Phosphorylation and ionic strength alter the LRAP-HAP interface in the N-terminus

Abstract

The conditions present during enamel crystallite development change dramatically as a function of time, including the pH, protein concentration, surface type, and ionic strength. In this work, we investigate the role that two of these changing conditions, pH and ionic strength, have in modulating the interaction of the amelogenin, LRAP, with hydroxyapatite (HAP). Using solid-state NMR dipolar recoupling and chemical shift data, we investigate the structure, orientation, and dynamics of three regions in the N-terminus of the protein: L(15) to V(19), V(19) to L(23), and K(24) to S(28). These regions are also near the only phosphorylated residue in the protein pS(16); therefore, changes in the LRAP-HAP interaction as a function of phosphorylation (LRAP(-P) vs LRAP(+P)) were also investigated. All of the regions and conditions studied for the surface immobilized proteins showed restricted motion, with indications of slightly more mobility under all conditions for L(15)(+P) and K(24)(-P). The structure and orientation of the LRAP-HAP interaction in the N-terminus of the phosphorylated protein is very stable to changing solution conditions. From REDOR dipolar recoupling data, the structure and orientation in the region L(15)V(19)(+P) did not change significantly as a function of pH or ionic strength. The structure and orientation of the region V(19)L(23)(+P) were also stable to changes in pH, with the only significant change observed at high ionic strength, where the region becomes extended, suggesting this may be an important region in regulating mineral development. Chemical shift studies also suggest minimal changes in all three regions studied for both LRAP(-P) and LRAP(+P) as a function of pH or ionic strength, and also reveal that K(24) has multiple resolvable resonances, suggestive of two coexisting structures. Phosphorylation also alters the LRAP-HAP interface. All of the three residues investigated (L(15), V(19), and K(24)) are closer to the surface in LRAP(+P), but only K(24)S(28) changes structure as a result of phosphorylation, from a random coil to a largely helical structure, and V(19)L(23) becomes more extended at high ionic strength when phosphorylated. These observations suggest that ionic strength and dephosphorylation may provide switching mechanisms to trigger a change in the function of the N-terminus during enamel development.

Figure 1

Changes in chemical shift as a function of preparation condition for the six samples vs. the standard condition, for the ¹³C' labeled L¹⁵, V¹⁹ and K²⁴ proteins. The most significant changes (extending outside of the red dashed lines) are observed for L¹⁵(+P) and V¹⁹(+P). Error bars on the Δ(chemical shift) are ±0.7 ppm.

Figure 2

Differences in ¹³C' chemical shift between phosphorylated and non-phosphorylated LRAP as a function of residue and condition. Phosphorylation produced significant chemical shift perturbations (above the red dashed line) for several conditions at L¹⁵ and V¹⁹. Error bars on the Δ(chemical shift) are ±0.7 ppm.

Figure 3

REDOR dephasing curves for V¹⁹L²³(+P) as a function of ionic strength (pH=7.4). The distance becomes measurably more extended, possibly to a β-sheet structure, at high ionic strength (red circles, 8.3 Å) when compared to standard conditions (blue diamonds, 5.9 Å), while the low ionic strength does not show a significant difference (green triangles, 6.2 Å). Error bars are only shown for one data set for clarity, but are representative of the errors in each data set. The standard canonical structures are also shown: dots (α-helix), dot-dash (random coil) and dash (β-sheet).

Figure 4

¹³C{¹⁵N} REDOR show that K²⁴S²⁸ changes structure as a function of binding, and as a function of phosphorylation. The solid lines are the best fit distances: (red 4.2 Å, blue 4.8 Å, and green 5.7 Å). Binding transforms the phosphorylated peptide from a perfect helix to a combination of two structures, helix and either random coil or β-sheet. Dephosphorylation extends the structure further and is consistent with a combination of random coil, helix and β-sheet.⁴ Data generated from deconvolution are also shown (open blue squares) with the dashed blue line fit of 4.6 Å.

Figure 5

Comparison of the ¹³C' chemical shifts for K²⁴S²⁸ at 23 °C. Bottom: K²⁴S²⁸(+P) lyophilized from solution. Middle: K²⁴S²⁸ (+P) bound to HAP under standard conditions, pH=7.4, IS=0.15M. Top: K²⁴S²⁸(−P) bound to HAP under standard conditions, pH=7.4, IS=0.15M. Dark blue: experimental data; red: simulated fit; green, pale blue and maroon: deconvoluted components.

Figure 6

¹³C{³¹P} REDOR dephasing curves for K²⁴(pS) bound to HAP under standard conditions showing that phosphorylation results in a closer association of K²⁴ 22 with the surface.

Figure 7

One-dimensional spectra shown for L¹⁵(−P) under each condition, as indicated. Left, −45 °C, right, 23 °C. Each sample has reduced signal to noise compared to the spectra when taken under frozen conditions, indicating mobility. The spectra at pH=7.4, IS=0.2 and 0.15 M also show an increased intensity in the isotropic peak (marked with an arrow) relative to the spinning side bands, suggesting even more mobility under these conditions, however in all cases, the data suggests that the mobility is restricted.