Purification and Characterization of glpX-Encoded Fructose 1,6-Bisphosphatase, a New Enzyme of the Glycerol 3-Phosphate Regulon of Escherichia coli

Abstract

In Escherichia coli, gene products of the glp regulon mediate utilization of glycerol and sn-glycerol 3-phosphate. The glpFKX operon encodes glycerol diffusion facilitator, glycerol kinase, and as shown here, a fructose 1,6-bisphosphatase that is distinct from the previously described fbp-encoded enzyme. The purified enzyme was dimeric, dependent on Mn²⁺ for activity, and exhibited an apparent K_m of 35 μM for fructose 1,6-bisphosphate. The enzyme was inhibited by ADP and phosphate and activated by phosphoenolpyruvate.

Growth of Escherichia coli on glycerol or sn-glycerol 3-phosphate (glycerol-P) as the sole carbon source is mediated primarily by the glp regulon (15). The glpFKX operon, one of the five operons of the regulon, encodes glycerol facilitator (glpF), glycerol kinase (glpK), and a protein of unknown function (glpX) (31, 32). It was initially reported that GlpX displays limited sequence similarity to the Synechococcus leopoliensis fructose 1,6-bisphosphatase (FBPase) (31). In our work, a more recent BLAST search revealed that GlpX manifests 39% identity to an FBPase of Synechococcus sp. strain PCC7492 (29). Until now, the only recognized E. coli FBPase was encoded by fbp (25). This FBPase (FBPase I) has only 10% identity to the amino acid sequence of glpX-encoded FBPase (FBPase II). E. coli FBPase I is dependent on Mg²⁺, is inhibited by low levels of AMP, is tetrameric (1), and is necessary for growth of E. coli on gluconeogenic substrates such as glycerol or succinate (10).

It is not clear why E. coli would maintain two distinct FBPases. FBPases can modulate the concentration of fructose 1,6-bisphosphate [Fru(1,6)P₂] and fructose 6-phosphate. These two regulatory hexoses affect glycolysis enzymes 6-phosphofructokinases I and II, pyruvate kinase I, and phosphoenolpyruvate (PEP) carboxylase (2, 4, 13, 20); glycogen synthesis enzyme ADP-glucose pyrophosphorylase (12); and carbon-source import pathway enzymes glycerol kinase and 1-phosphofructokinase (6, 15). Flux through the Embden-Meyerhof pathway in the direction of glycolysis or gluconeogenesis can be allosterically controlled at the enzyme level by other metabolites as well: PEP, ATP, ADP or AMP (9). The potential “futile cycle” of phosphofructokinases and FBPases is also alleviated by this regulation. Therefore, regulation of FBPases is important.

In this communication, the FBPase activity of the glpX gene product is documented. The glpX-encoded enzyme, FBPase II, was purified and characterized, enabling comparison of the attributes of E. coli FBPases in vitro. Further, a chromosomal insertion mutation in glpX was constructed to test the physiological effects of the glpX mutation on carbohydrate metabolism.

E. coli strains and cloning of glpX.

E. coli strains used in this study are listed in Table 1. Strains were grown in Luria broth (LB) supplemented with antibiotics as needed or in minimal medium (7) containing 0.4% glycerol or 0.2% glucose.

TABLE 1.

E. coli strains used in this study

Strain	Relevant genotype	Description and/or reference
MG1655	Wild-type isolate (F−rph-1 λ−)	3
TL524	MG1655 Δ(lacZYA–argF)U169	TcS derivative of TL504 (22)
CAG12185	argE86::Tn10	26
CAG12019	F−zjg-920::Tn10 rph-1 λ−	26
AM1	Hfr pfkA2 fhuA22 ompF627 (T2r) fadL701 (T2r) relA1 pit-10 spoT1 rrnB2 mcrB1 creC510	21
R6	F−edd-1 gnd-1 gnt pps-1 ΔargH1 iclR15(Con)	23
DF657	HfrC Δfbp287 fhuA22 spoT1 pit-10 relA1 λ−ompF627(T2r)	25
JB100	DF657 Δfbp287a zjg-920::Tn10	P1(CAG12019) into DF657
JB108	BL21(DE3) zjg920::Tn10 Δfbp287	P1(JB100) into BL21(DE3) (27)
JLD1001b	TL524 glpX::Spcr	This work
JLD1101	TL524 glpX::SpcrargE86::Tn10	P1(CAG12185) into JLD1001, select Tcr, score Spr
JLD1301	TL524 glpX::SpcrargE86::Tn10	P1(JLD1101) into TL524
JLD2401	TL524 glpX::Spcr	P1(JLD1301) into TL524, select Spr, score Tcr
JLD2402	TL524 glpX::Spcr Δfbp287 zjg-920::Tn10	P1(JB100) into JLD2401
JLD2403	TL524 glpX::Spcrfbp+zjg-920::Tn10	P1(JB100) into JLD2401
JLD2404	TL524 glpX+ Δfbp287 zjg-920::Tn10	P1(JB100) into TL524
JLD2405	TL524 glpX+fbp+zjg-920::Tn10	P1(JB100) into TL524

^aThe Δfbp287 deletion affects not only fbp but also the flanking genes yjfF, yjfG, ytfT, and ytfS. 

^bConstruction of the glpX insertion mutation: strain JLD1001 was formed by insertion of the Spc^r gene into the BsiWI site of glpX, using a plasmid or chromosomal integration selection system (16). pBS-glpX was constructed from SacI/BamHI-digested pBluescript KS(+) (Stratagene) ligated to the SacI/BamHI ∼3,400-bp fragment of pDW23 (32) encompassing the glpK, glpX, and fpr genes. Plasmid pBS-glpX::Spc^r was constructed by ligation of SmaI-excised Spc^r gene, from pHP45Ω (8), into the filled-in (now blunt) BsiWI site of glpX in pBS-glpX. pKO3-glpX::Spc^r was formed by ligation of the SacI (blunted)/SalI piece of pBS-glpX::Spc^r with SmaI/SalI-digested pKO3 (16). Plasmid pKO3-glpX::Spc^r contains genes glpK, glpX::Spc^r, and fpr. From pKO3-glpX::Spc^r, glpX::Spc^r was integrated into strain TL524 to form strain JLD1001. The integration site was verified by P1 transduction of JLD1001 to Tc^r, with strain CAG12185 as the P1 donor, to form strain JLD1101. Comparison of PCR and Southern analyses of DNA of strains JLD2402 and JLD2403, containing the insertional mutation, with JLD2404 and JLD2405 (glpX⁺), also confirmed integration of the Spc^r gene in the glpX gene (data not shown). 

The glpX gene was PCR amplified from chromosomal DNA of strain MG1655 using the primer pair acgtgaaTTCCCCTGTGCTACACTTCG (hybridizes to a region 48 bp upstream of the initiation codon) and acgttctagaTTGCCTGTTACCCAATCAGC (hybridizes to a region 102 bp downstream of the termination codon), containing sequence mismatches (lowercase) and restriction sites EcoRI and XbaI (underlined). The PCR product was ligated into XbaI/EcoRI-cut pZE14 (18) forming pJB100. This plasmid was introduced into strain DF657 (Δfbp) to test for complementation of the Fbp⁻ (glycerol-negative) phenotype. DF657(pJB100) was capable of slow growth on glycerol. Therefore, increased expression of glpX from the high-copy-number plasmid pJB100 complements the Fbp⁻ phenotype.

Overexpression and purification of FBPase II.

To facilitate overexpression and purification of FBPase II, a pT7-7 (28) derivative containing glpX was constructed (pJB300B). The NdeI-SalI fragment that was inserted into the same sites of pT7-7 to form pJB300B was obtained by PCR amplification using primers acgtaccaattgaggagatatacaTATGAGACGAGAACTTGCCATC and acgtgtcgacTGCCTTATCTTCGTTCTCCG, with DNA from strain MG1655 as the template (sequence mismatches are lowercase and restriction sites are underlined). The expected nucleotide sequence of glpX in pJB300B was confirmed. Overexpression of glpX was induced in JB108(pJB300B) with 200 μM isopropyl-β-d-thiogalactopyranoside (IPTG) for 2 h in a 200-ml LB culture with an optical density at 600 nm of 0.6. Cells were harvested by centrifugation, resuspended in 2.5 ml of 20 mM Tricine (pH 7.7)–50 mM KCl–1 mM MgCl₂–1 mM dithiothreitol (DTT)–0.5 mM EDTA, lysed by sonication, and centrifuged at 140,000 × g for 30 min. Nucleic acids were precipitated by addition of polyethylenimine (0.05%, vol/vol) and centrifugation at 10,000 × g for 20 min. FBPase II activity was precipitated by addition of 0.4 g of (NH₄)₂SO₄ ml⁻¹. The (NH₄)₂SO₄ pellet was dissolved in 2.3 ml of 50 mM Tris-HCl (pH 7.7)–0.2 mM MgCl₂–0.1 mM EDTA and fractionated by anion exchange chromatography on a Q HR15 column (Waters) at room temperature using a gradient of 0.075 to 0.5 M NaCl in 50 mM Tris-HCl (pH 7.7)–0.2 mM MgCl₂–0.1 mM EDTA. Peak fractions were pooled and precipitated with 0.5 g of (NH₄)₂SO₄ ml⁻¹, redissolved, and then dialyzed against 20 mM Tricine (pH 7.7)–1 mM MgCl₂–0.1 mM DTT–15% glycerol. All steps of the purification procedure were monitored by coupled spectrophotometric assay. A total of 2.8 mg of protein was purified 4.4-fold, with a 55% recovery and specific activity of 4.2 U mg⁻¹. Analysis of purified product by sodium dodecyl sulfate-polyacrylamide gel electrophoresis showed that the enzyme was >95% pure. Protein concentrations were determined as described by Bradford (5) with bovine serum albumin as the standard.

Optimal conditions for enzyme activity.

Using a coupled spectrophotometric assay (Fig. 1), the highest activity was obtained with 8 μg of FBPase II in a 1-ml reaction mixture containing 0.5 to 2 mM MnCl₂, 50 to 125 mM KCl, 0.02 M Tricine (pH 7.7), and 1.5 mM Fru(1,6)P₂. Omission of Mn²⁺ resulted in >90% loss of activity. Replacing Mn²⁺ with 1 mM ZnCl₂, CaCl₂, FeSO₄, or CuSO₄ resulted in almost complete loss of activity. Replacement with 3 to 10 mM MgCl₂ resulted in <10% of the activity present with Mn²⁺. The presence of 1 mM DTT had no effect on FBPase II activity.

FIG. 1.

(A) Dependence of FBPase II activity on concentration of Fru(1,6)P₂. Each data point is presented as an average of two to four determinations, with standard errors (error bars). Coupled spectrophotometric assay was used with a solution containing 8 μg of FBPase II, 0.2 U of phosphoglucose isomerase, 0.4 units of glucose-6-phosphate dehydrogenase, 20 mM Tricine (pH 7.7), 50 mM KCl, 1 mM MnCl₂, and 0.25 mM NADP in a final volume of 1 ml at room temperature (24). The rate of reduction of NADP to NADPH was determined by monitoring absorbance at 340 nm. (B) Lineweaver-Burk replot of selected data points from panel A.

Molecular mass determination.

The molecular mass of the GlpX subunit estimated by using sodium dodecyl sulfate-polyacrylamide gel electrophoresis was 40 kDa, which is near the expected size based on the deduced amino acid sequence (36 kDa). Gel filtration chromatography was performed using a Waters Protein Pak Glass 300SW column in 20 mM Tricine (pH 7.7)–100 mM KCl with 1 mM MnCl₂ or 1 mM MgCl₂. Standards included ribonuclease A (17 kDa), carbonic anhydrase (29 kDa), ovalbumin (45 and 90 kDa), bovine serum albumin (66 and 132 kDa) and phosphorylase b (97.4 kDa). Times of protein elution were determined by monitoring the absorbance at 280 nm. At high, medium, and low concentrations of FBPase II (320, 160, and 40 μg of FBPase II, respectively, in 200 μl of loading volume), the protein eluted at 90, 80, and 47 kDa, respectively. FBPase II is most likely a dimer at higher concentrations, in contrast to the tetrameric fbp-encoded FBPase I (1).

Catalytic properties.

A K_m of ∼35 μM and V_max of ∼3.3 U mg⁻¹ were determined for FBPase II using Fru(1,6)P₂ levels below 0.25 mM (Fig. 1). Substrate inhibition was observed at high Fru(1,6)P₂ concentrations. FBPase II has lower affinity for Fru(1,6)P₂ compared to FBPase I (apparent K_m of 5 μM [1]). Based on specific activities of the purified enzymes, the deduced turnover number for FBPase II is about seven times less than that of FBPase I (19).

Substrate specificity of FBPase II.

Substrate specificity was determined by measuring the rate of enzyme-catalyzed production of inorganic phosphate from various potential substrates (see footnote d to Table 2). Fru(1,6)P₂ was the best substrate found, while fructose 1-phosphate and ribulose 1,5-bisphosphate served less well as substrates. The apparent K_m and V_max of FBPase II with fructose 1-phosphate were 1 mM and 1.4 U mg⁻¹, respectively, giving a V_max/K_m almost 70 times lower than that obtained using Fru(1,6)P₂. Ribulose 1,5-bisphosphate at 1 mM produced 15% of the phosphatase activity obtained with 1.5 mM Fru(1,6)P₂. Glucose 6-phosphate, fructose 6-phosphate, mannose 6-phosphate, glucose 1,6-bisphosphate, sedoheptulose 1,7-bisphosphate, sorbitol 6-phosphate, and glycerol-P were not significant substrates (all present at 1 mM [data not shown]).

TABLE 2.

Effectors of FBPase II with Fru(1,6)P₂ substrate

Effector (concn [mM])	Reaction rate (%)
None	100a
KH2PO4 (0.5)	68a
Mn-AMP (1)	100a
Mn-ADP (1.25)	50b
Mn-ADP (1.9)	38b
Mn-ATP (1.25)	88b
Fructose 1-phosphate (2)	63c
PEP (1)	170d

^aEnzyme activity was determined by coupled spectrophotometric assay containing 8 μg of FBPase, 0.05 mM Fru(1,6)P₂, and effector. Reaction rates were compared to the rate with no effector added, which was 2.3 U mg⁻¹ for this preparation. 

^bEnzyme activity was determined by coupled spectrophotometric assay containing 8 μg of FBPase, 0.05 mM Fru(1,6)P₂, effector, and Mn²⁺ in 1 mM excess of effector concentration. Reaction rates were compared to the rate with no effector added, which was 1.5 U mg⁻¹ for this preparation. 

^cEnzyme activity was determined as explained in footnote b but with 1 mM Mn²⁺. 

^dEnzyme activity was determined by inorganic phosphate quantitation assay, containing 2 μg of FBPase, 0.6 mM Fru(1,6)P₂, and PEP. Reaction rates were compared to the rate with no effector added, which was 2.3 U mg⁻¹. Reaction mixtures (0.5 ml) containing 20 mM Tricine (pH 7.7), 50 mM KCl, 1 mM MnCl₂, Fru(1,6)P₂, and enzyme were incubated at room temperature. Aliquots were removed for phosphate determination using the malachite green-ammonium molybdate reagent (14). One unit of enzyme activity is defined as the amount of enzyme catalyzing the formation of 1 μmol of product in 1 min. 

Effectors of FBPase II.

Fructose 1-phosphate, inorganic phosphate, ADP, and ATP inhibited FBPase II activity (Table 2). Fructose 1-phosphate and inorganic phosphate inhibited activity competitively, with apparent K_is of 1 and 0.35 mM, respectively. Fructose, fructose 6-phosphate, and glucose 6-phosphate were not inhibitors (data not shown). AMP had no effect on enzyme activity, but ATP and especially ADP inhibited activity at low Fru(1,6)P₂ (≤0.1 mM). PEP almost doubled enzyme activity (Table 2), while addition of dihydroxyacetone phosphate, glycerol, or glycerol-P (all at 1 mM) did not affect activity (not shown). The response of FBPase II to effector molecules PEP, ADP, and ATP associates it with many enzymes of the Embden-Meyerhof pathway that are regulated by PEP, Fru(1,6)P₂, ATP, ADP, or AMP (9). The properties of FBPase II are distinct from those of FBPase I, an enzyme exquisitely sensitive to inhibition by AMP (50% inhibition by 15 μM AMP [1]). AMP inhibition of FBPase I is alleviated by PEP, although higher concentrations of PEP are inhibitory (>1 mM PEP).

Physiological role of glpX.

Genetic studies were undertaken in an attempt to discern the physiological role of glpX. First, the roles of the two FBPases were investigated using congenic strains with the four possible combinations of wild-type and defective fbp and glpX alleles: JLD2402 (Δfbp glpX::Spc^r), JLD2403 (fbp⁺ glpX::Spc^r), JLD2404 (Δfbp glpX⁺), and JLD2405 (fbp⁺ glpX⁺) (Table 1). The fbp⁺ glpX::Spc^r strain grew as well as the wild type on LB or glucose, fructose, succinate, or glycerol minimal medium, aerobically or anaerobically. Apparently, FBPase II is not crucial to cell growth under these conditions when FBPase I is present. Strains lacking FBPase I (e.g., Δfbp glpX⁺ strains DF657 and JLD2404) grew normally on glucose or fructose medium but were unable to grow on minimal medium supplemented with glycerol or other gluconeogenic substrates, indicating that chromosomal glpX⁺ does not compensate for the loss of fbp expression. (However, increased expression of glpX from multicopy plasmid pJB100 complemented the Fbp⁻ phenotype.)

The Δfbp glpX⁺ strain (JLD2404) was able to revert to slow growth on glycerol minimal medium enriched with 0.03% Casamino acids and 2.5 mM KNO₃ (with or without O₂). These revertants lost the ability to grow on glycerol when transduced to glpX::Spc^r. The Δfbp glpX::Spc^r strain JLD2402 did not show any reversion under the same conditions. A possible explanation for reversion of Δfbp glpX⁺ strains to a glycerol-positive phenotype is a mutation yielding elevated glpX expression.

The effect of the glpX and fbp mutations on glycogen accumulation was tested by exposure of the four congenic strains to iodine vapor following growth on Kornberg medium containing 0.4% glycerol instead of glucose (17). Glycogen accumulated to higher levels in both fbp⁺ strains (maroon colony color) compared to levels found in the Δfbp mutants (orange colony color). Therefore, accumulation of glycogen to wild-type levels requires FBPase I but is not affected by the glpX mutation.

We hypothesized that the inability of Δfbp glpX⁺ strains to grow on gluconeogenic substrates may be due to low-level expression of glpX combined with competing hexose catabolic pathways that would prevent accumulation of hexose sufficient for cell growth. To test this possibility, the Δfbp allele was introduced by cotransduction with zjg920::Tn10 into a pfkA2 strain deficient in glycolysis (phosphofructokinase I; strain AM1) and into an edd gnd strain deficient in the Entner-Douderoff pathway (6-phosphogluconate dehydrase and 6-phosphogluconate dehydrogenase; strain R6). In both cases, transductants unable to grow on glycerol were obtained at the expected frequency, demonstrating that glpX function is insufficient to provide hexose for growth of these strains where presumed competing pathways are impaired. We did observe that a pfkA glpX::Spc^r strain grew more slowly with glucose or glycerol than did the pfkA glpX⁺ parent strain AM1, indicating a functional importance of FBPase II in this strain.

Three classes of FBPases.

Bacterial FBPases are members of three different orthologous groups (30), including a group of glpX-encoded FBPases (class II). Some bacteria that contain orthologs of FBPase II include Klebsiella aerogenes (91% identity), Yersinia pestis (81%), Haemophilus influenzae (67%), Bacillus subtilis (49%), Clostridium acetobutylicum (45%), Mycobacterium tuberculosis (43%) and Synechococcus sp. strain PCC7942 (39% [29]). There is little similarity of these class II FBPases to the other two orthologous groups of FBPases, those of FBPase I (E. coli) (class I), and those of a very divergent FBPase of B. subtilis (class III [11]). From the sequence data available, no bacterial genome has a combination of class I and class III FBPases, although combinations of classes I and II and of classes II and III are seen.

In conclusion, glpX-encoded FBPase II is still somewhat of an unknown entity in the overall framework of E. coli metabolism. FBPase II is conserved in many other bacteria and is the only known FBPase in some organisms (e.g., M. tuberculosis). The enzyme is distinct in many aspects from FBPase I. However, the specificity and high affinity of FBPase II for substrate and its regulation by PEP and ADP suggest that the enzyme functions with FBPase I in the central pathways of carbohydrate metabolism.