Coevolution of both Thermostability and Activity of Polyphosphate Glucokinase from Thermobifida fusca YX

Polyphosphate glucokinase (PPGK) is an enzyme that transfers a terminal phosphate group from polyphosphate to glucose, producing glucose 6-phosphate. A petri dish-based double-layer high-throughput screening strategy was developed by using ultrathermostable Phytagel as the first layer instead of agar or agarose, followed by a redox dye-based assay for rapid identification of ultrathermostable PPGK mutants. The best mutant featuring both great thermostability and high activity could produce glucose 6-phosphate from glucose and polyphosphate without in vitro ATP regeneration.

ABSTRACT

Thermostability and specific activity of enzymes are two of the most important properties for industrial biocatalysts. Here, we developed a petri dish-based double-layer high-throughput screening (HTS) strategy for rapid identification of desired mutants of polyphosphate glucokinase (PPGK) from a thermophilic actinobacterium, Thermobifida fusca YX, with both enhanced thermostability and activity. Escherichia coli colonies representing a PPGK mutant library were grown on the first-layer Phytagel-based plates, which can remain solid for 1 h, even at heat treatment temperatures of more than 100°C. The second layer that was poured on the first layer contained agarose, substrates, glucose 6-phosphate dehydrogenase (G6PDH), the redox dye tetranitroblue tetrazolium (TNBT), and phenazine methosulfate. G6PDH was able to oxidize the product from the PPGK-catalyzed reaction and generate NADH, which can be easily examined by a TNBT-based colorimetric assay. The best mutant obtained after four rounds of directed evolution had a 7,200-fold longer half-life at 55°C, 19.8°C higher midpoint of unfolding temperature (T_m), and a nearly 3-fold enhancement in specific activities compared to those of the wild-type PPGK. The best mutant was used to produce 9.98 g/liter myo-inositol from 10 g/liter glucose, with a theoretical yield of 99.8%, along with two other hyperthermophilic enzymes at 70°C. This PPGK mutant featuring both great thermostability and high activity would be useful for ATP-free production of glucose 6-phosphate or its derived products.

IMPORTANCE Polyphosphate glucokinase (PPGK) is an enzyme that transfers a terminal phosphate group from polyphosphate to glucose, producing glucose 6-phosphate. A petri dish-based double-layer high-throughput screening strategy was developed by using ultrathermostable Phytagel as the first layer instead of agar or agarose, followed by a redox dye-based assay for rapid identification of ultrathermostable PPGK mutants. The best mutant featuring both great thermostability and high activity could produce glucose 6-phosphate from glucose and polyphosphate without in vitro ATP regeneration.

INTRODUCTION

Polyphosphate glucokinase (PPGK) (E.C. 2.7.1.63) is an enzyme that transfers a terminal phosphate group from polyphosphate, (P_i)_n, to glucose, producing glucose 6-phosphate (G6P) (1–3). G6P is an important precursor of numerous high-value biochemicals, such as ginsenoside (4), fructose 1,6-bisphosphate (5), myo-inositol (6, 7), hydrogen (8, 9), bioelectricity (1, 3), and so on. Polyphosphate is a linear polymer of tens or even hundreds of orthophosphate residues linked by high-energy phosphoanhydride bonds. It is a preferable phosphoryl donor over ATP in industrial biocatalysis due to its low price, high thermostability, and high stability at a broad pH range (10). PPGKs have been discovered in numerous organisms, such as mycobacterial species (11), Propionibacterium shermanii (12), Microlunatus phosphovorus (13), Arthrobacter spp. (14), Corynebacterium glutamicum (15), diazotrophic cyanobacterial species (16, 17), and so on. However, nearly all reported PPGKs have been isolated from mesophilic sources, and the only thermophilic PPGK from Thermobifida fusca suffers from rapid deactivation at 50°C (i.e., half-life [t_1/2], 15 min) (2). A lack of thermostable PPGK hampers its potential application for ATP-free G6P production, especially working with other hyperthermophilic enzymes (7).

The thermostability of enzymes is one of the most important properties for industrial biocatalysts (18). This characteristic can be improved through rational design, directed evolution, or a combination of the two (19–21). Different from rational design, directed evolution is very effective to increase the thermostability of proteins without a known protein structure and catalytic mechanism (22–24). It usually involves two steps, construction of DNA mutant library and screening/selection. The success of directed evolution relies on the effectiveness of high-throughput screening (HTS), which can facilitate the rapid identification of positive mutants from a large mutant library (21, 25).

HTS for enhanced thermostability of enzyme mutants usually involves the deactivation of most mutants at an elevated temperature, followed by residual enzyme activity assays. The most straightforward method is based on microtiter plates, in which each well contains one colony. However, this method is labor-intensive and consumes many reagents and/or requires costly automated machines. In contrast, petri dish-based HTS approaches have the following benefits: higher screening capacity; less time and labor intensity used for colony picking, liquid cell culture, cell lysis, and enzyme activity assays; less reagent consumption; and no need for costly automated machine (26, 27). The key point of petri dish-based HTS techniques is the development of a visual signal of colonies growing on solid plates, such as halo (28), color change (22), and fluorescence (29). However, typical agar or agarose-based petri dishes cannot endure too high of a temperature treatment due to their low melting temperatures. A longstanding belief for enzymologists is that thermostability at high temperatures is incompatible with high catalytic activity at low temperatures due to mutually exclusive demands on enzyme flexibility (30). It is commonly asserted that the mechanism responsible for enhanced thermostability is a decrease in the flexibility of protein structure (i.e., increased rigidity). Consequently, a more thermostable enzyme is expected to be a slower catalyst at a lower temperature than a less stable enzyme (31, 32). Many studies have reported that improvements in thermostability are often obtained at the cost of enzyme activity, such as a decrease in k_cat and/or an increase in K_m (33, 34). However, more and more examples have demonstrated that mutants could be stabilized without sacrificing activities at low temperatures (30, 35–38). It is of theoretical and practical interest to investigate whether the activity-stability trade-off could be compromised by the coevolution of both properties via directed evolution.

In this study, we developed a petri dish-based double-layer screening method with a redox dye-based colorimetric assay for the rapid identification of thermostable PPGK mutants. Thermostable Phytagel, which can stand more than 100°C, was used for HTS. The coevolution of thermostability and activity of PPGK was achieved through four rounds of directed evolution. The best mutant was demonstrated in one-pot production of myo-inositol from glucose and polyphosphate with other two hyperthermophilic enzymes at 70°C.

RESULTS

A petri dish-based HTS method.

Previously, Zhang and his coworkers developed a simple petri dish-based double-layer screening for the identification of mutants of thermophilic NADP-dependent 6-phosphogluconate dehydrogenase with enhanced catalytic efficiencies on NAD⁺. In this screening, the reduced NADH can react with the redox dye TNBT, generating black TNBT formazan (27). This screening method cannot be applicable to ultrathermophilic enzymes due to low melting temperatures of agar or agarose. To address this limitation, we designed a two-step enzymatic reaction (Fig. 1A). The PPGK mutant library was transformed into E. coli TOP10 and then spread on the solid LB medium containing 1.2% Phytagel (Fig. 1B). Phytagel was used to replace agar or agarose because it retains its solid phase even at a temperature of up to 120°C for 1 h. Heat treatment was applied to decrease the background noise from mesophilic enzymes and reducing metabolites (e.g., NADPH and NADH) in E. coli, deactivate inherent PPGK mutants, and disrupt cell membranes (Fig. 1B). G6P generated by PPGK in the first reaction was used as the substrate of G6PDH for reducing NAD⁺ in the second reaction (Fig. 1A). Positive colonies expressing stable PPGK exhibited darker color haloes after heat treatment (Fig. 1C).

FIG 1.

Scheme of petri dish-based high-throughput screening (HTS) for rapid identification of thermostable PPGK mutants. (A) Mechanism of the HTS method, where glucose 6-phosphate generated by PPGK in the first reaction was used as the substrate of glucose 6-phosphate dehydrogenase (G6PDH) for reducing NAD⁺ in the second reaction. The reduced NADH reacts with TNBT in the presence of PMS, yielding a black TNBT-formazan. (B) Procedure of two-layer HTS. (C) Photo images of the mutant library containing thermostable PPGK mutants. The positive mutants featuring darker colonies with haloes are identified by red arrows.

The E. coli TOP10 is a good strain for mutant library construction because of the high transformation efficiencies (e.g., 10⁸ to 10⁹ CFU/μg plasmid DNA) (Table 1). However, its ability for recombinant protein expression is much lower than that for E. coli BL21(DE3) based on the pET expression system, which suffers from low transformation efficiency (e.g., 10⁶ CFU/μg plasmid DNA) (27). To fulfill high transformation efficiency for directed evolution in E. coli TOP10 and high protein expression in E. coli BL21(DE3), an inducible tac promoter was inserted into pET28a-ppgk (Fig. 2A). Plasmid pET28a-P_tac-ppgk consists of a strong inducible promoter, T7; a modest inducible promoter, tac; a lac operator; a ribosome binding site (RBS); and the ppgk gene. SDS-PAGE analysis shows that E. coli BL21(DE3) displayed 2.8-fold greater PPGK expression levels than E. coli TOP10 (Fig. 2B). This dual-promoter plasmid ensured high transformation efficiency of E. coli TOP10 featuring a modest amount of PPGK expression, resulting in a high library size in HTS, and it achieved a very high PPGK expression level in E. coli BL21(DE3) for easy PPGK expression and purification.

TABLE 1.

Strains, plasmids, and oligonucleotides used in this study

Strain, plasmid, or primer	Genotype, description, or sequencea	Source
Strains
E. coli BL21(DE3)	F− ompT hsdSB(rB− mB−) gal dcm (DE3)	Invitrogen
E. coli TOP10	F− mcrAΔ(mrr-hsdRMS-mcrBC) Φ80lacZΔM15 ΔlacX74 recA1 araD139 Δ(ara-leu)7697 galU galK rpsL (Strr) endA1 nupG	Invitrogen
Plasmids
pET28a		Invitrogen
pET28a-ppgk	Plasmid used for PPGK expression and a precursor of pET28a-Ptac-ppgk	This study
pET28a-Ptac-ppgk	Plasmid for expressing C-terminal His-tagged PPGK	This study
Primers
PPGK_F	5′-TAACTTTAAGAAGGAGATATACATATGGCATCTCGGGGACGGGTCGGGC-3′	This study
PPGK_R	5′-AGTGGTGGTGGTGGTGGTGCTCGAGGGCAGAGACCCGGTCACCCCCCAGC-3′	This study
28_back_F	5′-GCTGGGGGGTGACCGGGTCTCTGCCCTCGAGCACCACCACCACCACCACT-3′	This study
28_back_R	5′-GCCCGACCCGTCCCCGAGATGCCATATGTATATCTCCTTCTTAAAGTTAA-3′	This study
T7_tac_F	5′-GGCTCGTATAATGTGTGGAATTGTGAGCGGATAAC-3′	This study
T7_tac_R	5′-GATGATTAATTGTCAACCTATAGTGAGTCGTATTAATTTC-3′	This study
MPPGK-VF	5′-CGGGTCTCTPGCCCTCGAGCACCACCATCACCATCACTGAG-3′	This study
MPPGK-VR	5′-CGTCCCCGAGATGCCATATG-3′	This study
MPPGK-IF	5′-GGTCGGGCTGGGGATTGAC-3′	This study
MPPGK-IR	5′-GATGGTGGTGCTCGAGGGCAGAGACCCGGTCA-3′	This study

^aUnderlined nucleotide sequences indicate overlapped sequences.

FIG 2.

The dual T7-tac promoter plasmid for PPGK screening in E. coli TOP10 and protein expression in E. coli BL21(DE3). (A) The conceptual plasmid map of pET28a-P_tac-ppgk. The DNA sequences of P_tac, lac operator, and the RBS are shown as underlined, italic, and lowercase, respectively. (B) SDS-PAGE analysis of PPGK expression in E. coli TOP10 and E. coli BL21(DE3). M, protein marker; control, pET28a-P_tac; WT, pET28a-P_tac-ppgk; P, purified PPGK. PPGK is indicated with an arrow.

Evolution of thermostable PPGK mutants.

Error-prone PCR (ep-PCR) was performed under mutation rate-optimized conditions that can generate an average of one or two amino acid substitutions per gene, because most mutations are neutral or deleterious, and multiple mutations often lead to inactivation of the enzyme (39). The ep-PCR condition was optimized by picking up colonies randomly for sequencing. Approximately 4,000 to 10,000 colonies were subjected to the screening per round. Heat treatment conditions were optimized to deactivate most PPGKs, allowing a small fraction of positive mutants to be active. A few colonies with black haloes were observed after 30 min of color development (Fig. 1C). Approximately 100 to 150 ng of plasmid could be extracted from each positive colony, and about 60 to 80 colonies could be obtained by transforming the 50 ng of plasmid into chemically competent E. coli TOP10 cells. Plasmids of positive mutants were transformed to E. coli BL21(DE3) for recombinant protein expression, followed by PPGK purification and characterization of activity and thermostability. The gene encoding the mutant with the highest thermostability at each round was used as the parental template for the next round of evolution. The heat treatment conditions for the first-, second-, third-, and fourth-round screenings were 70, 80, 90, and 100°C for 1 h, respectively. The mutagenesis and screening process was repeated four times until the best positive mutant 4-1 containing five amino acid mutations (i.e., F30L, R34P, T72A, V88M, and A231T) can work at 70°C. The evolution of PPGK and their amino acid mutation sites are shown in Fig. 3.

FIG 3.

Lineage and protein-level mutations in the evolved PPGK mutants. The selected mutants as the starting points for the next round of evolution are marked with asterisk. Newly introduced mutations in each generation are in bold.

Characterization of PPGK mutants.

Positive mutants 1-2, 2-2, 3-1, and 4-1, along with wild-type PPGK, were purified and characterized. All mutants and the wild-type PPGK were incubated at 55°C for different lengths of time. The residual activities followed the first-order kinetics (Fig. 4A). The t_1/2 of the best mutant 4-1 at 55°C was prolonged to 10,080 min, which was approximately 7,200 times higher than that of the wild type (Table 2). The specific activities of wild-type PPGK and mutants were determined (Fig. 4B). The optimal temperature for the best mutant 4-1 was 65°C, which was 15°C higher than that of wild-type PPGK. The specific activities of mutant 4-1 at 50 and 65°C were 265 and 345 units per mg, respectively. Even at 25°C, the mutant exhibited a specific activity of 43 units per mg, 2.2 times higher than that of the wild-type PPGK.

FIG 4.

Comparison of the improved thermal stability of PPGK mutants and wild-type PPGK. (A) Inactivation time of the WT and mutants at 55°C. (B) Specific activities at temperatures from 25 to 75°C.

TABLE 2.

Thermostability and activity parameters of the wild type and mutants of PPGK

Enzyme	Relative activity at 50°C	t1/2 at 55°C (min)	Improvement of t1/2 at 55°C	Tm (°C)	Topt (°C)a
WT	100	1.4 ± 0.04	1	55.9 ± 0.1	50
1-2	120	3.2 ± 0.03	2.8	59.1 ± 0.2	50
2-2	150	240 ± 1.6	185	65.2 ± 0.2	55
3-1	290	2,200 ± 1.7	1,570	70.6 ± 0.1	60
4-1	295	10,080 ± 4.1	7,200	75.7 ± 0.2	65

^aT_opt, optical temperature.

The circular dichroism (CD) spectra (below 250 nm) of the wild type and mutants were measured, as shown in Fig. 5. The CD spectra of the mutants were similar to that of the wild-type protein (Fig. 5A), suggesting that they had similar overall secondary structures. They had absorbance peaks at 208 and 220 nm (Fig. 5A), indicating that PPGK is an enzyme with a high helical content (40). The thermal unfolding of each protein was evaluated by measuring the temperature-dependent CD response at 220 nm from 25 to 95°C. The midpoint of unfolding temperature (T_m) was determined by fitting the mean residual ellipticity versus temperature. The T_m for the wild-type PPGK was 55.9°C, and it was increased to 59.1, 65.2, 70.6, and 75.7°C for mutants 1-2, 2-2, 3-1, and 4-1, respectively (Table 2).

FIG 5.

Structure and thermal stability analysis of wild-type and mutants of PPGK. (A) CD spectra of WT and mutants at 25°C. (B) Thermal unfolding profiles of WT and mutants monitored by change in the mean residual ellipticity (mrw) at 220 nm as a function of temperature.

The apparent kinetic parameters of the wild-type PPGK and mutants were characterized at 50°C. The k_cat of mutant 4-1 was 162.5 ± 3.5 s⁻¹, 2.6 times that of wild type (61.7 ± 1.5 s⁻¹) (Table 3). Although the K_m of mutant 4-1 was increased two times to 215.2 ± 8.3 μM compared to that of the wild-type PPGK, the 4-1 mutant exhibited a k_cat/K_m value comparable to that of wild-type PPGK (Table 3).

TABLE 3.

Determination of apparent kinetic parameters at 50°C

Enzyme	Glucose parameter			Hexametaphosphate Km (μM)
	kcat (s−1)	Km (μM)	kcat/Km (s−1 · mM−1)
WT	61.7 ± 1.5	101.1 ± 9.4	610	3.8 ± 0.2
1-2	63.7 ± 1.2	114.6 ± 7.4	556	4.0 ± 0.3
2-2	101.2 ± 2.7	187.8 ± 12.6	539	5.6 ± 0.2
3-1	151.8 ± 4.0	216.8 ± 8.6	700	6.4 ± 0.3
4-1	162.5 ± 3.5	215.2 ± 8.3	755	4.3 ± 0.4

One-pot synthesis of inositol from glucose.

The best mutant, 4-1, was used with two other hyperthermophilic enzymes, inositol 1-phosphate synthase (IPS) from Archaeoglobus fulgidus, and inositol monophosphatase (IMP) from Thermotoga maritima for the production of inositol at 70°C. This in vitro synthetic pathway contained the following three cascade reactions (Fig. 6A): phosphorylation of glucose to G6P catalyzed by PPGK, isomerization of G6P to inositol 1-phosphate catalyzed by IPS, and dephosphorylation of inositol 1-phosphate to inositol and inorganic phosphate catalyzed by IMP. The reaction temperature of 70°C was chosen because hyperthermophilic IPS had very low activities when the temperature was below 60°C (6, 7). The concentration of hexametaphosphate was optimized to ensure enough phosphoryl donors for converting glucose into G6P (data not shown). Also, the concentration of Mg²⁺ was optimized because high concentrations of polyphosphate and phosphate can form water-insoluble salts with magnesium ions (5, 41) and subsequently decrease the activity of the enzymes (data not shown). When the 4-1 mutant was used, 10 g/liter glucose disappeared rapidly, and there was no glucose left after 1 h (Fig. 6B). The titer of inositol continued increasing until its productivity slowed down after 4 h. The final titer of inositol was 9.98 ± 0.20 g at 24 h, suggesting the complete utilization of glucose. The yield reached 99.8%, which is much higher than the previous study based on a mesophilic enzyme cocktail, i.e., 17% (42). In contrast, when the wild-type PPGK was used, 0.53 ± 0.10 g glucose only was utilized in the first 30 min, and the final titer of inositol reached 0.53 ± 0.06 g only.

FIG 6.

One-pot production of myo-inositol from glucose and polyphosphate by using wild-type PPGK (black) and the best mutant 4-1 (red). (A) Scheme of the in vitro synthetic pathway from glucose to inositol. (B) Profiles of inositol and consumption of glucose by two enzyme cocktails where wild-type PPGK (black) and the best mutant 4-1 (red) were used.

DISCUSSION

Compared with other PPGKs in the literature, the most thermostable and active PPGK mutant was developed by directed evolution in this work (Table 4). After four-round directed evolution, the best mutant 4-1 containing five amino acid mutations (i.e., F30L, R34P, T72A, V88M, and A231T), exhibited a 7,200-fold improvement in half-life time of thermal deactivation at 55°C, a 19.8°C increase in apparent T_m, a 15°C increase in optimum temperature, and 2.95-fold higher specific activity at 50°C compared to the wild-type PPGK.

TABLE 4.

Comparison of enzymatic properties of characterized PPGKs^a

Organism	Sp act (U/mg) (pH, temp [°C])	Half-life (min) (pH, temp [°C])	Reference or source
Mycobacterium tuberculosis	203 (7.5, 30)	NA	11
P. shermanii	15.3 (7.5, 30)	NA	12
M. phosphovorus	NA	<5 (7.5, 25)	13
Arthrobacter sp.	220 (7.0, 30)	5 (7.5, 40)	14
C. glutamicum	21.7 (7.4, 30)	20 (7.4, 50)	15
Anabaena sp. strain PCC 7120	107 (7.5, 28)	NA	16
Nostoc sp. strain PCC7120	229.1 (9.0, 30)	26.7 (9.0, 50)	17
T. fusca	64.6 (7.5, 50)	15 (7.5, 50)	2
T. fusca	285.3 (7.5, 50)	10,080 (7.5, 50)	This study

^aNA, not applicable.

An HTS method for facilitating the identification of thermostable PPGK mutants was established based on the color development of the product of PPGK, G6P. Different from typical solid agar-based media, highly thermostable Phytagel, which can remain solid for 1 h at 120°C, was used for screening in directed evolution. To the best of our knowledge, Phytagel was normally used for the cultivation and isolation of hyperthermophilic bacteria (43) or plant tissue (44). Herein, we adopted it for the application of Phytagel-based plates as a thermotolerant gelling agent to replace the use of agar plates at enzyme engineering for enhanced thermostability by directed evolution. This expansion made it feasible for screening hyperthermophilic enzyme mutants and decreasing the background noise from E. coli mesophilic enzymes and metabolites (e.g., NADPH and NADH). Also, glass plates instead of plastic dishes should be used in the screening procedure, because plastic dishes were deformed above 90°C. The second layer was composed of the substrates for the following cascade enzyme reaction coupled with a colorimetric assay. This strategy was an extension of the previous method for detecting dehydrogenase that generates NAD(P)H (27). This Phytagel-based HTS method could be widely used for screening other ultrathermostable enzymes.

Structural analysis provides some insights into the structural basis of both the thermostability and activity for these thermostabilizing mutations (Fig. 7). The best mutant, 4-1, had five mutation sites (F30L, R34P, T72A, V88M, and A231T). Four residues, F30L, R34P, V88M, and A231T, were located in or near the loop region and distributed near the protein surface of PPGK, while T72A was buried in the active cleft and within 5 Å of catalytic site D114 (Fig. 7A). This was in good agreement with the fact that both loop and surface regions are often targeted sites for the improvement of thermostability (37). The inspection of the structure model of the 4-1 mutant revealed possible mechanisms of thermostability. The replacement of a bulky and planar residue F30 with leucine may stabilize the protein by decreasing its desolvation-accessible surface area (45). R34P was responsible for stabilizing the protein for two reasons. (i) Prolines have restricted conformations compared to other residues and thus lower the entropy of the denatured state, which is thermodynamically stabilizing. This effect is pronounced in loops that are less structured and more flexible (46). (ii) F30L and R34P mutations may result in a cooperative effect due to their hydrophobic interactions. The Sδ atom of M88 makes hydrophobic interactions with A43, P74, and W92 to fill in the deeply buried cavity surrounded by several loops (47). The substitution of alanine to threonine at position 231 stabilized the protein by introducing a new hydrogen bond with E202. Interestingly, the sequence comparison of PPGKs regardless of hypothetically thermostable and mesophilic sources suggests that threonine is the most prevalent amino acid at position 231 (data not shown). Therefore, the A231T substitution could be considered a type of “back-to-consensus” mutation, which has been shown in several cases to increase protein stability (48). Different from the four surface mutations, T72A was located in a β-sheet near the active center (Fig. 7B). The mutation was validated to be responsible for enhanced activity and thermostability by experiments of back mutation (data not shown). The replacement of a big polar threonine with a small nonpolar alanine may enlarge the binding cavity of protein and bring a benefit for the product release (38, 49). Given the above-described analysis, back mutations of each mutated residue in the 4-1 mutant would be done in the future to precisely analyze the effect of specific mutation on enzymatic thermostability and activity. The PPGK is expected to be further thermostabilized by screening a greater number of thermostable mutants from the saturation mutagenesis libraries of thermostabilized residues, followed by the recombination of new beneficial sites.

FIG 7.

Structure analysis of mutants. (A) Distribution of the mutation sites (F30L, R34P, T72A, V88M, and A231T) and the active site D114, where mutation residues in the best variant 4-1 are indicated in spherical representation, and the residue of the active site is marked as sticks. (B) Local environment of mutation sites.

Industrial biocatalysts are expected to have both key characteristics of high activity and thermostability at the same time (50). However, studies of thermally adapted proteins led to the belief that there is an inherent trade-off between the rigidity necessary for stability and the flexibility required for activity (30, 31, 51, 52). Nevertheless, our result demonstrated that the two properties are not incompatible or even inversely correlated. The often-observed trade-off between stability and activity in enzymes evolved for different temperatures is more like a genetic drift during the course of evolution (30, 53). Because single mutations that improve two or more properties are rare, it might be more efficient to direct the evolution of multiple independent properties, followed by recombination of those properties that evolved separately (30). Extensive screening based on restrained conditions allowed the identification and incorporation of rare mutations that simultaneously improve both properties (29, 54).

In conclusion, this study showed the coevolution of both the thermostability and catalytic activity of PPGK by directed evolution. An HTS approach was developed for rapid identification of thermostable PPGK. The most thermostable PPGK mutant was demonstrated in theoretical yield production of myo-inositol from glucose, far better than mesophilic enzymes (42). This study suggested great potentials of directed evolution in obtaining high-performance BioBricks suitable for the in vitro synthetic biology platform.

MATERIALS AND METHODS

Chemicals and materials.

All chemicals were reagent grade or higher and purchased from Sigma-Aldrich (St. Louis, MO, USA) or Sinopharm (Shanghai, China), unless otherwise noted. Sodium hexametaphosphate was purchased from Dingshenghui (Tianjin, China). T. fusca YX genomic DNA was a gift from David Wilson at Cornell University (Ithaca, NY, USA). The Luria-Bertani (LB) medium with 50 μg/ml kanamycin was used for E. coli cell growth and recombinant protein expression. All enzymes for molecular biology experiments were purchased from New England BioLabs (NEB, Ipswich, MA, USA). Super GelRed was used as the DNA gel stain (US Everbright, Inc., Suzhou, China). The strains, plasmids, and oligonucleotides used in this study are listed in Table 1.

Construction of pET28a-P_tac-ppgk.

Plasmid pET28a-ppgk was constructed as follows. The inserted ppgk gene was amplified from T. fusca genomic DNA by using the PPGK_F/PPGK_R primer pair, and the linearized vector backbone was amplified from pET28a by using the 28_back_F/28_back_R primer pair. The insert and vector backbone were assembled into multimeric DNA by prolonged overlap extension-PCR (55). The PCR product was directly transformed into E. coli TOP10, yielding pET28a-ppgk. To make the dual-promoter plasmid pET28a-P_tac-ppgk, the linear backbone was amplified based on pET28a-ppgk by using a pair of 5′ phosphorylated primers, T7_tac_F/T7_tac_R, containing each half of the tac promoter, and was self-ligated using the NEB Quick Ligation kit. After transformation in E. coli TOP10, plasmid pET28a-P_tac-ppgk was obtained.

Mutant library construction.

Mutant libraries of ppgk gene were generated by error-prone PCR (ep-PCR) from pET28a-P_tac-ppgk as the template DNA with a pair of primers, MPPGK-IF and MPPGK-IR. MPPGK-IF was purchased as a 5′-end phosphorylated primer. The last 28 bp of the 5′ terminus of MPPGK-IR was the complementary sequence to the vector backbone. The ep-PCR solution with a total volume of 50 μl contained 1 ng/μl plasmid pET28a-P_tac-ppgk, 0.2 mM dATP, 0.2 mM dGTP, 1 mM dCTP, 1 mM dTTP, 5 mM MgCl₂, 0.05 mM MnCl₂, 0.4 μM primers (MPPGK-IF and MPPGK-IR), and 0.05 U/μl NEB Taq polymerase. The ep-PCR conditions were initial denaturation (120 s at 94°C), 18 cycles of denaturation (30 s at 94°C), annealing (30 s at 56°C), and extension (45 s at 68°C), and a final extension step (5 min at 68°C). The linear vector fragment was amplified from pET28a-P_tac-ppgk by using the NEB high-fidelity Q5 DNA polymerase with the MPPGK-VF and MPPGK-VR primer pair. The first 28 bp of the 5′ terminus of MPPGK-VF was the complementary sequence to the insert fragment. MPPGK-VR was purchased as a 5′-end phosphorylated primer. Fifty microliters of the reaction mixture solution contained pET28a-P_tac-ppgk (0.02 ng/μl), dinucleoside triphosphates (dNTPs; 0.2 mM each), primers (0.5 μM each), and the Q5 DNA polymerase (0.02 U/μl). The PCR was conducted as follows: initial denaturation (60 s at 98°C), 30 cycles of denaturation (20 s at 98°C), annealing (20 s at 60°C), and extension (72 s at 72°C), and a final extension step (5 min at 72°C). The PCR products were recovered from 0.8% (wt/vol) agarose gel by using a TIANgel Midi purification kit (Tiangen, Beijing, China).

The gel-purified insert and vector fragments were spliced through overlap extension PCR (OE-PCR) by using the NEB high-fidelity Q5 DNA polymerase (56). The reaction solution, with a total volume of 50 μl, contained dNTPs (0.2 mM each), 0.5 μM primers (each MPPGK-IF and MPPGK-VR), the insert fragment (2 ng/μl), the vector fragment (equimolar with the insert fragment), and the Q5 DNA polymerase (0.02 U/μl). OE-PCR was conducted as follows: initial denaturation (60 s at 98°C), 16 cycles of denaturation (20 s at 98°C), annealing (20 s at 60°C), and extension (2 min at 72°C), and a final extension step (5 min at 72°C). The OE-PCR product was simply purified by using the TIANquick Midi purification kit (Tiangen, Beijing, China). The purified linear plasmid DNAs were self-ligated by using the NEB Quick Ligation kit at 25°C for 5 min. The ligation product was transformed into the competent E. coli TOP10 cells.

Screening of a mutant library.

The double-layer screening of thermostable PPGK mutants was performed as follows. After the transformation of the plasmid mutant library, the E. coli TOP10 cells were spread on the 1.2% Phytagel LB medium containing 50 μg/ml kanamycin, with an expected colony number of 400 to 500 per petri dish. The glass dishes were incubated at 37°C overnight. For the first-round screening, the colonies on plates were heat treated at 70°C for 1 h to kill cells, deactivate E. coli mesophilic enzymes and unstable PPGK variants, and degrade metabolites and intracellular coenzymes. Eight milliliters of 0.5% (wt/vol) melted agarose solution (60°C) containing 50 mM HEPES buffer (pH 7.5) containing 50 μM tetranitroblue tetrazolium (TNBT), 10 μM phenazine methosulfate (PMS), 5 mM glucose, 1 mM sodium hexametaphosphate, 4 mM Mg²⁺, 0.5 mM NAD⁺, and 1 U/ml glucose 6-phosphate dehydrogenase (G6PDH) was carefully poured on the heat-treated colonies. After incubation at room temperature for 30 min, positive colonies were identified based on the formation of black haloes. The identified colonies embedded in the agarose gel were taken out by a sterile toothpick and then suspended in 100 μl of the P1 buffer of the Tiangen plasmid miniprep kit (Tiangen, Beijing, China), followed by the Tiangen plasmid purification protocol. A tiny amount of plasmid extracted from the agarose gel was transformed into chemically competent E. coli TOP10 cells for plasmid amplification, purification, and DNA sequencing. For the second-, third-, and fourth-round screenings, the plates were heat treated at 80, 90, and 100°C for 1 h, respectively.

Overexpression and purification of wild-type PPGK and mutants.

Plasmid pET28a-P_tac-ppgk containing wild-type ppgk gene or mutants was transformed to E. coli BL21(DE3) for protein overexpression in LB medium with 50 μg/ml kanamycin at 37°C. The protein expression was initiated by adding 0.1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) until the absorbance at 600 nm reached ∼0.6 to 0.8. The protein expression was induced at 37°C for 6 h. Cell pellets were harvested by centrifugation and then were resuspended in a 50 mM HEPES buffer (pH 7.5) containing 0.5 M NaCl and 20 mM imidazole. After sonication, the supernatant was loaded onto the column packed with nickel-charged Sepharose 6 Fast Flow resin (GE, Sweden) and then eluted with 50 mM HEPES buffer (pH 7.5) containing 0.5 M NaCl and 250 mM imidazole. The mass concentration of protein was determined by the Bradford assay, with bovine serum albumin as the standard. The PPGK expression levels and purified PPGK were checked by SDS-PAGE and analyzed by using densitometry analysis of the Image Lab software (Bio-Rad, Hercules, CA, USA).

PPGK activity assays.

PPGK activity was measured by using a discontinuous approach, as described elsewhere (2). For the determination of the optimal temperature, 2 μg/ml PPGK activity was measured from 25°C to 75°C. The activity was measured based on the generation of G6P from polyphosphate and glucose in a 50 mM HEPES buffer (pH 7.5) containing 4 mM Mg²⁺, 5 mM glucose, and 1 mM sodium hexametaphosphate for 15 min. The formation of G6P was measured in 50 mM HEPES buffer (pH 7.5) containing 1 mM NAD⁺ and 0.5 U/ml G6PDH at 25°C for 10 min. An increase in absorbance at 340 nm due to the formation of NADH (ε₃₄₀ = 6.22 mM⁻¹ · cm⁻¹) was measured by a UV spectrophotometer. One unit of PPGK activity was defined as one micromole G6P generated per minute.

For the determination of the half-life, samples containing 10 μg/ml PPGK were incubated at 55°C for different times. The samples were withdrawn and chilled on ice for 10 min, and the residual activity was determined at 25°C. The natural logarithm of the residual activity was a linear function of the inactivation time: ln (residual activity) = −k_dt, where k_d was the inactivation rate and t is the inactivation time. The half-life (t_1/2) was calculated as: t_1/2 = ln(2)/k_d.

For determining apparent kinetic parameters, 2 μg/ml PPGK activities were measured as described above, whereas the concentration of glucose was changed from 5 to 5,000 μM and the concentration of hexametaphosphate was changed from 0.1 to 100 μM. The apparent K_m and k_cat values on glucose and hexametaphosphate were calculated based on Michaelis-Menten nonlinear regression by using GraphPad Prism 5 (GraphPad Software, Inc., La Jolla, CA, USA).

Circular dichroism analysis.

Circular dichroism spectra were measured on Chirascan (Applied Photophysics, Surrey, UK) at a protein concentration of 0.2 mg/ml in a 1-mm path-length quartz cuvette in 50 mM HEPES buffer (pH 7.5). To eliminate signal baseline drift, the spectropolarimeter and xenon lamp were warmed up for at least 15 min prior to each experiment. Full wavelength data were collected at 25°C, with the wavelengths ranging from 200 to 250 nm. A temperature ramp was applied to the sample at a rate increase of 1°C per min from 25 to 95°C. Ellipticity data of the enzymes in 220 nm were collected, and the integration time for data acquisition was set to 5 min, from which the spectrum of a buffer blank was subtracted. The mean residue ellipticity [θ] at 220 nm was plotted against temperature, and the temperature midpoint of the unfolding curve was determined by data fitting to the Boltzmann model by using GraphPad Prism 5 (GraphPad Software, La Jolla, CA, USA).

Structure analysis.

The three-dimensional structure modeling of wild-type PPGK from T. fusca and mutants was built by SWISS-MODEL (http://swissmodel.expasy.org/) based on the crystal structure of the Arthrobacter sp. PPGK (PDB 1WOQ). The wild-type of PPGK from T. fusca shared 59.8% sequence identity with the template of Arthrobacter sp. PPGK. The stereochemical quality of the homologue structure was checked by PROCHECK (http://services.mbi.ucla.edu/SAVES/). The residue interactions were analyzed by Discovery Studio 3.5 Client (Accelrys, San Diego, CA, USA).

Production of inositol.

The production of inositol from glucose was conducted with 1 ml of the reaction volume containing 100 mM HEPES buffer (pH 7.5), 40 mM MgCl₂, 10 g/liter glucose, 20 mM sodium hexametaphosphate, 12.5 μg/ml PPGK, 2 U/ml inositol 1-phosphate synthase (IPS) from Archaeoglobus fulgidus, and 2 U/ml inositol monophosphatase (IMP) from Thermotoga maritima (6). The reaction mixture was mixed gently and incubated at 70°C. An aliquot (65 μl) of the reaction mixture sample was withdrawn and then mixed with 35 μl of 1.88 M HClO₄ to stop the reaction. The pH value of the reaction solution was adjusted to neutral with 13 μl of 5 M KOH. The production of inositol was monitored using high-performance liquid chromatography (HPLC; Shimadzu) equipped with a Bio-Rad HPX-87H column with 5 mM H₂SO₄ as a mobile phase and a refractive index detector (6). The residual glucose concentration was determined with the glucose detection kit based on glucose oxidase. G6P was determined by the G6P assay kit containing 100 mM HEPES buffer (pH 7.2), 1 U/ml G6PDH, 0.5 mM NAD⁺, and 5 mM MgCl₂.