TGF-β1 autocrine signalling and enamel matrix components

Transforming growth factor-β1 (TGF-β1) is present in porcine enamel extracts and is critical for proper mineralization of tooth enamel. Here, we show that the mRNA of latent TGF-β1 is expressed throughout amelogenesis. Latent TGF-β1 is activated by matrix metalloproteinase 20 (MMP20), coinciding with amelogenin processing by the same proteinase. Activated TGF-β1 binds to the major amelogenin cleavage products, particularly the neutral-soluble P103 amelogenin, to maintain its activity. The P103 amelogenin-TGF-β1 complex binds to TGFBR1 to induce TGF-β1 signalling. The P103 amelogenin-TGF-β1 complex is slowly cleaved by kallikrein 4 (KLK4), which is secreted into the transition- and maturation-stage enamel matrix, thereby reducing TGF-β1 activity. To exert the multiple biological functions of TGF-β1 for amelogenesis, we propose that TGF-β1 is activated or inactivated by MMP20 or KLK4 and that the amelogenin cleavage product is necessary for the in-solution mobility of TGF-β1, which is necessary for binding to its receptor on ameloblasts and retention of its activity.


Supplementary Note
Dynamic light scattering (DLS) In main Fig. 4, the light source of DLS was a solid-state laser with an output of 100 mW and a wavelength of 533 nm. The intensities at scattering angles, , from 30 to 100° with a step of 10° were simultaneously measured using eight optical fibers connected to eight photodiodes. The accumulation time to complete each measurement was 240 s. The second autocorrelation function, g (2) (, t), at time t was measured for use in calculating the inverse decay time, , of a particle: where K is an instrumental constant and i represents the i-th particle in the solution. The measurement was repeated consecutively four times in each run to estimate the error in The g (2) (, t) function was analyzed using an ALV-5000E correlation system. Both CONTIN and NNLS fittings were used to estimate  for each  to determine whether the results were independent of the analysis method. In the calculation of the intensity-percent and mass-percent fittings were compared to estimate the applicability.
The apparent translational particle diffusion coefficient, D, was calculated from the slope of the  vs. q 2 plot: where q is the scattering vector related to , and wavelength , and solution refractive index n: The hydrodynamic radius (particle radius), R H , was calculated using the Stokes-Einstein relationship: where k B , T, and h are the Boltzmann constant, the absolute temperature of the solution, and the solution viscosity, respectively. The h was measured using a rolling-ball type viscometer (Lovis 2000M, Anton-Parr, Co. Ltd.)

Extraction of TGF- in enamel matrix
Three TGF- isoforms (TGF-1 to -3) have been identified in mammalian. Those isoforms possess over 97% of amino acid sequence homology within mature bioactive molecule 1 . Each of isoforms has also very high amino acid sequence among species (see Supplementary Fig. S2).
Certain consensus sequences have been proposed for heparin-binding proteins, such as BBXB or BXBB, where B denotes a positively charged amino acid residue 2 . We found that the mature TGF-1 possesses one heparin-binding site (R372-K375) conserved in mouse, rat, human and pig (see Supplementary Fig. S2), but not in the mature TGF- or TGF-. Based on this finding, we tried to isolate TGF- isoforms in both N1 and AL extracts with heparin affinity chromatography.

Identification of amelogenin-TGF-1 complex in vivo
The identification of TGF-1 in vivo was determined by ELISA. The fractions (10 g each) isolated by heparin column and each of purified P173, P162, P148 and P103 amelogenins (6.25 In this study, we qualitatively modified a sandwich ELISA to know if amelogenin cleavage products bind to TGF-1 in vivo. We first used the amelogenin antibody as the detection antibody as well as Western blot analysis, but the significant contamination (see Supplementary Each amelogenin was fractionated by RP-HPLC with the same system as their purification. Each fraction was lyophilized, dissolved into 200 µL of water, and aliquots (50 µL) were used for the ALP-HPDL system. Amelogenins only were also incubated and fractionated by RP-HPLC as controls. Total amounts (ng) of original or bound TGF-1 in fractions obtained from RP-HPLC before and after in vitro binding experiments against one mg of P173, P162, P148 and P103 amelogenins were calculated from the standard CF-hTGF-1 (0.3 ng mL -1 ). The resulting data is shown in Supplementary Fig. S8 and main Table 1. (R&D Systems). The TGF-β1 receptor inhibitor, SB431542, was applied to a final concentration of 1 mM into the ALP-HPDL system for examination of the influence against the ALP-inducing activity increased by the application of samples. In controls, the ALP-inducing activity in HPDL cells was enhanced by rhTGF-β1.