Phosphodianion Activation of Enzymes for Catalysis of Central Metabolic Reactions.

The activation barriers ΔG⧧ for kcat/Km for the reactions of whole substrates catalyzed by 6-phosphogluconate dehydrogenase, glucose 6-phosphate dehydrogenase, and glucose 6-phosphate isomerase are reduced by 11-13 kcal/mol by interactions between the protein and the substrate phosphodianion. Between 4 and 6 kcal/mol of this dianion binding energy is expressed at the transition state for phosphite dianion activation of the respective enzyme-catalyzed reactions of truncated substrates d-xylonate or d-xylose. These and earlier results from studies on β-phosphoglucomutase, triosephosphate isomerase, and glycerol 3-phosphate dehydrogenase define a cluster of six enzymes that catalyze reactions in glycolysis or of glycolytic intermediates, and which utilize substrate dianion binding energy for enzyme activation. Dianion-driven conformational changes, which convert flexible open proteins to tight protein cages for the phosphorylated substrate, have been thoroughly documented for five of these six enzymes. The clustering of metabolic enzymes which couple phosphodianion-driven conformational changes to enzyme activation suggests that this catalytic motif has been widely propagated in the proteome.

A 12 kcal/mol intrinsic binding energy (IBE)[1] is observed for the substrate phosphodianion in reactions catalyzed by triosephosphate isomerase (TIM)[2−4] and glycerol 3-phosphate dehydrogenase (GPDH),[5,6] where <50% of the 12 kcal/mol dianion binding energy is expressed at the Michaelis complex and >50% is specifically expressed at the transition state for activation of the reaction of a phosphodianion-truncated substrate by a phosphite dianion.[6−8] β-Phosphoglucomutase (PGM) likewise utilizes the binding energy of the phosphite dianion to produce a 30 000-fold activation of the enzyme for transfer of a covalent phosphoryl reaction intermediate to the anomeric hydroxyl of β-d-xylopyranose.[9]

Glucose 6-phosphate (G6P) and 6-phosphogluconate (6PG) feature as substrates in the following reactions from Scheme : aldose–ketose isomerization catalyzed by glucose 6-phosphate isomerase (PGI),[10] hydride transfer catalyzed by glucose 6-phosphate dehydrogenase (G6PDH),[11] and oxidative decarboxylation catalyzed by 6-phosphogluconate dehydrogenase (6PGDH).[12−14] We report that (i) d-xylose and d-xylonate are poor substrates for catalysis by PGI and G6PDH or by 6PGDH, respectively; (ii) the phosphodianion of whole G6P or 6-phosphogluconate substrates provides ≥11 kcal/mol stabilization of the respective enzymatic transition states; and (iii) between 30 and 50% of this total dianion binding energy is recovered as HPO32– activation of the enzyme-catalyzed reactions of phosphodianion-truncated substrates d-xylose (PGI and G6PDH) or d-xylonate (6PGDH). The utilization of dianion binding interactions in catalysis by this tight cluster of six enzymes (Scheme ), which function at ancient metabolic pathways,[15,16] provides evidence that large phosphodianion and phosphite binding energies are intrinsic to a catalytic motif that appeared early in evolution and which was propagated to glycolytic enzymes and enzymes that pivot intermediates of glycolysis toward the production of pentose phosphates or lipids (Scheme ).

Scheme 1

Enzymes with Transition States Strongly Stabilized by Interactions to the Substrate Phosphodianion or Phosphite Dianion

Red: earlier work;[6,9] blue, this work.

The sources for the chemicals and enzymes used in these studies are reported in the Supporting Information (SI). Ec6PGDH from Escherichia coli, LmG6PDH from Leuconostoc mesenteroides, and ScPGI from Saccharomyces cerevisiae were shown to each give a single major band by sodium dodecyl sulfate polyacrylamide gel electrophoresis (Figure S1). The experimental protocols for the following enzyme assays are described in the SI: Ec6PGDH-, LmG6PDH-, and ScPGI-catalyzed reactions of physiological substrates; the oxidative decarboxylation of d-xylonate catalyzed by Ec6PGDH; and LmG6PDH-catalyzed oxidation of d-xylose. The ScPGI-catalyzed isomerization of d-xylose (Scheme ) was monitored by coupling formation of the product d-xylulose to oxidation of NADH catalyzed by sorbitol dehydrogenase.[17]

Scheme 2

Reactions of Whole [(kcat/Km)SPi] and Truncated [(kcat/Km)S] Substrates, and HPO32–-Activated Reactions of Truncated Substrates (kcat/KXylKHPi) Catalyzed by (A) Ec6PGDH, (B) LmG6PDH, and (C) ScPGI

Figure S2A–C shows Michaelis–Menten plots of v/[E] against [6PG], [G6P], and [F6P], respectively, for the reactions catalyzed by Ec6PGDH, LmG6PDH, and ScPGI. These plots give the values of (kcat/Km)SPi for reactions of whole substrates reported in Table , where (kcat/Km)SPi for the ScPGI-catalyzed reaction of G6P was calculated from kcat/Km = 2.38 × 106 M–1 s–1 for isomerization of F6P to form G6P and Keq = 3.45 for this isomerization reaction.[19]Figure S3A–C shows plots of v/[E] against [d-xylonate] and [d-xylose] for reactions catalyzed by Ec6PGDH (d-xylonate), LmG6PDH (d-xylose), and ScPGI (d-xylose). The slopes of the linear correlations from Figure S3A,B are equal to (kcat/Km)S for the reaction of the truncated substrates. Values of (kcat)S = 0.0047 s–1 and (kcat/Km)S = 3.6 × 10–4 M–1 s–1 (Table ) were determined from the fit of data from Figure S3C to the Michaelis–Menten equation. The contribution of dianion binding interactions to transition-state stabilization, (ΔG⧧)Pi = RT ln[(kcat/Km)SPi/(kcat/Km)S], are reported in Table .[1]

Table 1

Kinetic Parameters at pH 7.5 and 25 °C for Enzyme-Catalyzed Reactions of Whole and Phosphodianion Truncated Substrates (Scheme ), the Total Phosphodianion Binding Energies, and the Dianion Binding Energy Utilized for Enzyme Activation

	(kcat/Km)SPi	(kcat/Km)S		(ΔG⧧)Pi (kcal/mol),h,j	ΔGHPi⧧  (kcal/mol),i,j
enzyme	(M–1 s–1)e,f	(M–1 s–1)e,f	(M–2 s–1)f,g	[IBE]T	[IBE]HPi	[IBE]HPi/[IBE]T
ScPGIa	(6.9 ± 0.2) × 105	(3.6 ± 0.2) × 10–4	0.52 ± 0.10	12.6 ± 0.1	4.3 ± 0.1	0.34
	kcat = 400 s–1
LmG6PDHa	(2.0 ± 0.2) × 106	(8.3 ± 0.1) × 10–3	53.0 ± 0.3	11.4 ± 0.1	5.2 ± 0.1	0.46
	kcat = 320 s–1
Ec6PGDHb	(8.4 ± 0.4) × 105	(9.9 ± 0.2) × 10–3	140 ± 1	10.8 ± 0.1	5.7 ± 0.1	0.53
	kcat = 12 s–1
OMPDCc	1.1 × 107	0.026	12000	11.7 ± 0.1	7.7 ± 0.1	0.66
TIMd	2.2 × 108	0.062	2700	13.0 ± 0.1	6.3 ± 0.1	0.48
GPDHd	4.6 × 106	0.050	16000	10.8 ± 0.1	7.5 ± 0.1	0.69

SPi = G6P; S = d-xylose.

SPi = 6-PG; S = d-xylonate.

SPi = orotidine 5′-monophosphate (OMP); S = 1-(β-d-erythrofuranosyl)orotic acid.[6,18]

SPi = dihydroxyacetone phosphate; S = glycolaldehyde.[2,6]

Kinetic data for the catalyzed reactions of whole or truncated substrate (see SI).

The quoted uncertainty is the standard error obtained from the least-squares fit of experimental data to the appropriate kinetic equation.

Third-order rate constant for the phosphite dianion-activated reaction of truncated substrate (Scheme ).

(ΔG⧧)Pi = RT ln[(kcat/Km)SPi/(kcat/Km)S].

See eq .

The approximate uncertainties are calculated from the standard errors in the kinetic parameters.

Scheme 3

Activation of the Catalyzed Reactions of d-Xylonate (Ec6PGDH) or d-Xylose (LmG6PDH and ScPGI) by HPO32–

The cofactor NADP is bound to LmG6PDH and Ec6PGDH and is reduced to NADPH for the enzyme-catalyzed reactions of d-xylose or d-xylonate.

The Ec6PGDH-, LmG6PDH-, and ScPGI-catalyzed reactions of dianion-truncated substrates are strongly activated by phosphite dianion. Note that only Ec6PGDH-catalyzed oxidation of d-xylonate by NADP was monitored. We have not determined if this enzyme catalyzes the subsequent decarboxylation reaction. We first fit plots of v/[E] against [HPO32–] to the full rate equation for Scheme (not shown). However, the uncertainties in the kinetic parameters obtained from these fits (not shown) range from 25 to 100%, because the data do not clearly define the value for KXyl for weakly bound d-xylose or d-xylonate (Kxyl ≫ [Xyl], Scheme ). We report these data (Figure A–C) as plots of (v – v0)/[E] against [HPO32–] for Ec6PGDH-, LmG6PDH-, and ScPGI-catalyzed reactions at different fixed concentrations of d-xylonate (1A) or d-xylose (1B and 1C), where (v – v0) is the difference in the initial velocity for reactions in the presence and absence of HPO32–. The HPO32–-activated reactions of Ec6PGDH and LmG6PDH were carried out at saturating [NADP] = 1 mM. Data from Figure show that essentially the same enzyme activation is observed for reactions at 0.5 mM (open symbols) and 1.0 mM (solid symbols) [NADP]. The solid lines from Figure A–C show the fits of these kinetic data to eq , derived for Scheme , with the assumption that KXyl ≫ [Xyl] and using the derived values of (kcat/KHPi)obs (eq ).

Figure 1

Effect of increasing [HPO32–] on (v – v0)/[E] for (A) Ec6PGDH-catalyzed reactions of d-xylonate and (B and C) LmG6PDH- and ScPGI-catalyzed reactions of d-xylose, respectively. (A) Reactions at the following [d-xylonate]: 4.1, 10, 12.4, 16.5, 20, and 24 mM. (B and C) Reactions at the following [d-xylose]: 10, 20, 30, 40, and 50 mM. The solid and open symbols for panels A (10 mM d-xylonate) and B (50 mM d-xylose) show data for reactions at 1.0 and 0.5 mM NADP, respectively.

Figure A shows plots of (kcat/KHPi)obs, determined for the corresponding plots from Figure , against [Xyl] for the reactions catalyzed by Ec6PGDH and LmG6PDH, while Figure B shows a similar plot for reactions catalyzed by ScPGI. The slopes, (kcat)S•HPi/KXylKHPi (eq ), for these linear plots are reported in Table . The absence of detectable curvature for plots from Figure shows that KXylKHPi ≫ KXyl[HPO32–] (eq ), and that there is no significant accumulation of ternary [E·Xyl·HPO32–] complexes. By contrast, robust binding of phosphorylated substrates is observed (Km = 10–100 μM, SI) because of the entropic advantage to reactions of the whole substrates compared with the corresponding substrate pieces.[18,20,21]

Figure 2

Effect of increasing [d-xylonate] or [d-xylose] on (kcat/KHPi)obs for dianion activation of the catalyzed reactions of truncated substrates (Scheme ). (A) Ec6PGDH-catalyzed reactions of d-xylonate (●) and LmG6PDH-catalyzed oxidation of d-xylose by NADP (■). (B) ScPGI-catalyzed isomerization of d-xylose (●).

Table also reports (1) the total transition-state stabilization from binding interactions with phosphite dianion (intrinsic binding energy [IBE]HPi), calculated from eq derived for Scheme ; (2) the fraction ([IBE]HPi/[IBE]T) of the total intrinsic phosphodianion binding energy [IBE]T that is expressed at the transition state for the phosphite dianion-activated reaction of truncated substrate (Scheme ); and (3) the corresponding kinetic parameters determined for isomerization, decarboxylation, and hydride-transfer reactions catalyzed by TIM, orotidine 5′-monophosphate decarboxylase (OMPDC), and GPDH, respectively.[2,5,6,18]ScPGI catalyzes both ring-opening of cyclic sugar phosphates and subsequent isomerization of the acyclic sugar,[22] while Ec6PGDH catalyzes the coupled hydride-transfer and decarboxylation reactions of 6PG. We have not determined the transition-state stabilization for the individual enzymatic reaction steps from interactions with the substrate phosphodianion or phosphite dianion piece. We note earlier reports of phosphite dianion activation of both enzyme-catalyzed hydride transfer[5,23,24] and decarboxylation reaction[18] of phosphodianion-truncated substrates.

Scheme 4

Ground-State (KHPi) and Transition-State (KHPi⧧]) Binding of HPO32– to Ec6PGDH, LmG6PDH, and ScPGI

The data from Table and from an earlier study on catalysis by PGM[9] define a cluster of six metabolic enzymes (Scheme ) that show strong activation by HPO32– for the catalytic turnover of substrates, d-xylose, d-xylonate, or glycolaldehyde. Ec6PGDH, LmG6PDH, and ScPGI show smaller values of kcat/Km for the catalyzed reactions of both whole and truncated substrates, compared with the corresponding kinetic parameters for the TIM- and GPDH-catalyzed reactions of dihydroxyacetone phosphate and glycolaldehyde substrates. However, similar total intrinsic phosphodianion binding energies of 11–13 kcal/mol are observed for these five enzymes and for OMPDC (Table ). Apparently, 13 kcal/mol, which corresponds to a >1010-fold rate acceleration, represents an operational limit for transition-state stabilization from catalytic protein–dianion interactions. A larger fraction of the total dianion binding energy, ≤70% compared with ≤50%, is recovered in the activation of reactions of glycolaldehyde compared with d-xylonate or d-xylose by HPO32–. This corresponds to a larger specificity in the utilization of HPO32– binding energy for the ground state relative to the transition state in enzymes that catalyze the reactions of 6-carbon versus 3-carbon substrates.[7,25]

We have proposed a model for enzyme activation in which unliganded catalysts exist in flexible, open, but inactive conformations, and protein–dianion interactions are utilized to stabilize a fully active, rigid, and closed enzyme.[3,25] This model is supported by the results from experimental and computational studies on TIM,[20,26,27] OMPDC,[28,29] and GPDH.[30,31] It provides a rationalization for dianion activation of reactions catalyzed by 6PGDH[14] and PGI[32] because substrate binding to these enzymes gives rise to sizable phosphodianion-driven conformational changes. It may hold for catalysis by G6PDH; however, we are not aware of X-ray crystal structures for G6PDH which show the bound cofactor and substrate positioned to undergo hydride transfer.

Evolutionary pressure to optimize energy production from nutrients has driven TIM to perfection in catalysis of the isomerization of triosephosphate.[33] This catalytic perfection is presumably reflected in the structure for the iconic TIM barrel.[34] We propose that enzyme catalysis, with utilization of phosphodianion binding energy to drive an enzyme-activating conformational change, appeared early in protein evolution and that this powerful catalytic motif has been replicated in the evolution of metabolic pathways (Scheme ) and of enzymes that serve a host of cellular functions.