A biosensor study indicating that entropy, electrostatics, and receptor glycosylation drive the binding interaction between interleukin-7 and its receptor

Abstract

The interaction between interleukin-7 (IL-7) and its α-receptor, IL-7Rα, plays fundamental roles in the development, survival, and homeostasis of B- and T-cells. N-Linked glycosylation of human IL-7Rα enhances its binding affinity for human IL-7 300-fold versus that of the nonglycosylated receptor through an allosteric mechanism. The N-glycans of IL-7Rα do not participate directly in the binding interface with IL-7. This biophysical study involves dissection of the properties of binding of IL-7 to both nonglycosylated and glycosylated forms of the IL-7Rα extracellular domain (ECD) as functions of salt, pH, and temperature using surface plasmon resonance (SPR) spectroscopy. Interactions of IL-7 with both IL-7Rα variants display weaker binding affinities with increasing salt concentrations primarily reflected by changes in the first on rates of a two-step reaction pathway. The electrostatic parameter of the IL-7-IL-7Rα interaction is not driven by complementary charge interactions through residues at the binding interface or N-glycan composition of IL-7Rα, but presumably by favorable global charges of the two proteins. van't Hoff analysis indicates both IL-7-IL-7Rα interactions are driven by large favorable entropy changes and smaller unfavorable (nonglycosylated complex) and favorable (glycosylated complex) enthalpy changes. Eyring analysis of the IL-7-IL-7Rα interactions reveals different reaction pathways and barriers for the transition-state thermodynamics with the enthalpy and entropy changes of IL-7 binding to nonglycosylated and glycosylated IL-7Rα. There were no discernible heat capacity changes for the equilibrium or transition-state binding thermodynamics of the IL-7-IL-7Rα interactions. The results suggest that the unbound nonglycosylated IL-7Rα samples an extensive conformational landscape relative to the unbound glycosylated IL-7Rα, potentially explaining the switch from a "conformationally controlled" reaction (k(1) ∼ 10(2) M(-1) s(-1)) for the nonglycosylated interaction to a "diffusion-controlled" reaction (k(1) ∼ 10(6) M(-1) s(-1)) for the glycosylated interaction. Thus, a large favorable entropy change, a global favorable electrostatic component, and glycosylation of the receptor, albeit not at the interface, contribute significantly to the interaction between IL-7 and the IL-7Rα ECD.

Figure 1

Ribbon diagrams of the complex structures of IL-7 bound to nonglycosylated (green, 3di2.pdb) and glycosylated (magenta, 3di3.pdb) forms of the IL-7Rα (7). The nonglycosylated complex was superimposed onto the glycosylated complex. The 4 α-helical bundle of IL-7 is labeled helix A (HA), B (HB), C (HC), and D (HD) accordingly. The N-glycans attached to N29^7R, N45^7R, and N131^7R are colored as cyan CPK groups for the carbon atoms. The only histidine residues observed in the IL-7/IL-7Rα interface involve H78⁷ and H33^7R. The histidine side chains are drawn as white CPK groups for the carbon atoms. Oxygen and nitrogen atoms are colored red and blue, respectively. This picture was created and rendered using PyMOL (58).

Figure 2

Examples of the SPR binding kinetics of IL-7 to both nonglycosylated (EC, A and B) and glycosylated (CHO, C and D) forms of the IL-7Rα in 10 mM HEPES (pH 7.4), 150 mM NaCl, 3 mM EDTA, and 0.005% Tween-20 at 298 K. The black curves are trimmed and buffer subtracted binding sensorgrams. Two-fold serial dilutions of IL-7 were performed starting at 2.5 µM for the nonglycosylated IL-7/IL-7Rα (EC) interaction and 100 nM for the glycosylated IL-7/IL-7Rα (CHO) interaction. Also displayed is the global analysis of the sensorgrams analyzed to an one-step (A and C) or a two-step (B and D) binding reaction model using ClampXP (12) and depicted as red curves. The residuals of the global fitting analysis for each binding mechanism are plotted above the sensorgrams. Note that the residual y-axes scales for the two-step binding reactions models are smaller in A and C than the residual y-axes scales for the one-step binding reaction models in B and D.

Figure 3

Binding interactions of IL-7 to nonglycosylated and glycosylated IL-7Rα and IL-4 to IL-4Rα as a function of salt at pH 7.4 and 298 K. The IL-4/IL-4Rα data were taken from Sebald and coworkers (15). (A) The equilibrium binding constants (K_a) plotted against increasing NaCl concentrations for nonglycosylated IL-7/IL-7Rα (EC, filled circles), glycosylated IL-7/IL-7Rα (CHO, filled squares), and IL-4/IL-4Rα (filled triangles). The data displayed a linear correlation between K_a with increasing NaCl concentrations. (B) Natural logarithm of k₁ versus NaCl concentration expressed as 1/(1+κa) displaying an electrostatic enhancement for the IL-7/IL-7Rα (CHO) and IL-4/IL-4Rα interactions. The data were fit using linear regression with the slopes (−U/RT) relating to the electrostatic energy of interaction to the first association phase (k₁) and the intercept giving the basal on-rate, ${{\displaystyle k}}_1^0$. The electrostatic contributions (−U/RT) and ${{\displaystyle k}}_1^0$ rate constants are listed in Table 1.

Figure 4

Binding kinetic constants, k₁, k₋₁, k₂, and k₋₂, of IL-7 to nonglycosylated and glycosylated IL-7Rα as a function of pH at 150 mM NaCl and 298 K. (A) Bar graphs of k₁ constants versus pH. (B) Bar graphs of the k₋₁ constants versus pH. (C) Bar graphs of k₂ constants versus pH. (D) Bar graphs of the k₋₂ constants versus pH.

Figure 5

Equilibrium binding constants of IL-7 to nonglycosylated and glycosylated IL-7Rα and IL-4 to IL-4Rα as a function of temperature at 150 mM NaCl and pH 7.4. (A) van't Hoff analysis plotted as lnK_d versus 1/T (K⁻¹). Data fit well using linear regression for nonglycosylated IL-7/IL-7Rα (EC, filled circles), glycosylated IL-7/IL-7Rα (CHO, filled squares), and IL-4/IL-4Rα (filled triangles). (B) Bar graphs of the equilibrium thermodynamic parameters obtained from van't Hoff analysis for these interactions. These parameters are listed in Table 2. The ΔG^o values are reported at 298 K.

Figure 6

Eyring analysis of the kinetic rate constants of IL-7 to nonglycosylated (filled circles) and glycosylated (filled squared) IL-7Rα as a function of temperature at 150 mM NaCl and pH 7.4. The data were plotted as ln(k_xh/k_BT) versus 1/T (K⁻¹) where k_x are the individual rate constants (k₁ (A), k₋₁ (B), k₂ (C), and k₋₂ (D)), h is Plank’s constant, and k_B is Boltzman’s constant (59). Data fit well using linear regression. The slopes and intercepts yield transition-state enthalpies (ΔH^o‡/R) and entropies (ΔS^o‡/R) for the individual association and disassociation pathways and are listed in Table 3.

Figure 7

Open book view of the electrostatic surfaces and potentials for the IL-7/IL-7Rα (A) and IL-4/IL-4Rα (B) complexes. The linear Poisson-Boltzmann equation was solved for both protein complexes using APBS (33) with 150 mM monovalent salt at 298 K. The solvent accessible surface areas for each protein are displayed and colored blue (+5 kT/e) and red (−5 kT/e). The electrostatic potential gradients for each protein are displayed as a blue (+2 kT/e) and red (−2 kT/e) mesh. The structures and electrostatic potentials were displayed using PyMOL (58). The binding interfaces for both complexes are highlighted with dashed circles. The second view of IL-7 below the view with the dashed circle is rotated 90° in the vertical direction to illustrate the large positive surface on the top of the cytokine.

Figure 8

Binding reaction pathways determined from Eyring analysis for IL-7/IL-7Rα interactions for ΔG^o (A), ΔH^o (B), and −TΔS^o (C). The blue and red lines are for the nonglycosylated and glycosylated pathways, respectively. The transition-state thermodynamic values for the interactions are listed in Table 3.

Figure 9

Circular dichroism spectra of unbound forms of IL-7Rαs (EC and S2) and the human growth hormone receptor (hGHR) ECDs. The spectra display the far-UV regions of the nonglycosylated IL-7Rα (EC, blue), glycosylated IL-7Rα (S2, red), and hGHR (green).