Figure 1.
Functional domains and cleavage products of the tooth enamel protein amelogenin.
Besides the 180 amino acid full length amelogenin, the developing enamel matrix contains two major fragments, the tyrosine-rich amelogenin peptide (TRAP, AA1–45) and the leucine-rich amelogenin peptide (LRAP, AA1–33 & 155–180). For our amelogenin structure determination using 3D-NMR, three fragments were generated: Amel-M (AA1–92), Amel-N (AA34–154, non-LRAP), and Amel-C (AA86–180). Based on our structural data, we now distinguish between four major amelogenin domains, (i) the TRAP domain (AA1–45), (ii) the coil domain (AA46–125), (iii) the PXX repeat domain (AA126–164), and the hydrophilic C-terminal domain (AA165–180).
Figure 2.
Two-dimensional 1H-15N HSQC spectra of amelogenin fragments Amel-N, Amel-C, and Amel-M.
Two-dimensional HSQC spectroscopy was conducted under identical conditions (pH = 5.5) for all three amelogenin fragments, Amel-N, Amel-M, and Amel-C. Individual peaks were labeled with residue names and numbers, either directly or with arrows. His-tag residue peaks were marked with an asterisk. Aliased or folded peaks were labeled with an (f). Y26 (7.60, 120.22) was not shown in the figure. 1H–15N HSQC spectra of individual fragments are shown as follows Amel-N (A), Amel-C (B), Amel-M (C).
Figure 3.
Summary of chemical shift deviations, J-couplings, heteronuclear NOEs and inter-residue NOE signals.
(A and B) δCα and δHα chemical shift differences were obtained by subtracting published amino acid residue average chemical shifts (BMRB Database Statistics) from measured Cα and Hα chemical shifts of amelogenin assigned amino acid residues. C, JHNα(3) coupling values of amelogenin amino acid residues. D, Summary display of heteronuclear NOE values. The L15-W45 regions featured average NOE values of 0.6 while the central regions of the amelogenin molecule displayed NOEs between 0.5 and −0.2 (0.3 average), suggesting that the L15-W45 region contained a relatively rigid structure while the M1-N14 and L46-D180 regions were more flexible in comparison. E, Summary display of Interresidue NOE signals. The interresidue NOEs were classified into dNN, daN and dbN signals and each class included (i, i+1), (i,i+2) and (i, i+3) subcategories. dαN(i,i+1) and dβN(i,i+1) signals were denser between amino acids 100 to 170. Interresidue NOEs equal to or more than four amino acids apart (dxN(i, j)) and interresidue NOE that were three amino acid apart (dNN(i,i+3), dαN(i,i+3), dβN(i,i+3)) were mainly observed at the amelogenin N-terminus.
Figure 4.
Solution NMR structure of the amelogenin TRAP region (AA1-45).
A, Structural diagrams based on backbone traces from 6 selected conformers with lowest target functions calculated using the DYANA software program. Superimposed conformers from amelogenin TRAP region (AA1-45). B, Ribbon diagram representation of amelogenin TRAP region structure. The two regions (S9-V19) and (K24-I30) adopted α-helix like secondary structure and interacted with each other through Y17 and W25 side-chains. The region between V19-L23 formed a turn.
Table 1.
Statistics for Mouse Amelogenin Structure Determination.
Figure 5.
Solution NMR structure of the central and c-terminal amelogenin (AA46-180).
A, Alignment of ten selected conformers from the TRAP-neighboring region (AA46-85) with lowest target energies calculated using the DYANA software program. This fragment was characterized by turns and coils interrupted by an α-helix at V53-Q56 and an unusual 310 – helix at P74-Q76. B, Backbone ribbon representation and side chain heteroatom representation of one TRAP neighbor region (AA46-85) lowest energy structure. The position of the α-helix and rare 310 – helix are highlighted. C, Alignment of ten selected Amel-C conformers (AA86-180) with lowest target energies, once more calculated using the DYANA software program. Note the repetition of fairly similar PXX conformers along the polyproline type II helix region. D, Backbone ribbon representation and side chain heteroatom representation of one Amel-C lowest energy structure.
Figure 6.
Solution NMR structure of the entire full-length mouse amelogenin (M180).
(A and B), Alignment of 10 lowest energy conformers (A) and line structure (B) of the full-length mouse amelogenin (M180) based on full-length amelogenin NMR constraints calculated by DYANA. Helices are demarked by cartoons. Note the position of the 310 – helices at P74–Q76.
Figure 7.
Amelogenin domains and full-length molecule: Interactions and assembly behavior.
(A and B), Interaction between fragments and the full-length amelogenin as revealed by solution NMR. There were only weak interactions between Amel-N and Amel-C at T21, S28, and Q32 (chemical shift) while all others overlapped (A). Comparison of Amel-M (green color) and full length mouse amelogenin (red color) two-dimensional 1H-15N HSQC spectra at neutral pH (B). The Amel-M amino acid residue peaks which were absent or dramatic reduced in fAmel spectra were marked with hollow bars, including Y34, S36, Y37, G38, Y39, E40, G43, G44, W45, L46, H47, H48, Q49, I51, V53, L54, S55, Q56, H58. Peaks which were present in both Amel-M and full length spectra were marked with solid bars, including T63, L64, H68, H69, V72, V73, A75, Q76, A80, Q82, Q83, M85, V88, G90, H91, S93, M94, T95, T97, Q101, N103, I104, S107, A108, Q109, F112, Q113, Q117, I121, S125, H126, Q127, M129, Q130, Q132, S133, L135. Weak or lost HSQC peaks between AA34-62 indicate that the N-terminal Amel-M region may be involved in amelogenin nanosphere assembly (B). Analytical ultracentrifugation was used to demonstrate the self-assembling properties of Amel-N, Amel-M, Amel-C and of the full-length M180 amelogenin at pH 5.5 and pH 7 (C–J). C, the full-length M180 fragment displayed a monomeric peak at 26.1 kDa and heterogeneous assemblies with sizes between 1050 kDa to 1240 kDa at pH 7.0. (D–F), two of the three amelogenin fragments employed in this study Amel-M and Amel-C, were mostly monomeric (E,F), while Amel-N also featured high molecular weight self-assemblies in addition to a single peak (D). (G–J), at pH 5.5, all three fragments Amel-N, Amel-M and Amel-C as well as the full-length M180 revealed singular peaks and no higher molecular weight assemblies, suggesting monomeric distribution of these fragments and of the full-length amelogenin at pH 5.5.
Figure 8.
Model of basic amelogenin assembly.
Chemical shift analyses in conjunction with ultracentrifugation results from this study identify the the amelogenin N-terminus as the major nanosphere assembly site (A). In addition, our AFM and TEM data reveal effective nanosphere diameters as 25–30 nm (B). Based on the analytical ultracentrifugation data presented here, we have calculated a total of 50 molecules per nanosphere (1050 kDa, major peak) and assumed 11.9 nm as the length of the full-length amelogenin length based on our NMR constraint structure (C). The structural model presented here (D) was directly generated by MolMol based on these parameters and molecules were visualized using the VMD software based on N-terminal ipsilateral interactions and weak C-terminal ipsilateral interactions. Assemblies were arranged as spheres around a hollow center region with C-termini pointing toward the outside of the assembly.