Mutagenesis of the Glucose-1-Phosphate-Binding Site of Potato Tuber ADP-Glucose Pyrophosphorylase

Abstract

Lysine (Lys)-195 in the homotetrameric ADP-glucose pyrophosphorylase (ADPGlc PPase) from Escherichia coli was shown previously to be involved in the binding of the substrate glucose-1-phosphate (Glc-1-P). This residue is highly conserved in the ADPGlc PPase family. Site-directed mutagenesis was used to investigate the function of this conserved Lys residue in the large and small subunits of the heterotetrameric potato (Solanum tuberosum) tuber enzyme. The apparent affinity for Glc-1-P of the wild-type enzyme decreased 135- to 550-fold by changing Lys-198 of the small subunit to arginine, alanine, or glutamic acid, suggesting that both the charge and the size of this residue influence Glc-1-P binding. These mutations had little effect on the kinetic constants for the other substrates (ATP and Mg²⁺ or ADP-Glc and inorganic phosphate), activator (3-phosphoglycerate), inhibitor (inorganic phosphate), or on the thermal stability. Mutagenesis of the corresponding Lys (Lys-213) in the large subunit had no effect on the apparent affinity for Glc-1-P by substitution with arginine, alanine, or glutamic acid. A double mutant, S_K198RL_K213R, was also obtained that had a 100-fold reduction of the apparent affinity for Glc-1-P. The data indicate that Lys-198 in the small subunit is directly involved in the binding of Glc-1-P, whereas they appear to exclude a direct role of Lys-213 in the large subunit in the interaction with this substrate.

ADPGlc PPase (EC 2.7.7.27) catalyzes an important regulatory step in the biosynthesis of starch in plants and of glycogen in bacteria (Preiss, 1988, 1991, 1997; Preiss and Sivak, 1996). This enzyme mediates the synthesis of ADPGlc and PPi from Glc-1-P and ATP. The product, ADPGlc, serves as the activated glucosyl donor in α-1,4-glucan synthesis. ADPGlc PPase from higher plants is heterotetrameric and encoded by two different genes (Smith-White and Preiss, 1992; Preiss and Sivak, 1996), whereas the enzyme from enterobacteria and cyanobacteria is homotetrameric in structure. The small subunit of higher-plant ADPGlc PPases is highly conserved, whereas the similarity among different large subunits is lower (Smith-White and Preiss, 1992). It has been speculated that the two plant subunits were originally derived from the same gene. This gene was duplicated during evolution, and then the two polypeptides diverged in sequence. Both subunits are required for optimal activity (Iglesias et al., 1993; Ballicora et al., 1995).

Studies based on a wide range of sources have shown that ADPGlc PPase is regulated by effectors derived from the dominant carbon-assimilation pathway in the organism. The enzyme from higher plants (Preiss, 1988, 1991), green algae (Sanwal and Preiss, 1967; Ball et al., 1991), and cyanobacteria (Iglesias et al., 1991) is mainly activated by 3PGA and inhibited by Pi. ADPGlc PPases from enteric bacteria are activated by Fru-1,6-bisP and inhibited by AMP (Preiss and Romeo, 1989).

The potato (Solanum tuberosum L.) tuber ADPGlc PPase consists of two different subunits of 51 and 50 kD (Okita et al., 1990). The cDNAs encoding the large and small subunits of the potato tuber ADPGlc PPase have been expressed in Escherichia coli, and yielded a recombinant heterotetrameric enzyme with properties similar to those of the native enzyme purified from potato tuber (Ballicora et al., 1995). It was found that the homotetrameric enzyme composed of only small subunits exhibited catalytic activity, with allosteric regulatory properties different from those of the heterotetrameric enzyme. In contrast, the large subunit by itself has negligible catalytic activity (Ballicora et al., 1995). It was proposed, therefore, that the major function of the large subunit is to modulate the sensitivity of the small subunit to allosteric regulation by Pi and 3PGA, and the major function of the small subunit is catalysis. However, the precise role of the small and large subunits is not well understood. Recent studies show that the putative activator sites in the small subunit are more important in the regulation of the potato tuber enzyme than the homologous residues in the large subunit (Ballicora et al., 1998).

Chemical modification and site-directed mutagenesis studies have determined that Lys-195 in ADPGlc PPase from E. coli is involved in binding of the substrate Glc-1-P (Parsons and Preiss, 1978; Hill et al., 1991). This residue is highly conserved in the bacterial enzymes as well as in both the small and large subunits of plant ADPGlc PPases (Preiss and Sivak, 1996). However, no studies have been done to investigate the function of this highly conserved Lys residue in plant ADPGlc PPases, which consist of two different subunits. The expression system for potato tuber ADPGlc PPase provides a useful tool with which to characterize the role of the corresponding Lys residues in the small subunit (Lys-198) as well as in the large subunit (Lys-213) and to determine if they retain the same function as in the homotetrameric E. coli enzyme.

MATERIALS AND METHODS

Reagents

ATP, ADPGlc, Glc-1-P, Man-1-P, Gal-1-P, GlcUA-1-P, 3PGA, and PPi were purchased from Sigma. ³²PPi and [8-¹⁴C]ATP were purchased from DuPont-NEN. [U-¹⁴C]Glc-1-P was from ICN. [α-³⁵S]dATP and the in vitro mutagenesis kit were from Amersham. Enzymes for DNA manipulation and sequencing were from New England Biolabs and United States Biochemical, respectively. Oligonucleotides were synthesized and purified by the Macromolecular Structure Facility at Michigan State University. All other reagents were of the highest available commercial grade.

Bacterial Strains and Media

Escherichia coli strain TG1 (F′traD36 lacI^qΔ[lacZ]M15 proA⁺B⁺/supE Δ[hsdM-mcrB]5[r_k⁻m_k⁻McrB⁻] thi Δ[lac-proAB]) was used for site-directed mutagenesis. E. coli mutant strain AC70R1–504 (Carlson et al., 1976), which exhibited negligible ADPGlc PPase activity, was used for expression of the potato tuber ADPGlc PPase gene (Ballicora et al., 1995). Both E. coli strains were grown in Luria-Bertani medium.

Site-Directed Mutagenesis

For mutagenesis, the gene for the small subunit of potato tuber ADPGlc PPase was subcloned as an EcoRI fragment from pML 10, a plasmid containing the cDNA gene of the small subunit (Ballicora et al., 1995), into the EcoRI site of M13mp18RF. The gene for the large subunit was subcloned as an XbaI fragment from pMON17336, a plasmid containing the cDNA gene of the large subunit (Iglesias et al., 1993; Ballicora et al., 1995), into the XbaI site of M13mp19RF. For the small subunit after mutagenesis, the mutated EcoRI fragment was exchanged with the unmutated EcoRI fragment in pML 10; for the large subunit, the mutated XbaI fragment was exchanged with the unmutated XbaI fragment in pMON17336. Site-directed mutagenesis experiments were performed according to a previously described method (Sayers et al., 1988) using the in vitro site-directed mutagenesis kit from Amersham. Three heterotetrameric mutant enzymes with a single substitution of Arg, Ala, and Glu at Lys-198 of the small subunit were designated as S_K198RL_wt, S_K198AL_wt, and S_K198EL_wt, respectively. The mutant enzymes with the substitution of Arg, Ala, and Glu at Lys-213 of the large subunit were designated as S_wtL_K213R, S_wtL_K213A, and S_wtL_K213E, respectively. The double-mutant enzyme, in which both Lys-198 of the small subunit and Lys-213 of the large subunit were replaced with Arg, was designated as S_K198RL_K213R. The oligonucleotides used to create the desired mutations are shown in Figure 1. Before expression of the mutant enzymes, the entire coding regions of these mutant alleles were sequenced to verify that there were no undesired mutations.

Figure 1.

Nucleotide sequence and encoded protein sequence of the potato tuber ADPGlc PPase gene in the region of Lys-198 in the small subunit and Lys-213 in the large subunit. The synthetic oligonucleotides used for site-directed mutagenesis at these positions are shown beside the corresponding mutants they created. The codons for position 198 in the small subunit and the anticodons for position 213 in the large subunit are underlined.

Expression and Purification of Mutant and Wild-Type Enzymes

The single mutant enzymes were obtained by co-expressing the mutated plasmid pML 10 (or pMON17336) with unmutated plasmid pMON17336 (or pML 10) in E. coli mutant strain AC70R1-504. The double-mutant enzyme was obtained by co-expressing the two mutated plasmids in AC70R1-504. Mutant enzyme S_K198RL_wt was expressed as described previously (Ballicora et al., 1995). The other mutant and wild-type enzymes were expressed in the same manner except that the concentration of isopropyl-β-d-thiogalactopyranoside was increased from 10 μm to 0.5 mm for induction. An improved procedure over the one in the previous study (Ballicora et al., 1995) was used for the purification of the wild-type and mutant enzymes. In the hydrophobic-interaction-chromatography step, the enzyme was loaded onto the column in the presence of 1.2 m ammonium sulfate instead of 1.3 m potassium phosphate buffer. After the heat treatment step, a 50% saturation ammonium-sulfate-precipitation step was added, after which the pellet was dissolved in a minimal volume of extraction buffer (100 mm Hepes-NaOH, pH 8.0, 5 mm MgCl₂, 1 mm EDTA, and 20% [w/v] Suc) and dialyzed against the same buffer overnight. The dialyzed enzyme was centrifuged at 20,000g for 15 min at 4°C to remove insoluble material before being loaded onto the DEAE Fractogel column (EM Separations Technology, Gibbstown, NJ).

Assay of ADPGlc PPase

Assay I

In the pyrophosphorolysis direction, enzyme activity was assayed according to a previously described method (Morell et al., 1987). The reaction mixture contained 80 mm glycyl-Gly, pH 8.0, 2 mm ADPGlc, 5 mm MgCl₂, 3 mm DTT, 2 mm ³²PPi (1000–2000 cpm nmol⁻¹), 3 mm 3PGA, 10 mm NaF, 200 μg mL⁻¹ BSA, and enzyme in a total volume of 250 μL. The assay conditions for the mutant enzymes were identical to those for the wild-type enzymes except that the amounts of MgCl₂ and ADPGlc were altered for some mutant enzymes to obtain maximal activity. The amount of MgCl₂ was increased to 10 mm for S_K198AL_wt and S_K198EL_wt, and to 20 mm for S_K198RL_wt and S_K198RL_K213R. For the S_K198EL_wt and S_K198RL_wt enzymes, 3 mm ADPGlc was used.

Assay II

In the ADPGlc-synthesis direction, enzyme activity was measured according to a previously described method (Preiss et al., 1966). The reaction mixture contained 100 mm Hepes-NaOH, pH 8.0, 0.5 mm [U-¹⁴C]Glc-1-P (1000–3000 cpm nmol⁻¹), 1.5 mm ATP, 5 mm MgCl₂, 3.0 mm 3PGA, 3 mm DTT, 200 μg mL⁻¹ BSA, 0.3 unit of inorganic pyrophosphatase, and enzyme in a final volume of 200 μL. For assay of the S_K198RL_wt, S_K198AL_wt, S_K198EL_wt, and S_K198RL_K213R mutant enzymes, [8-¹⁴C]ATP (about 200–500 cpm nmol⁻¹) instead of [¹⁴C]Glc-1-P was used to monitor the synthesis of ADPGlc, and the amounts of Glc-1-P and MgCl₂ were increased to 30 and 20 mm, respectively, to obtain maximal activity. For the double-mutant S_K198RL_K213R enzyme, the 3PGA concentration was increased to 10 mm in addition to the changes mentioned above. Because of the high content of Glc-1-P in the assay mixture, the time for the alkaline phosphatase digestion was extended to overnight. Control experiments showed that the product, [¹⁴C]ADPGlc, was stable during the overnight digestion.

Kinetic Studies

For determination of kinetic parameters, the concentration of the substrate or effectors tested was systematically varied with the other substrates and effectors fixed at a saturating concentration, as described for assays I and II. Kinetic data were plotted as initial velocity versus substrate or effector concentration. The kinetic constants A_0.5, S_0.5, and I_0.5, which correspond to the concentration of activator, substrate, or inhibitor giving 50% of maximal activation, velocity, or inhibition, respectively, as well as the interaction coefficient n_H were obtained from a computer program using a nonlinear, iterative, least-squares fitting to a modified Michaelis-Menten equation (Canellas and Wedding, 1980).

Sugar-Phosphate Specificity

Reactions were performed in the ADPGlc-synthesis direction (assay II) with different sugar phosphates substituted for Glc-1-P. [8-¹⁴C]ATP (about 200–500 cpm nmol⁻¹) was used to monitor the assay. The reaction time was extended to 30 min at 37°C. Under the conditions with Glc-1-P as the substrate, the production of ADPGlc was shown to be linear for both wild-type and mutant enzyme S_K198AL_wt. The time for the alkaline phosphatase digestion was extended to 48 h. The extent of the reaction was controlled by varying the amount of enzyme used for the assay.

Thermal Stability

Enzyme samples were diluted to give the same final protein concentration (0.3 mg mL⁻¹) in a final volume of 20 μL. Dilution buffer was 50 mm Hepes-NaOH, pH 8.0, 5 mm MgCl₂, 1 mm EDTA, 20% (w/v) Suc, and 1 mg mL⁻¹ BSA. The enzyme samples were heated for 5 min in a water bath equilibrated at 60°C, then immediately placed on ice. The enzyme activities were assayed in the ADPGlc-synthesis direction, as described in “Assay II.”

Protein Assay

Protein concentration was measured using the Pierce bicinchoninic acid reagent (Smith et al., 1985) with BSA as the standard.

Protein Electrophoresis and Immunoblotting

SDS-PAGE was performed according to the method of Laemmli (1970) on 10% polyacrylamide slab gels. After electrophoresis, proteins on the gel were visualized by staining with Coomassie brilliant blue R-250 or electroblotted onto a nitrocellulose membrane (Burnette, 1981). After electroblotting the nitrocellulose membrane was treated with affinity-purified rabbit anti-spinach leaf ADPGlc PPase IgG, and the antigen-antibody complex was visualized by treatment with alkaline phosphatase-linked goat anti-rabbit IgG followed by staining with purple alkaline phosphatase-substrate precipitating reagent (Boehringer Mannheim).

RESULTS

Expression and Purification of Mutant Enzymes

Wild-type and mutant enzymes of potato tuber ADPGlc PPase were identified by immunoblotting with antibody prepared against the spinach leaf ADPGlc PPase. It has been shown that the small subunit of the potato tuber enzyme cross-reacts significantly with the spinach leaf ADPGlc PPase antibody (Okita et al., 1990). Under the same conditions the expression level of all of the mutant enzymes was similar to that of the wild type based on the results of immunoblotting. The apparent sizes of these mutant polypeptides were the same as that of the wild type, having a molecular mass of about 50 kD. Mutant enzyme S_K198RL_wt was purified to more than 85% homogeneity, as estimated from about 4 μg of protein electrophoresed on SDS-PAGE. All of the other enzymes were purified to greater than 95% homogeneity.

Kinetic Characterization of S_K198L_wt Mutant Enzymes

The apparent affinity for Glc-1-P decreased dramatically when Lys-198 in the small subunit was mutated to either Arg, Ala, or Glu. The S_0.5 values for Glc-1-P of the S_K198RL_wt, S_K198AL_wt, and S_K198EL_wt enzymes were about 135-, 400-, and 550-fold higher, respectively, than that of the wild-type enzyme (Table I). Substitution of Lys-198 in the small subunit by Ala and Glu resulted in such large S_0.5 changes that they could not be accurately determined (Fig. 2A). The highest Glc-1-P amount used in the assay mixture was 40 mm, which is about two times higher than the S_0.5 of the Ala mutant and slightly higher than the S_0.5 of the Glu mutant, due to a solubility problem. The values for n_H were changed from 1.1 for the wild type to 1.3 to 1.8 for the mutant enzymes. The changes of the V_max value in the synthesis direction were 2-fold or less, and in the pyrophosphorolysis direction, they were less than 4-fold, except for mutant S_K198EL_wt. This suggests that Lys-198 in the small subunit has virtually no role in the rate-determining step of catalysis. This is consistent with the observation on the corresponding Lys-195 of the E. coli ADPGlc PPase (Hill et al., 1991). Replacement with Glu not only caused the largest increase of S_0.5 value for Glc-1-P, but also substantially decreased the catalytic efficiency relative to the wild-type enzyme.

Table I.

Comparison of apparent affinity for substrates of potato tuber wild-type and mutant ADPGlc PPases

ADPGlc PPase	Synthesis Direction		Pyrophosphorolysis Direction
	S0.5 for Glc-1-P	Vmax	S0.5 for ADPGlc	Vmax
	mm	units mg−1	mm	units mg−1
Wild type	0.057  ± 0.003 (1.1)	48  ± 1	0.20  ± 0.01 (1.3)	55  ± 1
SK198RLwt	7.7  ± 0.1 (1.3)	24  ± 1	0.49  ± 0.02 (1.1)	16  ± 1
SK198ALwt	22.0  ± 2.5 (1.5)	46  ± 3	1.3  ± 0.1 (1.0)	20  ± 1
SK198ELwt	31.1  ± 2.7 (1.8)	1.7  ± 0.1	2.1  ± 0.2 (1.0)	1.4  ± 0.1
SwtLK213R	0.044  ± 0.002 (1.1)	27  ± 1	0.36  ± 0.03 (1.8)	37  ± 1
SwtLK213A	0.037  ± 0.001 (1.0)	25  ± 1	0.46  ± 0.02 (1.7)	31  ± 1
SwtLK213E	0.036  ± 0.001 (0.9)	31  ± 1	0.68  ± 0.01 (1.8)	45  ± 1
SK198RLK213R	5.6  ± 0.1 (1.5)	24  ± 1	0.72  ± 0.01 (1.9)	14  ± 1

Reactions were performed in either the synthesis (assay II) or the pyrophosphorolysis direction (assay I) as described in Methods. Data represent the average of two identical experiments ± the average difference of the duplicates. The values in parentheses are the Hill interaction coefficients (n_H). One unit of enzyme activity is expressed as the amount of enzyme required to form 1 mol of ADPGlc per minute at 37°C assayed in either the synthesis or the pyrophosphorolysis direction.

Figure 2.

Glc-1-P dependence for wild-type and mutant enzymes. A, For wild-type (•), S_K198RL_wt(▴), S_K198AL_wt(▪), and S_K198EL_wt(○), 100% activity corresponds to 1.2, 16.6, 33.7, and 39.0 nmol 10 min⁻¹, respectively. B, For wild-type (•) and S_wtL_K213R (□), 100% activity corresponds to 1.2 and 1.6 nmol 10 min⁻¹, respectively. Initial velocities of the enzymes were determined in the ADPGlc-synthesis direction (assay II), as described in “Materials and Methods,” with the concentration of Glc-1-P being varied. The amounts of the wild-type, S_K198RL_wt, S_K198AL_wt, S_K198EL_wt, and S_wtL_K213R proteins were about 2.5 × 10⁻³, 70 × 10⁻³, 80 × 10⁻³, 2.4 × 10⁻³, and 5.5 × 10⁻³ μg, respectively.

Considering the large effect of the mutation on the S_0.5 of Glc-1-P, it was not surprising to see that the apparent affinity of ADPGlc was affected (Table I). Indeed, the S_0.5 value of ADPGlc was increased 2.5- to 10- fold. The kinetic constant for the other substrate, PPi, was increased 3- to 6-fold that for the mutant enzymes (Table II). However, these changes were relatively small compared with the change of the S_0.5 value for Glc-1-P. Overall, the various mutations at position 198 in the small subunit caused little or no alteration in the apparent affinities for the other substrates (ATP and Mg²⁺) and activator (3PGA) (Table II). The data suggest that the conformations of these ligand-binding sites were relatively unchanged. The 2- to 7-fold increase of the I_0.5 value for the inhibitor Pi is relatively small, but the inhibitor site may be in close proximity to the Glc-1-P-binding site.

Table II.

Kinetic parameters of the potato tuber wild-type and mutant ADPGlc PPases

ADPGlc PPase	Synthesis Direction			I0.5 for Pi	Pyrophosphorolysis Direction
	S0.5		A0.5 for 3PGA		S0.5 for PPi
	for ATP	for Mg2+
	μm	mm	mm		μm
Wild type	76  ± 2 (1.6)	2.2  ± 0.1 (3.7)	0.14  ± 0.01 (0.9)	1.4  ± 0.1 (1.4)	41  ± 3 (1.0)
SK198RLwt	119  ± 1 (1.5)	3.7  ± 0.1 (2.1)	0.39  ± 0.04 (1.3)	2.8  ± 0.2 (1.3)	210  ± 10 (1.5)
SK198ALwt	130  ± 4 (1.3)	2.8  ± 0.1 (2.5)	0.15  ± 0.01 (1.2)	5.0  ± 0.1 (1.1)	135  ± 5 (1.5)
SK198ELwt	102  ± 24 (1.4)	2.6  ± 0.1 (1.7)	0.07  ± 0.01 (1.3)	10.0  ± 0.2 (1.2)	240  ± 10 (1.6)
SwtLK213R	125  ± 6 (1.5)	2.0  ± 0.1 (3.6)	0.37  ± 0.01 (0.8)	0.7  ± 0.1 (1.0)	43  ± 2 (0.8)
SwtLK213A	129  ± 1 (1.4)	2.1  ± 0.1 (3.4)	0.32  ± 0.03 (1.0)	0.9  ± 0.1 (1.3)	28  ± 2 (1.5)
SwtLK213E	170  ± 8 (1.4)	1.9  ± 0.1 (3.9)	0.36  ± 0.02 (0.9)	0.7  ± 0.1 (1.2)	33  ± 1 (1.3)
SK198RLK213R	320  ± 20 (1.7)	3.5  ± 0.2 (1.8)	1.6  ± 0.1 (1.6)	1.3  ± 0.1 (1.2)	230  ± 10 (1.2)

Reactions were performed in either the synthesis (assay II) or the pyrophosphorolysis direction (assay I) as described in Methods. Data represent the average of two identical experiments ± the average difference of the duplicates. The values in parentheses are the Hill interaction coefficients (n_H). The 3PGA concentration used was 3 mm for the wild-type enzyme and the single-mutant enzymes, and 3PGA at 10 mm was used for the double-mutant enzyme.

Kinetic Characterization of S_wtL_K213 Mutant Enzymes

The apparent affinity for Glc-1-P was not affected when Lys-213 in the large subunit was replaced with Arg, Ala, or Glu (Table I). As shown in Figure 2B, there was no difference between the wild-type and S_wtL_K213R enzyme in terms of Glc-1-P dependence. This was also true for both the S_wtL_K213A and S_wtL_K213E enzymes (data not shown). This is in sharp contrast to the effect caused by mutations on Lys-198 in the small subunit (Fig. 2A). Although the apparent affinity of Glc-1-P was not affected, the apparent affinity for ADP-Glc decreased 2- to 4-fold (Table I). In general, mutations on Lys-213 of the large subunit caused small changes (less than 4-fold) to the kinetic constants for substrates (ATP and Mg²⁺), activator (3PGA), and inhibitor (Pi) (Table II).

Kinetic Characterization of S_K198RL_K213R Mutant Enzyme

When both Lys-198 in the small subunit and Lys-213 in the large subunit were replaced with Arg, the S_0.5 value for Glc-1-P was about 100-fold higher than that of the wild-type enzyme (Table I). Considering the 135-fold increase of the S_0.5 value in mutant enzyme S_K198RL_wt, the double mutation did not cause a further decrease in the apparent affinity of Glc-1-P over the single mutation. In either direction of the assay, the V_max of the double-mutant enzyme was essentially the same as that of the single-mutant enzyme, S_K198RL_wt.

The double mutation did not cause much alteration in the apparent affinities for Mg²⁺ and Pi (Table II); however, the A_0.5 value for 3PGA increased 11-fold relative to the wild type. This effect seemed to be additive, because the A_0.5 value for both S_K198RL_wt and S_wtL_K213R increased 3-fold. The 6-fold increase of the S_0.5 value for PPi was similar to the effect seen in the single-mutant enzyme, S_K198RL_wt. The other minor effects were 4-fold increases of the S_0.5 values for both ATP and ADP-Glc (Tables I and II).

Sugar-Phosphate Specificity

Because Lys-198 on the small subunit was implicated in Glc-1-P binding, it was of interest to determine whether the specificities for the substrates of the S_K198AL_wt mutant enzymes had been changed. To analyze systematically the contribution of each specific hydroxyl group in the binding of substrate to the active site of wild-type and mutant enzymes, a variety of compounds were used in which sugar moieties differ from Glc stereochemically or by substitution or elimination of hydroxyl groups at different positions. In the measurement of the activity of mutant enzyme S_K198AL_wt, large amounts of enzyme had to be used to obtain accurate measurements. The results are summarized in Table III. When 6-deoxy-Glc-1-P and 6F-Glc-1-P were used to substitute for Glc-1-P, the activity of the wild-type and S_K198AL_wt enzyme decreased 3.1- to 4.5-fold and 17- to 34-fold, respectively. When the other analogs were used to substitute for Glc-1-P, the activity of the wild-type and S_K198AL_wt enzymes decreased more than 14- and 150-fold, respectively. Thus, both enzymes showed the largest tolerance toward the elimination or substitution of the hydroxyl group at C-6 of the Glc molecule, followed by the changes at C-2, C-3, and C-4. In no case was there a substrate analog that showed a significantly enhanced reactivity with the mutant enzyme compared with the wild type. Substitution of a carboxyl group at C-6 indicates that GlcUA-1-P is not a substrate. The sugar-phosphate analogs also showed the same pattern of effect for the mutant enzyme S_K198RL_wt (data not shown). The results indicate that the hydroxyl groups at C-2, C-3, and C-4 probably played much more important roles than the C-6 hydroxyl group in the binding process. Overall, replacement of Lys-198 with either Ala or Arg did not cause any broadened changes in substrate specificity for sugar-1-phosphate.

Table III.

Specificity of sugar phosphates as substrates for wild-type and mutant enzyme S_K198AL_wt

Substrate	Substrate Concentration	Wild Type	SK198ALwt
	mm	units mg−1
Glc-1-P	2	48  ± 1	1.5  ± 0.2
	10	55.8  ± 4.7	10.4  ± 0.9
6-deoxy-Glc-1-P	2	13.6  ± 0.3	0.085  ± 0.001
	10	12.5  ± 0.1	0.62  ± 0.04
6F-Glc-1-P	2	13.0  ± 0.5	0.044  ± 0.002
	10	18.1  ± 0.6	0.47a
2F-Glc-1-P	2	1.0	0.007  ± 0.001
	10	4.0  ± 0.1	0.018  ± 0.03
Man-1-P	2	1.7  ± 0.1	0.010  ± 0.001
	10	4.1  ± 0.2	0.058  ± 0.003
3-deoxy-Glc-1-P	2	0.5  ± 0.1	0.010  ± 0.006
	10	1.3  ± 0.1	0.017  ± 0.002
3F-Glc-1-P	2	0.05  ± 0.01	≤0.002
	10	0.15  ± 0.01	≤0.002
Gal-1-P	2	0.16  ± 0.03	≤0.003
	10	0.28  ± 0.04	≤0.002
GlcUA-1-P	2	≤0.01	≤0.001
	10	≤0.01	≤0.02

Reactions were performed in the synthesis direction as described in “Materials and Methods,” with the presence of sugar phosphates as indicated. Data represent the average of two duplications ± sd. One unit of enzyme activity is expressed as the amount of enzyme required to form 1 μmol of ADPGlc per minute at 37°C assayed in either the synthesis or the pyrophosphorolysis direction. The lower limit of detection of enzyme activity for the wild type is 0.01 unit mg⁻¹, and for the mutant enzyme S_K198AL_wt it is 0.001 unit mg⁻¹ when a sufficient amount of enzyme was used in the assay, as indicated in the text.

^aSingle determination. 

Thermal Stability of Wild-Type and Mutant Proteins S_K198AL_wt, S_K198EL_wt, S_wtL_K213A, and S_K198RL_K213R

After heat treatment at 60°C for 5 min, the activity of the wild-type enzyme remained unchanged, whereas the S_K198AL_wt, S_K198EL_wt, S_wtL_K213A, and S_K198RL_K213R enzymes retained 104%, 67%, 94%, and 105% activity, respectively. Therefore, substitution of Lys-198 with a negatively charged Glu made the protein more susceptible to heat inactivation. Nevertheless, neither residue 198 in the small subunit nor residue 213 in the large subunit was critical for the stability of the native folded state of the potato tuber enzyme.

DISCUSSION

According to the results presented here, we can conclude that Lys-198 of the small subunit of potato tuber ADPGlc PPase is primarily involved in Glc-1-P binding. The 135- to 550-fold increases in the S_0.5 value for Glc-1-P when this residue was replaced by other amino acids explains the high conservation of this Lys in plant and bacterial ADPGlc PPases. The Lys residue probably is required for the proper substrate binding to ADPGlc PPase under physiological concentrations of Glc-1-P. Although Lys-198 is critical in interacting with Glc-1-P, it is obviously not essential for thermal stability. From the modest effect on V_max values and the kinetic constants for ATP, Mg²⁺, 3PGA, and Pi (Tables I and II), Lys-198 is probably neither involved in the rate-limiting step of the catalytic mechanism nor responsible for maintaining the native conformation of the enzyme. The n_H of Glc-1-P of the mutant enzymes was increased to 1.3 to 1.8, compared with 1.1 for the wild type. However, because both the S_K198AL_wt and S_K198EL_wt mutant enzymes had high S_0.5 values for Glc-1-P, it was impossible to perform kinetic studies for them under saturated concentrations of Glc-1-P. Only 50 to 70% V_max could be attained based on the estimation from the Lineweaver-Burk plot. Therefore, the n_H values could be overestimated for these two enzymes. Nevertheless, it has been observed for some enzymes that a single mutation causing decreased affinity for a ligand results in an increase in cooperativity (Stebbins and Kantrowitz, 1992; First and Fersht, 1993). The phenomenon was explained by the theory of preexisting cooperativity (First and Fersht, 1993).

As the substitution of Lys-198 varied from basic to neutral to acidic amino acids, the apparent affinities for Glc-1-P decreased. There seems to be a highly specific requirement for the presence of a Lys residue in terms of its charge, size, and shape in the active site to allow optimal binding of substrate. Even the most conservative substitution of an Arg resulted in a mutant enzyme with 135-fold lower apparent affinity for Glc-1-P, suggesting that charge alone is insufficient to account for proper interaction with the substrate. Arg, being a slightly larger amino acid than Lys, may sterically interfere with substrate binding.

In contrast to the effects observed for the mutations of Lys-198 of the small subunit, mutations of Lys-213 of the large subunit had no effect on the S_0.5 of Glc-1-P. When both residues were replaced by Arg, the effect on the apparent affinity for Glc-1-P was similar to that obtained with the single Arg substitution in the small subunit, ruling out a direct role of Lys-213 in binding of the substrate. As indicated in the introduction, the two mutated Lys residues and their surrounding sequences are highly conserved in the ADPGlc PPase family. A sequence search of the large subunit of tuber ADPGlc PPase revealed no consensus sequence other than the region surrounding Lys-213. Therefore, it is unlikely that Glc-1-P binds to an alternative site on the large subunit. This seems to be consistent with the proposed function of this subunit, i.e. modulating the allosteric regulation of the small subunit by 3PGA and Pi, with no direct role in catalysis. It is worth noting that this Lys residue is replaced by Gln in the large subunit of ADPGlc PPase from wheat endosperm (WE7) (Smith-White and Preiss, 1992), which may reflect the relative unimportance of this residue in the large subunit. In one small-subunit isozyme of ADPGlc PPase from Vicia faba L. seeds, VfAGPP, this Lys is replaced by Asn. In the other isozyme, VfAGPC, the Lys residue is retained.

Both small-subunit genes are expressed in identical temporal and spatial patterns (Weber et al., 1995); however, nothing is known about the comparative kinetics of the two enzymes. Recent binding experiments on the potato tuber enzyme by equilibrium dialysis (Y. Fu and J. Preiss, unpublished results) showed that ADPGlc bound to four sites per tetrameric enzyme. Unfortunately, experiments to determine the number of binding sites of Glc-1-P were unsuccessful because of the interference of ADP-Glc produced in the binding procedure; however, there is a possibility that Glc-1-P may bind to the large subunit, but with no catalysis after the binding event. In any case, the data provide further evidence that the main function of the small subunit is catalysis, as suggested by a previous study (Ballicora et al., 1995). At this stage, further studies are necessary to clarify the precise role of the large subunit.

The Lys-195 region (FVEKP) of E. coli ADPGlc PPase is not only conserved in potato tuber ADPGlc PPase (FAEKP), but is also identical to the Man-1-P-binding site of phospho-Man isomerase-GMP-d-Man pyrophosphorylase (May et al., 1994). Furthermore, this motif or a closely related sequence (GVEKP, IVEKY, KVIKP, or FKEKP) is found in many enzymes with the common characteristic of catalyzing the synthesis of a nucleoside diphosphate sugar from sugar phosphate and nucleoside triphosphate (Table IV). Therefore, FVEKP may be part of a sugar-phosphate-binding motif for this class of sugar nucleotide pyrophosphorylases. Of course, other sequences must dictate the sugar specificity, e.g. for Man-1-P or Glc-1-P.

Table IV.

Conservation of the sugar-phosphate-binding motif in sugar nucleotide PPases

Enzyme	Source	Sugar-Phosphate-Binding Motifa	Ref.
ADPGlc PPase	E. coli	IIEFVEKPAN	Hill et al. (1991); Preiss and Sivak (1996)
PMI/GMP	Pseudomonas aeruginosa	VQS*****DE	May et al. (1994)
PMI/GMP	Xanthomonas campestris	VER*****LA	Köplin et al. (1992)
GDPMan PPase	E. coli	RT*****NL	Marolda and Valvano (1993)
GDPMan PPase	Salmonella typhimurium	VAE*****DI	Jiang et al. (1991)
UDPGlc PPase	E. coli	PMVG****KA	Hossain et al. (1994)
UDPGlc PPase	Bacillus subtilis	VKN*****PK	Varnó et al. (1993)
UDPGlc PPase	Solanum tuberosum L.	TLKI***Y**	Katsube et al. (1990)
UDPGlc PPase	Dictyostelium discoideum	ETNK*I**YK	Ragheb and Dottin, 1987)
CDPGlc PPase	Yersinia pseudotuberculosis	VRS*K***KG	Thorson et al. (1994)

^aLys-195 in E. coli ADPGlc PPase and Lys-175 in P. aeruginosa PMI/GMP, which were shown to bind to Glc-1-P and Man-1-P, respectively, are underlined. Asterisks signify the same amino acid as in the E. coli ADPGlc PPase sequence. 

Various sugar-1-phosphate analogs with sugar moieties differing from Glc at each hydroxyl group were tested as substrates for wild-type and mutant enzyme S_K198AL_wt (Table III). No broadened specificities for the mutant enzyme were observed. This probably suggests that Lys-198 only participates in forming an ionic bond between its positively charged ε-amino group and the negatively charged phosphate group of Glc-1-P. Those hydroxyl groups may interact with the side chains of the other residues in the active site, i.e. by hydrogen bonding, to anchor the substrate correctly. Therefore, those analogs tested would have similar effects on the wild-type as well as the mutant enzymes.

Abbreviations:

ADPGlc PPase

ADP-Glc pyrophosphorylase

6F-Glc-1-P

6-fluoro-Glc-1-P

3-PGA

3-phosphoglycerate