules, hybrid genes of agaA and agaB, designated agaA1 and agaB1, respectively, were used. The enzyme activities in crude extracts of T. thermophilus OF1053GD harboring pOF2811 with agaA1 or pOF2812 with agaB1 are shown in Table 2. For the comparison, enzyme activities in T. thermophilus OF1053GD harboring pOF5714 (agaA) and pOF1176 (agaB2) from our previous work (8), with the α-galactosidase genes downstream of the slpA promoter (16), are shown. Purified AgaA and AgaB2 displayed specific activities similar to those of AgaA1 and AgaB1, respectively (data not shown). As expected, a roughly 10-fold lower level of enzyme activity was measured in cells containing the α-galactosidase genes downstream of the tac promoter. Moreover, the strain OF1053GD harboring the agaA1 plasmid pOF2811 exhibited stable growth in a liquid minimal melibiose medium (0.4%). In contrast, OF1053GD without a plasmid or harboring the agaB1 gene in the plasmid pOF2812 remained in the lag phase for a period of days (Fig. 3). Also, growth at 67°C on minimal agar plates containing 0.2 and 0.1% melibiose was observed only with strain OF1053GD carrying the agaA1 gene in plasmid pOF2811 (Table 3) following incubation for 5 days. Thus, the requirements for the genetic selection were achieved.

• Open in new tab
• Download powerpoint

FIG. 3.

Growth of T. thermophilus OF1053GD at 67°C, in minimal medium containing 0.4% melibiose as a single carbohydrate source. ▪, OF1053GD/pOF2811 (agaA1); ⧫, OF1053GD/pOF2812 (agaB1); •, OF1053GD without plasmid.

Thermoadaptation of AgaB1, MNNG mutagenesis, and screening. T. thermophilus OF1053GD harboring pOF2812 (agaB1) was incubated at 67°C in an aqueous minimal medium containing 0.4% melibiose. The cells divided once and then remained in the lag phase. Although the cells did not grow at this temperature during the first days, no death phase was observed, according to cell titer quantification (data not shown). Slow growth was observed after 11 days of incubation. The cells were plated out and tested for growth on minimal agar medium. A thermoadapted strain harboring a plasmid, designated pOF2812 M1, was selected for further examination. Subsequently, the plasmid was isolated from its host strain and reintroduced into a plasmid-free strain of OF1053GD.

Another approach for selecting thermoadapted mutants involved reinforcement with mutagens. Following treatment with MNNG, as explained in Materials and Methods, mutants were plated and screened on a minimal agar medium containing 0.2% melibiose. Plasmid mutagenesis is commonly applied by adding the mutagenic agent to the culture medium (28). The highest number of mutants was obtained with a relatively high dose (0.15 mg ml⁻¹), probably due to instability of MNNG in rich medium and at high temperatures. Four independent mutants, capable of growth, were isolated following incubation at the restricted temperature of 67°C for 6 days. The corresponding plasmids, designated pOF2812 M2 to M5, were isolated from their host strain and reintroduced into a plasmid-free strain of OF1053GD. The cells carrying these different plasmids, including pOF2812 M1, with the spontaneous mutation, were all capable of growing at 67°C on a minimal melibiose medium, thus confirming the plasmid-dependent phenotype. Further, α-galactosidase activity at 50 and 60°C in crude extracts of the plasmid-harboring strains was examined. All enzymes displayed higher activity at 60 than at 50°C, in contrast to the wild-type enzyme AgaB1.

Sequencing and characterization of the mutant enzymes.The sequencing of the mutant genes, designated agaM1 to agaM5 (from pOF2812 M1 to M5, respectively), revealed a single nucleotide replacement in all cases. In agaM1 to agaM4, A-1064, according to agaB1 signature, was replaced with G. This mutation is located, as expected, in the BsmI-HindIII fragment of the gene and corresponds to a Glu355→Gly amino acid replacement. In agaM5, the same nucleotide (A-1064) was replaced with T, which corresponds to a Glu355→Val amino acid replacement. These base substitutions are inconsistent with the generally accepted effect of MNNG mutagenesis, which most commonly leads to GC→AT transitions (27). Nevertheless, control plating indicated the mutagenic effect of MNNG. No colonies were detected during the mutagen approach, on minimal agar medium following 6 days of incubation, where cells without a mutagenic treatment had been plated as a control. On the other hand, the A1064→G transition in pOF2812 M1, which occurred spontaneously, was detected following a prolonged incubation at a restricted temperature. Although MNNG-induced A→G transitions or A→T transversions are generally infrequent, such mutations induced with alkylating agents are common in E. coli alkA mutants, which are deficient for 3-methyladenine DNA glycosylase II, involved in induced DNA repair activities. In most cases these mutations are umuC⁺ dependent, i.e., resulting from SOS mutagenic processing (6). Thus, it is possible that deficient specific DNA repair activities in T. thermophilus OF1053GD, induced by exposure of MNNG, caused the base substitutions at A/T site 1064 to occur. Also, this site may be a substitution hot spot, similar to what is known to occur in E. coli (5).

The enzymatic properties of AgaM4 (Glu→Gly) and AgaM5 (Glu→Val) were examined and compared with those of the wild-type enzyme (AgaB1). Figure 4 shows the effect of temperature on the activity and stability (panels A and B, respectively). The optimum hydrolyzing activity of AgaM4 (following 15 min of preincubation at respective temperatures) was determined to be at 65°C, about 15°C higher than for the wild-type enzyme (50°C). The optimum activity of AgaM5 was determined to be at 60°C (Fig. 4A). The temperature where 50% of the activity remains following a 30-min incubation (T[1/2]) was determined to be 71°C for the mutant enzymes, at least 3°C higher than for the AgaB1 enzyme (Fig. 4B). Although this difference is not large, these results were reproducible. The slightly enhanced stability observed for the mutant enzymes can, however, hardly account for the increase in the temperature optimum for the hydrolyzing activity. Table 4 summarizes the K_m values of AgaA, AgaB1, AgaM4, and AgaM5 for the hydrolysis of pNP-α-galactopyranoside at 25, 50, and 60°C. The K_m of AgaB1 increases substantially along with a rise in temperature, whereas the K_m values of the mutant enzymes remain below 0.5 mM. Thus, the hydrolyzing rate of AgaB1 at high temperatures is affected by the drastic change in substrate affinity. Table 5 summarizes the kinetic parameters of AgaA, AgaB1, and the mutant enzymes for melibiose and raffinose hydrolysis. The K_m values for AgaM4 and AgaM5 are much lower than for AgaB1 and resemble the values of AgaA. Further, the mutant enzymes are superior to the progenitor enzyme (AgaB1) in regard to catalytic properties and are even considerably better than AgaA as reflected by the V_max/K_m values in Table 5 for hydrolysis at 37°C.

• Open in new tab
• Download powerpoint

FIG. 4.

Effect of temperature on the pNP-α-galactopyranoside hydrolyzing activity (A) and stability (B) of AgaA (▪), AgaB1 (⧫), AgaM4 (▴), and AgaM5 (•). Standard pNP-α-galactosidase assays at pH 6.5 were performed using enzymes in E. coli crude extracts of protein concentration 1 mg ml⁻¹. Temperature stability was determined by measuring the residual α-galactosidase activity after incubation of enzymes at a temperature range of 25 to 80°C as described in Materials and Methods. T[1/2], the temperature where the remaining activity is 50% following 30 min of incubation, is indicated by arrows. All activity tests were done in triplicate. The maximum variation from the mean values (shown) was less than 5%.

To verify that the described single-nucleotide replacements at position 355 are determinative for the new enzyme phenotypes, the observed point mutations were introduced into the agaB gene. This was carried out using our routinely applied E. coli expression vector construct (agaB in pBTac1) and a conventional site-directed mutagenesis (25). The resulting mutant enzymes were produced, isolated from E. coli, and characterized. The mutants derived thereby were identical to AgaM4 and AgaM5, respectively (data not shown). The results indicate that amino acid residue 355 influences the kinetic properties and stability of the B. stearothermophilus α-galactosidases. It seems unlikely that the modest increase in thermostability displayed by the mutants (Fig. 4B) is responsible for the dramatic difference in K_m values. It seems more likely that the residue at position 355 affects the affinity for the substrate more directly. A simple, plausible explanation may be that the binding or access to the binding site is obstructed by the large side chain of the glutamate residue but not by the smaller side chains of the residues found in AgaA (Ala) and the mutants (Gly and Val). Yet in absence of a three-dimensional structure of the enzyme, this remains speculative. According to the nature of the amino acid residue 355 in AgaA as well as in the thermoadapted mutants, nonpolar side chains or uncharged polar side chains of amino acid 355 may be compulsory for the enzyme to acquire the properties suitable for α-galactoside hydrolysis at high temperatures (60 to 70°C). Figure 5 shows an amino acid alignment of bacterial α-galactosidase partial sequences which correspond to the BsmI-HindIII fragment of agaA and agaB. All of them contain Ala in position 355 (according to the AgaB1 sequence numbering), except AgaB1 and the α-galactosidase from Bacillus halodurans (accession number NP_243089 ), which contains Gly as observed in AgaM1 to AgaM4. Further analysis of the evolutionary potential of α-galactosidases may be attempted, e.g., by applying this system for selection of mutants at temperatures over 70°C.

• Open in new tab
• Download powerpoint

FIG. 5.

Alignment of α-galactosidase partial amino acid sequences, residues 326 to 387 according to AgaB1 amino acid signature. The sequence region corresponds approximately to the BsmI-HindIII fragment of AgaA and AgaB. Amino acid sequences of bacterial α-galactosidases belonging to family 36 and with a subunit size of about 80 kDa (7) were retrieved from protein databases according to their accession numbers and aligned by using CLUSTALX version 1.8 (30). The sequences are as follows: AgaB1, (the sequence region is identical to the corresponding region in AgaB from B. stearothermophilus [B. stearotherm] KVE39, accession no. AY013287 ); AgaA1 (the sequence region is identical to the corresponding region in AgaA from B. stearothermophilus KVE39, accession no. AY013286 ); AgaN from B. stearothermophilus NUB3621 (accession no. AF130985 ); RafA from E. coli, accession no. P16551 ; Aga from Streptococcus mutans, accession no. P27756 ; MelA from Thermoanaerobacter ethanolicus, accession no. CAA69852 ; AgaR from Pediococcus pentosaceus, accession no. L32093 ; α-galactosidase from Yersinia pestis, accession no. NP_405164 ; α-galactosidase from Streptococcus pneumoniae, accession no. NP_346329 ; α-galactosidase from B. halodurans, accession no. NP_243089 . Identical residues are indicated with black boxes (shade threshold, 100%) and gray boxes (shade threshold, 80%). The residue corresponding to amino acid 355 in AgaB1 is marked with a solid triangle.

Concluding remarks and general discussion.Thermophilic selection systems, such as the one described in this paper, can be useful for isolation of enzyme variants with improved thermostability. Yet, substantial thermostabilization of mesophilic proteins by means of in vitro or in vivo evolution systems may be a difficult task since the routes around multidimensional thermodynamic obstacles to suitable thermostable structures are difficult to find and evolving structures tend to get caught in thermodynamic troughs. Another potential application for thermophilic in vivo evolution systems is to establish thermostable enzyme variants with improved specific activity at elevated temperatures for relevant substrates. In fact, this may be especially pragmatic since thermophilic enzymes sometimes show lower specific activity than their mesophilic or less thermophilic counterparts at their respective temperature optima. We have seen this in the case of α-galactosidases where the thermostable α-galactosidases from Thermus spp. display, in general, a lower specific activity for some substrates than their less thermostable counterparts from B. stearothermophilus, E. coli, and Streptococcus mutans at respective temperature optima (7; also unpublished results). Yet do we need a thermophile to promote such mutants; i.e., can an E. coli selection system serve this purpose? We have observed that selection of enzyme variants with improved catalytic activity can be dependent on the growth temperature of the selector organism; e.g., in previous work in this laboratory, an attempt was made to isolate variants of AgaB with improved catalytic activity by applying mutagenesis, using E. coli and selecting for growth at 37°C. Enzyme variants were isolated which displayed improved catalytic activity, but only at moderate temperatures (37 to 50°C) (A. Gehweiler et al., unpublished results). Mutants with improved catalytic activity at elevated temperatures (>60°C) were obtained only by applying the thermophilic selection system described in this paper.

Although various α-galactosidases of eubacterial and eukaryotic sources have been studied extensively regarding their biochemical and physical properties and a number of sequences are available in databases, the structure and detailed catalytic mechanisms remain to be solved. Hence, for improvement of enzyme properties, rational design and protein engineering by site-directed mutagenesis are restricted. In this study we employed an in vivo system to investigate the evolutionary potential of an α-galactosidase. We succeeded in isolating mutants which supported the growth of the thermophilic host cells at an elevated temperature (67°C) on a selective medium. Characterization of the mutants revealed improved substrate affinity and catalytic capacity at elevated temperatures besides slightly enhanced thermostability. The results show that thermophilic bacteria can principally be applied for a thermoadaptation of glycoside-hydrolyzing enzymes. Thus, such in vivo evolution systems should be applicable to other industrially important enzymes as well, and in combination with in vitro evolution techniques, such as gene shuffling, should present a powerful method for obtaining thermoactive and thermostable enzymes. The prerequisite is the capability of the thermophilic host to import and metabolize the relevant substrates and the evolutionary potential of the particular enzyme.