Molecular architecture of E. coli purine nucleoside phosphorylase studied by analytical ultracentrifugation and CD spectroscopy

Abstract

Purine nucleoside phosphorylase (PNP) is a key enzyme of the nucleoside salvage pathway and is characterized by complex kinetics. It was suggested that this is due to coexistence of various oligomeric forms that differ in specific activity. In this work, the molecular architecture of Escherichia coli PNP in solution was studied by analytical ultracentrifugation and CD spectroscopy. Sedimentation equilibrium analysis revealed a homohexameric molecule with molecular mass 150+/-10 kDa, regardless of the conditions investigated-protein concentration, 0.18-1.7 mg/mL; presence of up to 10 mM phosphate and up to 100 mM KCl; temperature, 4-20 degrees C. The parameters obtained from the self-associating model also describe the hexameric form. Sedimentation velocity experiments conducted for broad protein concentration range (1 microg/mL-1.3 mg/mL) with boundary (classical) and band (active enzyme) approaches gave s(0)20,w=7.7+/-0.3 and 8.3+/-0.4 S, respectively. The molecular mass of the sedimenting particle (146+/-30 kDa), calculated using the Svedberg equation, corresponds to the mass of the hexamer. Relative values of the CD signal at 220 nm and the catalytic activity of PNP as a function of GdnHCl concentration were found to be correlated. The transition from the native state to the random coil is a single-step process. The sedimentation coefficient determined at 1 M GdnHCl (at which the enzyme is still fully active) is 7.7 S, showing that also under these conditions the hexamer is the only catalytically active form. Hence, in solution similar to the crystal, E. coli PNP is a hexameric molecule and previous suggestions for coexistence of two oligomeric forms are incorrect.

Figure 1.

Sedimentation coefficient distributions (g(s⁰_20,w)) obtained for E. coli PNP. Sedimentation velocity experiments were carried out in (A) 50 mM Tris-HCl (pH 7.0) for 1.3 mg/mL (– –) and 0.03 mg/mL (–) PNP and in 50 mM phosphate buffer (pH 7.0) with 1M GdnHCl for 0.6 mg/mL (- - -) and 0.03 mg/mL (○) PNP and (B) for 1.3 mg/mL PNP in 50 mM Tris-HCl (pH 7.0) without (– –) and with (–) 0.5 M NaCl, with 50 mM sodium phosphate (○), or with 0.5 M NaCl and 50 mM sodium phosphate (- - -). Centrifugation was at 60,000 rpm and 20°C.

Figure 2.

(A) CD spectra obtained for 1.75 mg/mL PNP in 50 mM phosphate buffer (pH 7.0) without (–) and with 1M (- - -), 2M (▴), and 4M (○) guanidine hydrochloride (GdnHCl) and for 4 M GdnHCl without PNP (*). (B) Influence of a denaturant (GdnHCl) on enzymatic activity and secondary structure of E. coli PNP. Relative values of CD signal at 220 nm (□) and enzymatic activity of PNP with 0.5 mM m⁷Guo as a nucleoside substrate (•) are presented as a function of GdnHCl concentration.

Figure 3.

Active enzyme centrifugation experiment for 20 μg/mL PNP and the substrate MESG. Centrifugation was performed at 40,000 rpm and 20°C. Absorbance distribution, corresponding to the concentration of the product of phosphorolysis of MESG, observed while the enzyme layer moves along the cuvette filled with the reaction mixture containing both substrates (300 μM MESG and 50 mM phosphate) was measured at 360 nm. Radial absorption profiles were obtained in 5-min intervals. Inset shows the plot of ln(r) vs. time (r is the position of the midpoint of the sedimenting boundary) (▪) and the linear fit of these data (—).

Figure 4.

Active enzyme centrifugation experiment for 20 μg/mL PNP and substrate m⁷Guo. Centrifugation was performed at 40,000 rpm and 25°C. Absorbance distribution, corresponding to the concentration of the product of phosphorolysis of m⁷Guo, observed while the enzyme layer moves along the cuvette filled with the reaction mixture containing both substrates (300 μM m⁷Guo and 50 mM phosphate) was measured at 260 nm. Radial absorption profiles were obtained in 7.5-min intervals. The inset shows plot of ln(r) vs. time (r is the position of the midpoint of the sedimenting boundary) (▪) and the linear fit of these data (—).

Figure 5.

Band centrifugation of E. coli PNP (3.4 mg/mL). Centrifugation was performed at 40,000 rpm and 20°C. Protein was loaded on 50 mM HEPES buffer (pH 7.0) with 150 mM KCl. Absorbance distribution of protein observed while enzyme layer moves along the cuvette was measured at 280 nm. Radial absorption profiles were obtained in 15-min intervals. The inset shows plot of ln(r) vs. time (r is the maximum of a spectrum) (▪) and the fit of these data (—).

Figure 6.

Sedimentation equilibrium experiment of E. coli PNP in 50 mM Tris-HCl buffer (pH 7.5) with 100 mM KCl and 0.1 mM DTT at 20°C. Centrifugation was performed at 10,200 rpm. The enzyme was loaded at 0.46 mg/mL (A), 1.08 mg/mL (B), and 1.7 mg/mL (C). Protein concentration gradient in equilibrium measured at 280 nm (shown as open circles at bottom) is shown together with the single-species model fit (line). (Top) Residuals (A_exp–A_mod) for this model (fitted simultaneously to all three data sets). Similar curves and residual plots were obtained for the self-association model (i.e., coexistence of monomers and N-mer oligomer). The following parameters for these two models were obtained: single-species model, M = 150,746 Da; self-association model, M_monomer = 25,988 Da, N = 5.8, K_a = 4.0 × 10⁸⁷ M⁻⁵.