Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Oct 6.
Published in final edited form as: Arch Oral Biol. 2018 Jun 6;93:187–194. doi: 10.1016/j.archoralbio.2018.06.001

Identification of major matrix metalloproteinase-20 proteolytic processing products of murine amelogenin and tyrosine-rich amelogenin peptide using a nuclear magnetic resonance spectroscopy based method

Garry W Buchko a,b,*, Rajith M J Arachchige a, Jinhui Tao a, Barbara J Tarasevich a, Wendy J Shaw a
PMCID: PMC7537640  NIHMSID: NIHMS1632260  PMID: 29960917

Abstract

Objective:

The aim of this study was to identify major matrix metalloproteinase-20 (MMP20) proteolytic processing products of amelogenin over time and determine if the tyrosine-rich amelogenin peptide (TRAP) was a substrate of MMP20.

Design:

Recombinant 15N-labeled murine amelogenin and 13C,15N-labeled TRAP were incubated with MMP20 under conditions where amelogenin self-assembles into nanospheres. Digestion products were fractionated by reverse-phase high-performance liquid chromatography at various time points. Product identification took advantage of the intrinsic disorder property of amelogenin that results in little change to its fingerprint 1H-15N heteronuclear single-quantum coherence nuclear magnetic resonance spectrum in 2% acetic acid upon removing parts of the protein, allowing cleavage site identification by observing which amide cross peaks disappear.

Results:

The primary product in five out of the six major reverse-phase high-performance liquid chromatography bands generated after a 24 hour incubation of murine amelogenin with MMP20 were: S55-L163, P2-L147, P2-E162, P2-A167, and P2-R176. After 72 hours these products were replaced with five major reverse-phase high-performance liquid chromatography bands containing: L46-A170, P2-S152, P2-F151, P2-W45, and short N-terminal peptides. TRAP was completely digested by MMP20 into multiple small peptides with the initial primary site of cleavage between S16 and Y17.

Conclusions:

Identification of the major MMP20 proteolytic products of amelogenin confirm a dynamic process, with sites towards the C-terminus more rapidly attacked than sites near the N-terminus. This observation is consistent with nanosphere models where the C-terminus is exposed and the N-terminus more protected. One previously reported end-product of the MMP20 proteolytic processing of amelogenin, TRAP, is shown to be an in vitro substrate for MMP20.

Keywords: amelogenesis imperfecta, amelogenesis, enamel, biomineralization, TRAP, MMP20

1. Introduction

Amelogenesis is a highly coordinated, dynamic, biomineralization process that is governed by the complex interplay between matrix proteins and solutes. The foremost enzyme to process the matrix proteins is matrix metalloproteinase-20 (MMP20) (Bartlett & Simmer, 1999; Ryu et al., 1999), a protease present at the initial stage of amelogenesis through to the early maturation-stage (Simmer & Hu, 2002). Mutations to MMP20 that inactivate the protein (Gasse et al., 2013; Kim et al., 2005) contribute to the autosomal recessive non-syndromic condition amelogenesis imperfecta (Wright et al., 2011), a disorder with clinical phenotypes that include hypomaturation, hypocalcification, and hypoplasia. The ablation of MMP20 has similar consequences in MMP−/− transgenic mice, thin and brittle enamel with an absent or malformed rod pattern (Caterina et al., 2002). The overexpression of MMP20 in MMP20+/+ transgenic mice also leads to improperly formed enamel (Shin et al., 2014). Data suggests that this is due to excess β-catenin released by the MMP20 cleavage of cadherin cell-cell junctions in fibroblasts and epithelial cells near the ameloblasts (Shin, Suzuki, Guan, Smith, & Bartlett, 2016). The translocation of the β-catenin to ameloblast nuclei promotes premature cell migration that is responsible for the formation of soft enamel.

One of the primary targets of MMP20 is amelogenin, the dominant protein in the first stage of amelogenesis. The full length protein is not the only amelogenin initially present as the RNA transcripts undergo extensive alternative splicing to generate amelogenin proteins of different lengths and combinations (Bartlett et al., 2006; Gibson et al., 1991). Just as it has been suggested that these amelogenin splice-variants may play a role in amelogenesis (Fincham, Moradian-Oldak, & Simmer, 1999), it has also been suggested that the cleavage products of amelogenin may perform essential functions in the process (Bartlett & Simmer, 1999; Yang, Sun, Ma, Fan, & Moradian-Oldak, 2011). Consequently, identifying the major MMP20 cleavage products of amelogenin may provide insights into understanding amelogenesis. Using a combination of sodium dodecyl sulfate-polyacrylamide gel electrophoresis and reverse-phase chromatography/mass spectrometric analyses, many of the MMP20 cleavage products of recombinant porcine and murine amelogenin have been identified in vitro and shown to mostly agree with the cleavage products observed in vivo (Ryu et al., 1999). However, mass spectrometry can be too sensitive, detecting the minor as well as the major products of proteolysis, with quantification difficult (Scalbert et al., 2009).

Here, a combination of high-performance liquid chromatography and nuclear magnetic resonance (NMR) spectroscopy was used to identify primarily the major products of MMP20 cleavage of amelogenin at a concentration of three mg/mL (0.13 mM) at pH 7.4, a condition shown to produce nanospheres (Bromley et al., 2011). Solution NMR spectrometry is especially useful for identifying amelogenin cleavage products because, as an intrinsically disordered protein (Delak et al., 2009), there is little change in the fingerprint 1H-15N heteronuclear single-quantum coherence spectrum of amelogenin in 2% acetic acid upon removing (Zhang, Ramirez, Liao, & Diekwisch, 2011), altering (Buchko, Lin, Tarasevich, & Shaw, 2013; Buchko & Shaw, 2015), or recombining (leucine-rich amelogenin protein, LRAP) (Buchko, Tarasevich, Roberts, Snead, & Shaw, 2010; Tarasevich et al., 2015) parts of the protein. Furthermore, the major product will be readily identifiable over any minor products due to concentration dependent intensity differences for each product in an 1H-15N heteronuclear single-quantum coherence spectrum. In addition to our studies with full length amelogenin, a method to prepare recombinant, 15N- and 13C-labelled tyrosine-rich amelogenin peptide (TRAP) is also described and used to unambiguously show that TRAP is a substrate for MMP20 in vitro. TRAP was first identified in ameloblasts in 1981 (Fincham, Belcourt, Termine, Butler, & Cothran, 1981) and is one of the major products of MMP20 digestion of amelogenin (Nagano et al., 2009; Ryu et al., 1999).

2. Materials and methods

2.1. Materials

Full-length recombinant murine amelogenin (M179, P2-D180, the N-terminal methionine is removed in Escherichia coli by methionine aminopeptidase (Bonde & Bulow, 2012)) without an affinity tag was 15N-labeled, expressed, and purified from E. coli following a previously described protocol (Buchko & Shaw, 2015). The catalytic domain of recombinant human MMP20 was purchased from Enzo Life Sciences, Inc. (Farmingdale, NY) and 4–12% bis-tris polyacrylamide gels from Invitrogen (Carlsbad, CA). All chemicals were purchased from Research Products International Corporation (Mount Prospect, IL) except trifluroacetic acid (TFA), glacial acetic acid, and HPLC grade acetonitrile (Sigma Chemical Company, St. Louis, MI).

2.2. Cloning, expression, and purification of murine TRAP

The DNA encoding the primary amino acid sequence of murine TRAP (M1-W45) with a 21-residue, N-terminal extension (MAHHHHHHMGTLEAQTQGPGS-) for metal-affinity purification (nTRAP), was codon-optimized for E. coli, chemically synthesized, and inserted into the expression vector pJexpress414 (DNA2.0, Menlo Park, CA). The recombinant plasmid was then transformed into competent E. coli BL21(DE3) cells (Novagen, Madison, WI) following a heat shock method. To obtain uniformly 15N-, 13C-labeled nTRAP, the transformed cells (37 °C) were grown in 750 mL of minimal medium (Miller) containing 15NH4Cl (1 mg/mL), D-[13C6]glucose (2.0 mg/mL), NaCl (50 μg/mL), MgSO4 (120 μg/mL), CaCl2 (11 μg/mL), Fe2Cl3 (10 ng/mL) and the antibiotic ampicillin (200 μg/mL). When the cell culture attained an OD600 reading of ~0.8, it was transferred to a 25 °C incubator and protein expression induced with isopropyl β-D-1-thiogalactopyranoside (0.026 μg/mL). Approximately four hours later the cells were harvested by mild centrifugation and frozen at −80 °C.

To purify nTRAP, the frozen pellet was thawed, resuspended in ~ 25 mL of 8 M urea, 100 mM NaPO4, pH 8.2, and passed three times through a French press (SLM Instruments, Rochester, NY). Following 60 s of sonication the insoluble cell debris was pelleted by centrifugation (10 °C) at 25,000g for 60 min in a JA-20 rotor (Beckman Instruments, Fullerton, CA). The supernatant was then applied onto a ~ 25 mL Ni-NTA affinity column (Qiagen, Valencia, CA) pre-equilibrated with solubilization buffer (8 M guanidinium hydrochloride, 100 mM NaPO4, pH 8.2). The column was washed stepwise by gravity (room temperature) with 40, 40, 40, 20, and 50 mL of buffer (6 M guanidinium hydrochloride, 100 mM NaPO4, pH 8.2) containing 0, 10, 20, 50, and 500 mM imidazole, respectively. Recombinant TRAP eluted exclusively in the 500 mM imidazole fraction. This fraction was further purified by making it acidic with 50 μL of TFA (0.1%) and applied onto an XBridge Preparative C18 (5 μm, 10×250 mm) reverse-phase column (Waters, Milford, MA) pre-equilibrated with 100% Buffer A (0.05% TFA in water). The following gradient scheme was applied with Buffer B (70% aqueous CH3CN in 0.05% TFA) and a flow rate of 2.5 mL/min: Step 1 – 0.5 Column Volume (CV), 100% Buffer A; Step 2 – 0.2 CV, linear gradient 0 to 45% Buffer B; Step 3 – 4.0 CV, linear gradient 45 to 85% Buffer B). The fractions containing nTRAP (~68% Buffer B) were pooled, frozen (−80 °C), and lyophilized. To effect removal of the N-terminal tag, lyophilized nTRAP was resuspended in 3C protease cleavage buffer (150 mM NaCl, 20 mM Tris, pH 7.6) at a concentration of ~ 1 mg/mL (yields were ~ 15 mg/L minimal media) and 3C protease added (1 μg per 50 μg target protein). This solution was cloudy and digestion under the standard protocol was extremely slow (standing overnight at 4 °C) (Choi et al., 2011). To accelerate cleavage, the reaction was performed at room temperature (~ 22 °C) with gentle agitation over 48 hours with a second addition of 3C protease (1 μg per 50 μg target protein) after the first 24 hours. The reaction mixture was then centrifuged, supernatant removed (little product present), and the pellet re-dissolved in 0.5 to 1 mL of 2% acetic acid prior to HPLC purification using the protocol described below for the MMP20 digestion products of full length amelogenin. The final lyophilized product was weighed and a 0.25 mM solution prepared in NMR buffer (2% CD3CO2D, 7% D2O/91% H2O, pH 3.0) for NMR characterization (Buchko, Bekhazi, et al., 2008). Note that after 3C protease treatment the final product, rTRAP, contained four non-native residues (GPGS-) at the N-terminal and are labeled G1* through S4* in our discussions.

2.3. MMP20 digestions

For M179, a stock solution of 15N-labelled protein (~20 mg/mL) was prepared by gently shaking purified protein in water for two days. The solution was filtered through a 0.2 μm microspin filter unit (Lida, Kenosha, WI) and the concentration measured by ultra-violet absorption spectroscopy using a 280 nm extinction coefficient of 20,300 M−1cm−1 (calculated). From this M179 stock solution a 3 mg/mL (0.13 mM) solution of M180 was prepared in 1 mL of MMP20 reaction buffer (25 mM Tris, pH 7.4). Following confirmation of the pH with a meter reading, 3 μg of MMP20 was added and the reaction gently agitated at room temperature using a Fisher Scientific Multi-purpose tube rotator (10 rotations per minute). To monitor the reaction, a 25 μL aliquot was removed and suspended in 500 μL of 2% acetic acid prior to HPLC analysis. Upon observing little change in HPLC profiles after 72 hours, the solution was spun down by centrifugation and the supernatant removed. The pellet was resuspended in 1 mL of 2% acetic acid and the supernatant made 0.05% in TFA prior to HPLC analysis.

For rTRAP, 2 mg of 15N- and 13C-labelled protein was suspended in 1 mL of MMP20 reaction buffer. At pH 7.4, rTRAP was largely insoluble and visible as a snowy white precipitate. To this solution 2 μg of MMP20 was added and the reaction gently agitated at room temperature as described for M179. To monitor the reaction, a 25 μL aliquot was removed and suspended in 500 μL of 2% acetic acid prior to HPLC analysis. After 24 hours the reaction solution was centrifuged and the supernatant removed. The supernatant was made 0.05% in TFA and the precipitate dissolved in 500 μL of 2% acetic acid prior to loading on to the reverse-phase HPLC column. A second reaction in the same proportions, but at a volume of 500 μL, was performed and allowed to progress for 120 hours. The reaction mixture was treated following the description for the 24 hour digestion.

2.4. Reverse-phase HPLC purification of MMP20 and 3C protease digestion products

The protease digestion products of full length amelogenin, nTRAP, and rTRAP were applied onto an RESOURCE RPC 3 ml (15 μm, 6.4 × 100 mm) reverse-phase column (GE Healthcare, Piscataway, NJ) with the following purification gradient generally applied after loading the material using 100% Buffer A (Buffer A = water in 0.05% TFA, Buffer B = 70% aqueous CH3CN in 0.05% TFA, flow rate 3.0 mL/min: Step 1 – 0.5 CV, 100% Buffer A; Step 2 – 20.0 CV, linear gradient 0 to 100% Buffer B). Fractions were pooled, frozen at −80 °C, lyophilized, and resuspended in an NMR buffer consisting of 2% CD3CO2D, 7% D2O/91% H2O, pH 2.8). Because there was a small difference between the retention times of cleaved and uncleaved nTRAP, after lyophilization of the pooled fractions containing rTRAP it was sometimes necessary to redissolve in 2% acetic acid and reapply to the reverse-phase column to effect more complete purification. To better resolve late eluting MMP20 digestion product of M179 the following Steps were employed to generate a shallower gradient: Step 1 – 0.5 CV, 100% Buffer A; Step 2 – 0.2 CV, linear gradient 0 to 40% Buffer B; Step 3 – 20.0 CV, linear gradient 40 to 60% Buffer B.

2.5. NMR data collection and processing

Products purified by reverse-phase HPLC were lyophilized and resuspended in 250 – 300 uL of NMR buffer (2% CD3CO2D, 7% D2O/91% H2O, pH 3.0). Using Shegimi tubes, two-dimensional 1H-15N heteronuclear single-quantum coherence spectra were collected at 20 °C using Agilent or Bruker NMR spectrometers operating at a 1H resonance frequency of 600 or 750 MHz. Assignment of the MMP20 cleavage products of M179 were based primarily on comparison to the 1H-15N heteronuclear single-quantum coherence spectrum of previously assigned murine amelogenin containing a non-removable N-terminal tag MRGSHHHHHHGS- (Buchko, Tarasevich, Bekhazi, Snead, & Shaw, 2008). The same procedure was used to identify the major MMP20 cleavage products of rTRAP using the chemical shifts assignments reported here. For rTRAP, this required the further acquisition of three-dimensional HNCACB, HNCO, CC-TOCSY-NNH, and 15N-edited TOCSY data to make all the 1H, 13C, 15N chemical shifts assignments possible. These chemical shifts for rTRAP were deposited into the Biological Magnetic Resonance Data Bank database (http://www.bmrb.wisc.edu) under accession number BMRB-27415. To confirm the assignments for the analyzed MMP20 digestion products of rTRAP a three-dimensional HNCA data set was collected for each product.

3. Results and discussion

3.1. Analysis of the MMP20 digestion products of M179 with reverse-phase HPLC

The reaction was performed at a large M179:MMP20 ratio of 500:1 (weight:weight) to obtain enough product to readily follow the reaction with HPLC and analyze the products with solution state NMR spectroscopy. As shown in the time-course HPLC chromatograms of the digestion products in Figure 1A, M179, the major band with a retention time of 24.5 min (0 h, black), is converted into many products. Most of the products elute with a retention time between 22 and 26 min. Subtle changes in the shape of the HPLC chromatogram with time, especially in the 22 – 26 min region, indicates that the proteolytic process is dynamic. To acquire enough material for NMR analysis the digestion products were analyzed in detail 24 and 72 hours after digestion. As shown in Figure 1B, a shallower HPLC elution gradient enabled the isolation of six different bands after a 24 hour reaction from the products eluting in the 22–26 min region in Figure 1A.

Fig. 1.

Fig. 1.

Open in a new tab

(A) HPLC chromatograms collected at various time points during the incubation of M179 with MMP20. The entire MMP20 digestion mixture was applied to the reverse-phase column by making the solution 2% in acetic acid. As described in Figure 2, bands labeled 1 and 2 are soluble in the initial reaction mixture while bands 3 – 5 are in the precipitate. (B) The HPLC chromatogram collected with a shallow elution gradient of the late eluting products of the MMP20 digestion of M179 after a 24 hour incubation. The products in bands labeled 1e through 6e represent earlier products generated by the digestion of M179 with MMP20 that were collected, lyophilized, and then analyzed by NMR spectroscopy.

Figure 2 A is the HPLC chromatogram after incubating M179 with MMP20 for 72 hours. The HPLC profile for the blue chromatogram was obtained by making the solution 2% in acetic acid prior to injection on the reverse-phase column. Visible white precipitate present in solution was observed to disappear. In general, aside from some intensity differences in the major bands, this 72 hour HPLC profile is similar to the profile observed after 48 hours (cyan chromatogram, Figure 1A). The red chromatogram in Figure 2A is the HPLC profile of the products remaining in solution upon centrifuging the reaction products after a 72 hour incubation. Two major products, numbered 1 and 2 in red, are present and most of the material that previously eluted in the 22 – 26 min region is gone. The pellet present after centrifugation was resuspended in 2% acetic and the HPLC chromatogram, shown using a shallower HPLC gradient in Figure 2B, contained only three bands numbered 3 – 5 in blue. In summary, out of the five major M179 cleavage products after a 72 hour incubation, two were soluble and three insoluble. These bands appear to corresponds to the products numbered 1 – 5 in the 48 hour (cyan) HPLC chromatogram in Figure 1A. Note that the HPLC profile at 24 hours shown in Figure 1A represents soluble product. Nothing significant was observed in the HPLC profile of the 24 hour digestion aliquot after centrifugation of the reaction mixture, removal of supernatant, and resuspension of any material that might be present with 2% acetic acid. Hence, digestion of M179 with MMP20 results in soluble cleavage products first with the appearance of insoluble cleavage products after 24 hours.

Fig. 2.

Fig. 2.

Open in a new tab

(A) HPLC chromatograms after a 72 hour incubation of M179 with MMP20. The blue chromatogram is the entire MMP20 digestion mixture solubilized with 2% acetic acid solution and the red chromatogram is the soluble fraction after centrifugation. The products in bands labeled 1 through 5 represent late products generated by the digestion of M179 with MMP20 that were collected, lyophilized, and then analyzed by NMR spectrometry. (B) The HPLC chromatogram of the products in the pellet after centrifugation of the 72 hour MMP20 digestion mixture of M179. The pellet was solubilized in 2% acetic acid. The data was collected with a shallower gradient, and therefore, the time in Figure B is shifted plus 10 minutes per point.

3.2. Cleavage-site identification from 1H-15N heteronuclear single-quantum coherence spectral comparisons

Two-dimensional 1H-15N heteronuclear single-quantum coherence spectra were collected for the major HPLC bands observed after incubation of M179 with MMP20 for 24 and 72 hours, bands 1e - 6e and 1 – 5, respectively. Because there is little change in the fingerprint 1H-15N heteronuclear single-quantum coherence spectrum of amelogenin in 2% acetic acid upon removing N- and/or C-terminal regions of amelogenin (Zhang et al., 2011), the site of cleavage by MMP20 could be identified by observing which amide cross peaks disappeared in the 1H-15N heteronuclear single-quantum coherence spectrum. The power of this method is illustrated in Figure 3, an overlay of the 1H-15N heteronuclear single-quantum coherence spectra for M179 and the major product in band 3. All the cross peaks in the two spectra essentially overlap with each other except for the 21 green cross peaks in the 1H-15N heteronuclear single-quantum coherence spectrum of band 3 circled in black. Three of these cross peaks are near a cross peak in the 1H-15N heteronuclear single-quantum coherence spectrum of M179 without an associated overlapping cross peak from the major product in band 3. The latter three cross peaks are circled cyan and correspond to L150, F151, and S152. The remaining 18 unique missing cross peaks correspond to the non-proline C-terminal amides of residues of M179, M153 to D180. Hence, the major product in band 3 is a protein corresponding to P2 – S152 with the three non-overlapped cross peaks residues with slightly perturbed amide chemical shifts due to the removal of the N-terminal region. In a similar fashion, it was possible to unambiguously identify the major products in most of the other HPLC bands, as indicated in Tables 1 and 2, except for bands 2e and 1. In the 1H-15N heteronuclear single-quantum coherence spectrum of band 2e, both expected N- and C-terminal amide cross peaks, L3 and D180 respectively, were observed. Because the retention time of band 2e was markedly different from M179 on the reverse-phase column, the two products likely consist of a cleavage product towards the N-terminal and a cleavage product towards the C-terminal of M179.

Fig. 3.

Fig. 3.

Open in a new tab

Overlay of the 1H-15N heteronuclear single-quantum coherence spectra for M179 (red) and the M179 HPLC digestion band 3 (green) identified as P2-S152. Spectra were recorded at a 1H resonance frequency of 600 MHz, 20 °C. Green amide cross peaks that do not superimpose on red cross peaks are circled in black and the assignments labeled. The chemical shift of three of the non-superimposable amide cross peaks (L150-S152) move and these red cross peaks are circled in cyan. Not shown are the downfield cross peaks for the ring amides of the three tryptophan residues present in both proteins.

Table 1.

Summary of the early (24 h) MMP20 digestion products of murine amelogenin.

HPLC Band Fraction Product
1e Soluble S55-L163
2e Soluble P2-X, Y-D180
3e Soluble P2-L147
4e Soluble P2-R176
5e Soluble P2-E162
6e Soluble P2-A167
Open in a new tab

Table 2.

Summary of the late (72 h) MMP20 digestion products of murine amelogenin.

HPLC Band Fraction Product
1 Soluble Very small peptide(s)
2 Soluble L46-A170
3 Precipitant P2-S152
4 Precipitant P2-F151
5 Precipitant P2-W45 (TRAP)
Open in a new tab

In conclusion, after 24 hours the starting material, M179, was not a major protein in any of the HPLC bands, having been cut at least once by MMP20. Only one of the HPLC bands contained a product that had been unambiguously cut twice, band 1e identified as S55 – L163. The majority of the other identified products were cleavages near the C-terminal of M179. This observation confirms previous conclusions suggesting that MMP20 favored cutting M179 towards the C-terminus first (Ryu et al., 1999). The peptide TRAP, a previously reported product of the MMP20 digestion of M179, was not observed until after longer digestion times (band 5). While the major products have been identified here by analysis of 1H-15N heteronuclear single-quantum coherence spectra, a number of minor products were generated as indicated most clearly in the HPLC chromatogram in Figure 2A that is populated with many weaker HPLC bands. Consequently, it is not surprising that previous mass spectral analyses of the MMP20 digestion products of M179 identified so many products.

Out of the 11 major HPLC bands that were analyzed by NMR, the 1H-15N heteronuclear single-quantum coherence spectrum of the products in band 1 was the most complicated of all the spectra and is reproduced in Figure 4. While the 1H-15N heteronuclear single-quantum coherence of band 1 contained many cross peaks, only a few of the cross peaks tracked closely to cross peaks present in the 1H-15N heteronuclear single-quantum coherence spectrum of M179. Moreover, the cross peaks surveyed a range of intensities, suggesting a mixture of products. Figure 5 is an sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel that includes the five major bands following the digestion of M179 with MMP20 for 72 hours. The lane containing band 1 contained no detectable products, even when heavily loaded (data not shown). Because the product of HPLC band 5 is P2-W45 (TRAP), a 44-residue peptide with a molecular weight of 5.2 kDa, the products in band 1 must all be smaller than 44-residues and below the detection limit of the sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel. A few of the more intense cross peaks in the 1H-15N heteronuclear single-quantum coherence spectrum of band 1 did trace closely to the assigned cross peaks in the 1H-15N heteronuclear single-quantum coherence spectrum of M179 and these are tentatively assigned in Figure 4. The remaining cross peaks that are perturbed relative to the 1H-15N heteronuclear single-quantum coherence spectrum of M179 likely represent amides near the termini of the peptide products. While the sequence of these peptide products cannot be identified on the basis of data collected on 15N-labeled samples alone, the spectrum suggests that TRAP, a previously identified product of MMP20 digestion of amelogenin, is also a substrate for MMP20. Because TRAP had previously been reported as an end product of MMP20 digestion of amelogenin (Nagano et al., 2009; Ryu et al., 1999; Yamakoshi, 2011), 13C-, 15N-labeled TRAP was prepared to unambiguously determine if MMP20 could hydrolyze TRAP, and to identify some of the major cleavage projects.

Fig. 4.

Fig. 4.

Open in a new tab

The 1H-15N heteronuclear single-quantum coherence spectrum for band 1, a soluble product of the 72 hour MMP20 digestion of M179, collected in 2% acetic acid, 20 °C, at a 1H resonance frequency of 750 MHz. Cross peaks that could be readily traced to assigned resonances in the 1H-15N heteronuclear single-quantum coherence spectrum of M179 are identified.

Fig. 5.

Fig. 5.

Open in a new tab

A 4–12% bis-tris sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel of M179 and the 72 hour MMP20 digestion products of M179 using the NuPAGE Bis-Tris Electrophoresis System with a MES-SDS running buffer. The well contents are labeled on the top of the gel and the molecular weight markers are indicated on the right side in kDa along with the position of TRAP.

3.3. Purification of rTRAP

While full length murine amelogenin without any fusion regions added to the native protein sequence was purified by exploiting its solubility in 2% acetic acid (Buchko & Shaw, 2015), attempts to use a similar method for a construct containing only the primary amino acid sequence of murine TRAP (M1-W45) were unsuccessful. It was suspected that the procedure failed because, unlike M179, TRAP was pervious to the 3,500 Dalton membrane filters used during the dialysis steps in 2% acetic acid. It was possible to isolate recombinant nTRAP, a construct containing a 21-residue N-terminal tag with a hexa-histidine region, by nickel affinity chromatography. Guanidium hydrochloride and other impurities were removed by semi-preparative reverse-phase HPLC. An indication of the purity of nTRAP after the two chromatography steps is shown by the HPLC chromatogram in Figure 6A (dotted black line) containing a single band.

Fig. 6.

Fig. 6.

Open in a new tab

(A) Overlay of the HPLC chromatograms of nTRAP (black dashed line) and following digestion with 3C protease (light blue) to generate rTRAP. (B) The HPLC chromatograms following the incubation of rTRAP with MMP20 for 24 (blue and red) and 120 hours (green). The chromatograms at 24 hour are for the products in the soluble fraction following centrifugation (red) and the pellet (blue) re-suspended in 2% acetic acid. After 120 hours, products were only observed in the soluble fraction following centrifugation. The products in HPLC bands labeled 1 through 4 were collected, lyophilized, and then analyzed by NMR spectroscopy.

Under the optimal conditions for 3C protease activity at pH 7.5, nTRAP visibly precipitated out of solution as previously observed for synthetically prepared TRAP at this pH (Tan, Leung, Moradian-Oldak, Zeichner-David, & Fincham, 1998). While the reaction was slow, as shown by the cyan HPLC chromatogram in Figure 5A, cleavage occurred over time with gentle agitation of the solution at room temperature. Two new bands appeared, a major band with a slightly longer retention time then observed for uncleaved nTRAP and a minor band at ~12.5 min. Analysis of the NMR data confirmed that the major digestion product was rTRAP. By elimination it was assumed that the minor band was the complementary 17-residue peptide produced by 3C protease cleavage (rTRAP contained four “scar” residues after cleavage). Note that efforts to push the cleavage reaction to completion with the addition of excess 3C protease were unsuccessful. The ratio of cleaved:uncleaved nTRAP never increased beyond the ratio illustrated in the HPLC chromatogram in Figure 6A (cyan), approximately 2:1. The reason is unlikely due to impurities or chemical modifications because isolated uncleaved nTRAP from the first reaction could be digested (but also not to completion) in a second reaction.

3.4. Chemical shift assignment of rTRAP and characterization of the major MMP20 digestion products

Aside from a couple residues near the N- and C-termini, the 1H-15N heteronuclear single-quantum coherence spectrum for rTRAP mapped closely onto the 1H-15N heteronuclear single-quantum coherence spectrum of M179 (data not shown). Assignments for these non-overlapping residues were made through the analysis of routine three-dimensional backbone NMR experiments with the complete assignments shown in the 1H-15N heteronuclear single-quantum coherence spectrum of rTRAP illustrated in Figure 7.

Fig. 7.

Fig. 7.

Open in a new tab

Overlay of the 1H-15N heteronuclear single-quantum coherence spectra for rTRAP (red) and the MMP20 digestion product in HPLC band 2 (blue) identified as G1*-S16. Spectra were recorded at a 1H resonance frequency of 750 MHz, 20 °C. Not shown are the cross peaks for the rTRAP ring amide of W24 at 129.3 (15N) and 10.11 (1H) ppm and W45 at 128.9 (15N) and 10.05 (1H) ppm. Side chain amide resonance cross peak pairs are connected by a dashed line and the “scar” residues remaining after removal of the N-terminal tag are identified with an asterisk.

At the pH (7.4) used during the incubation of rTRAP with MMP20, precipitate was visible. Likely for this reason, at a rTRAP:MMP20 ratio of 500:1 (weight:weight), digestion of rTRAP was slow as illustrated in Figure 6B. However, if soluble substrate is necessary for MMP20 enzymatic activity, then there was still some soluble rTRAP in solution because after 24 hours approximately one quarter of the starting peptide was gone. After 120 hours the solution was clear and the HPLC chromatogram (green, Figure 6B) showed that the original rTRAP had been completely consumed and digested into many products. Two-dimensional 1H-15N heteronuclear single-quantum coherence spectra were collected for HPLC bands 1 through 4 generated after 24 hour incubation of nTRAP with MMP20. Using the same NMR approach employed to identify the MMP20 digestion products of M179, the site at which MMP20 cleaved rTRAP were identified by observing which amide cross peaks disappeared in the 1H-15N heteronuclear single-quantum coherence spectrum of the product in HPLC bands. The power of this method is further illustrated in Figure 7, an overlay of the 1H-15N heteronuclear single-quantum coherence spectra for rTRAP (red) and the major product in band 1 (blue). The 1H-15N heteronuclear single-quantum coherence spectrum of band 1 clearly has fewer cross peaks then observed in the 1H-15N heteronuclear single-quantum coherence of rTRAP with cross peaks corresponding to the amides of residues G1* through Y12 overlapping well. Analysis of the three-dimensional HNCA data for HPLC band 1 enabled the assignment of the four non-overlapping amide cross peaks to residues I13 through S16 and identification of the product as G1*-S16. In a similar fashion the major product in the other three HPLC bands were identified as described in Table 3. The product in HPLC band 4 is Y17-W45, the complementary product to the G1*-S16 peptide. Note that Y17-W45 is the only non-soluble digestion product observed in the precipitant fraction, suggesting that MMP20 cleavage of rTRAP between residues Y17-W45 produces soluble peptide products.

Table 3.

Summary of the early MMP20 digestion products of murine rTRAP.

HPLC Band Fraction Product
1 Soluble G1*-S16
2 Soluble S28-W45
3 Soluble 6-residue
4 Precipitant Y17-W45
Open in a new tab

3.5. Comparison to reported MMP20 digestion products and potential biological significance

The primary amino acid sequence of the two most studied amelogenins, porcine and murine, are 80% identical and 89% similar with the murine sequence seven residues longer (Toyosawa, O’HUigin, Figueroa, Tichy, & Klein, 1998). Despite such similarities, in vitro studies show a greater number of MMP20 cleavage products for murine over porcine amelogenin with some cleavage sites only observed in the equivalent position in one of the two sequences (Ryu et al., 1999). Our NMR-based identification of the major products of the MMP20 digestion of murine amelogenin added to the number of known products generated by MMP20. Out of the nine unambiguously NMR-identified major products of the MMP20 proteolytic processing of M179 after 24 and 72 hour incubation, only two were previously identified in equivalent positions for porcine amelogenin by mass spectrometric methods: P2-F167 and TRAP (Bartlett & Simmer, 1999; Ryu et al., 1999). It has been suggested that some of the previously reported differences between murine and porcine may be related to incubation times with MMP20 (Bartlett & Simmer, 1999), a conclusion supported by our HPLC and NMR data. Differences in the solution conditions may also contribute to differences in the MMP20 digestion profiles as previous in vitro studies have shown that proteolytic processing of amelogenin by MMP20, with regards to kinetics and cleavage pattern, is affected by the chemistry of the solution (eg: calcium and phosphate content) (Khan et al., 2013).

An MMP20 digestion product common to our study and all previously reported studies is TRAP. This small, N-terminal peptide has been identified in ameloblasts during the later stages of amelogenesis (Fincham & Moradian-Oldak, 1993) and it may be one of the amelogenin cleavage products with regulatory functions in the process (Bartlett & Simmer, 1999). For example, the MMP20 digestion of a human amelogenin containing a P41T substitution was retarded relative to the wild type protein (Wu, Cen, Yan, & DenBesten, 2003). It was suggested that this retardation of TRAP formation may be responsible for the amelogenesis imperfecta observed with this mutation.

At a TRAP:MMP20 ratio of ~5:1 (weight:weight) using TRAP isolated from pig, Nagano et al. did not observe any appreciable cleavage of TRAP after 24 hour incubation in 50 mM Tris-HCl, 5 mM CaCl2, at pH 7.4 and 37 °C (Nagano et al., 2009). While our study was performed at a larger TRAP:MMP20 ratio of 1000:1 (weight:weight) and lower temperature, it was under similar buffer conditions (25 mM Tris-HCl, 5 mM CaCl2) at the same pH (7.4). The explanation for the difference may be a post-translational modification to amelogenin, phosphorylation at S16 (Takagi, Suzuki, Baba, Minegishi, & Sasaki, 1984; Fincham & Moradian-Oldak, 1993). The recombinant TRAP used in our experiments was not phosphorylated at S16 and the initial site of cleavage was observed between S16 and Y17. Hence, a possible biological role for phosphorylation at S16 may be to prevent MMP20 from digesting TRAP. Perhaps dephosphorylation of TRAP serves to affect its removal after it has completed its regulatory role.

4. Conclusions

It is clear from the time-dependent changes to the profiles of HPLC chromatograms in Figure 1A and the major products identified after 24 and 72 hours that the proteolytic processing of M179 by MMP20 is dynamic. After 24 hours there was no evidence for any remaining starting material; all but one of the HPLC bands contained a major product with the C-terminal residue, D180, still present (band 2e) and only one major product unambiguously identified contained a product cleaved near the N-terminus (band 1e, S55-L163). This is consistent with previous findings suggesting that sites towards the C-terminus of amelogenin were more rapidly attacked by MMP20 than sites near the N-terminus (Ryu et al., 1999). At 24 hours the unambiguous major C-terminal sites of hydrolysis observed are after L147, E162, L163, and R176. At 72 hours the unambiguous C-terminal sites of hydrolysis observed are after F151, S152, and A170. This pattern suggests C-terminal cleavage largely progresses from sites near the end of the C-terminus. This observation is consistent with nanosphere models (Du, Falini, Fermani, Abbott, & Moradian-Oldak, 2005) and fluorescence experiments on single tryptophan amelogenins that suggest the C-terminus is exposed in self-assembled nanospheres and the N-terminus more protected (Bromley et al., 2011). By attacking sites near the C-terminus of amelogenin nanospheres, sites near the N-terminus may slowly become exposed to proteolysis, a process that may be assisted by the precipitation of digestion products. Note that all the early major MMP20 cleavage products are soluble in solution (Table 1) with insoluble products, including TRAP, generated towards the latter stages of the incubation (Table 2). If solubility is a prerequisite for function, then the early products of the proteolytic processing of amelogenin by MMP20 could still have functional roles in orchestrating enamel formation. This improved understanding of the MMP20 proteolytic processing products of amelogenin may provide new insights into understanding the molecular basis for enamel formation, which in turn may lead to new repair or regeneration strategies for tooth enamel (Palmer, Newcomb, Kaltz, Spoerke, & Stupp, 2008).

Acknowledgements

This research was performed at the Pacific Northwest National Laboratory, a facility operated by Battelle for the U.S. Department of Energy, including access to the W.R. Wiley Environmental Molecular Sciences Laboratory, a national scientific user facility sponsored by the U.S. DOE Biological and Environmental Research program. We thank Dr. Sarah D. Burton for insightful discussions and assistance in collecting the NMR data.

Funding

This research was supported by NIH-NIDCH Grant number DE-015347.

Footnotes

Conflict of Interest

The authors report no conflict of interests regarding this manuscript.

References

  1. Bartlett JD, Ball RL, Kawai T, Tye CE, Tsuchiya M, & Simmer JP (2006). Origin, splicing, and expression of rodent amelogenin exon 8. Journal of Dental Research, 85(10), 894–899. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bartlett JD, & Simmer JP (1999). Proteinases in developing dental enamel. Critical Reviews in Oral Biology & Medicine, 10(4), 425–441. [DOI] [PubMed] [Google Scholar]
  3. Bonde JS, & Bulow L (2012). Use of human amelogenin in molecular encapsulation for the design of pH responsive microparticles. BMC Biotechnology, 12, 25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bromley KM, Kiss AS, Lokappa SB, Lakshminarayanan R, Fan D, Ndao M, … Moradian-Oldak J (2011). Dissecting amelogenin protein nanospheres: characterization of metastable oligomers. Journal of Biological Chemistry, 286(40), 34643–34653. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Buchko GW, Bekhazi J, Cort JR, Valentine NB, Snead ML, & Shaw WJ (2008). 1H, 13C, and 15N resonance assignments of murine amelogenin, an enamel biomineralization protein. Biomolecular NMR Assignments, 2(1), 89–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Buchko GW, Lin G, Tarasevich BJ, & Shaw WJ (2013). A solution NMR investigation into the impaired self-assembly properties of two murine amelogenins containing the point mutations T21-->I or P41-->T. Archives of Biochemistry and Biophysics, 537(2), 217–224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Buchko GW, & Shaw WJ (2015). Improved protocol to purify untagged amelogenin - Application to murine amelogenin containing the equivalent P70-->T point mutation observed in human amelogenesis imperfecta. Protein Expression and Purification, 105, 14–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Buchko GW, Tarasevich BJ, Bekhazi J, Snead ML, & Shaw WJ (2008). A solution NMR investigation into the early events of amelogenin nanosphere self-assembly initiated with sodium chloride or calcium chloride. Biochemistry, 47, 6571–6582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Buchko GW, Tarasevich BJ, Roberts J, Snead ML, & Shaw WJ (2010). A solution NMR investigation into the murine amelogenin splice-variant LRAP (Leucine-Rich Amelogenin Protein). Biochimica et Biophysica Acta, 1804(9), 1768–1774. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Caterina JJ, Skobe Z, Shi J, Ding YL, Simmer JP, Birkedal-Hansen H, & Bartlett JD (2002). Enamelysin (matrix metalloproteinase 20)-deficient mice display an amelogenesis imperfecta phenotype. Journal of Biological Chemistry, 277(51), 49598–49604. [DOI] [PubMed] [Google Scholar]
  11. Choi R, Kelley A, Leibly D, Hewitt SN, Napuli A, & Van Voorhis W (2011). Immobilized metal-affinity chromatography protein-recovery screening is predictive of crystallographic structure success. Acta Crystallographica Section F Structural Biology Crystallography Communications, 67(Pt 9), 998–1005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Delak K, Harcup C, Lakshminarayanan R, Sun Z, Fan Y, Moradian-Oldak J, & Evans JS (2009). The tooth enamel protein, porcine amelogenin, is an intrinsically disordered protein with an extended molecular configuration in the monomeric form. Biochemistry, 48(10), 2272–2281. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Du C, Falini G, Fermani S, Abbott C, & Moradian-Oldak J (2005). Supramolecular assembly of amelogenin nanospheres into birefringent microribbons. Science, 307(5714), 1450–1454. [DOI] [PubMed] [Google Scholar]
  14. Fincham AG, Belcourt AB, Termine JD, Butler WT, & Cothran WC (1981). Dental enamel matrix - Sequences of 2 amelogenin polypeptides. Bioscience Reports, 1(10), 771–778. [DOI] [PubMed] [Google Scholar]
  15. Fincham AG, & Moradian-Oldak J (1993). Amelogenin post-translational modifications: carboxy-terminal processing and the phosphorylation of bovine and porcine “TRAP” and “LRAP” amelogenins. Biochemical and Biophysical Research Communications, 197(1), 248–255. [DOI] [PubMed] [Google Scholar]
  16. Fincham AG, Moradian-Oldak J, & Simmer JP (1999). The structural biology of the developing dental enamel matrix. Journal of Structural Biology, 126(3), 270–299. [DOI] [PubMed] [Google Scholar]
  17. Gasse B, Karayigit E, Mathieu E, Jung S, Garret A, Huckert M, … Bloch-Zupan A (2013). Homozygous and compound heterozygous MMP20 mutations in amelogenesis imperfecta. Journal of Dental Research, 92(7), 598–603. [DOI] [PubMed] [Google Scholar]
  18. Gibson CW, Golub E, Ding WD, Shimokawa H, Young M, Termine J, & Rosenbloom J (1991). Identification of the Leucine-Rich Amelogenin Peptide (LRAP) as the translation product of an alternatively spliced transcript. Biochemical and Biophysical Research Communications, 174(3), 1306–1312. [DOI] [PubMed] [Google Scholar]
  19. Khan F, Liu H, Reyes A, Witkowska HE, Martinez-Avila O, Zhu L, … Habelitz S (2013). The proteolytic processing of amelogenin by enamel matrix metalloproteinase (MMP-20) is controlled by mineral ions. Biochimica et Biophysica Acta, 1830(3), 2600–2607. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kim JW, Simmer JP, Hart TC, Hart PS, Ramaswami MD, Bartlett JD, & Hu JCC (2005). MMP-20 mutation in autosomal recessive pigmented hypomaturation amelogenesis imperfecta. Journal of Medical Genetics, 42(3), 271–275. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Nagano T, Kakegawa A, Yamakoshi Y, Tsuchiya S, Hu JC, Gomi K, … Simmer JP (2009). Mmp-20 and Klk4 cleavage site preferences for amelogenin sequences. Journal of Dental Research, 88(9), 823–828. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Palmer LC, Newcomb CJ, Kaltz SR, Spoerke ED, & Stupp SI (2008). Biomimetic systems for hydroxyapatite mineralization inspired by bone and enamel. Chemical Reviews, 108(11), 4754–4783. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Ryu OH, Fincham AG, Hu CC, Zhang C, Qian Q, Bartlett JD, & Simmer JP (1999). Characterization of recombinant pig enamelysin activity and cleavage of recombinant pig and mouse amelogenins. Journal of Dental Research, 78(3), 743–750. [DOI] [PubMed] [Google Scholar]
  24. Scalbert A, Brennan L, Fiehn O, Hankemeier T, Kristal BS, van Ommen B, … Wopereis S (2009). Mass-spectrometry-based metabolomics: limitations and recommendations for future progress with particular focus on nutrition research. Metabolomics, 5(4), 435–458. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Shin M, Hu YY, Tye CE, Guan XM, Deagle CC, Antone JV, … Bartlett JD (2014). Matrix metalloproteinase-20 over-expression is detrimental to enamel development: A Mus musculus model. PLos ONE, 9(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Shin M, Suzuki M, Guan X, Smith CE, & Bartlett JD (2016). Murine matrix metalloproteinase-20 overexpression stimulates cell invasion into the enamel layer via enhanced Wnt signaling. Science Reports, 6, 29492. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Simmer JP, & Hu JCC (2002). Expression, structure, and function of enamel proteinases. Connective Tissue Research, 43(2–3), 441–449. [DOI] [PubMed] [Google Scholar]
  28. Takagi T, Suzuki M, Baba T, Minegishi K, & Sasaki S (1984). Complete amino acid sequence of amelogenin in developing bovine enamel. Biochemical and Biophysical Research Communications, 121(2), 592–597. [DOI] [PubMed] [Google Scholar]
  29. Tan J, Leung W, Moradian-Oldak J, Zeichner-David M, & Fincham AG (1998). The pH dependent amelogenin solubility and its biological significance. Connective Tissue Research, 38(1–4), 215–221; discussion 241–216. [DOI] [PubMed] [Google Scholar]
  30. Tarasevich BJ, Philo JS, Maluf NK, Krueger S, Buchko GW, Lin G, & Shaw WJ (2015). The leucine-rich amelogenin protein (LRAP) is primarily monomeric and unstructured in physiological solution. Journal of Structural Biology, 190(1), 81–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Toyosawa S, O’HUigin C, Figueroa F, Tichy H, & Klein J (1998). Identification and characterization of amelogenin genes in monotremes, reptiles, and amphibians. Proceedings of the National Academy of Science USA, 95(22), 13056–13061. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Wright JT, Torain M, Long K, Seow K, Crawford P, Aldred MJ, … Hart TC (2011). Amelogenesis imperfecta: genotype-phenotype studies in 71 families. Cells Tissues Organs, 194, 279–283. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Wu L, Cen G, Yan Y, & DenBesten P (2003). X-linked amelogenesis imperfecta may result from decreased formation of tyrosine rich amelogenin peptide (TRAP). Archives of Oral Biology, 48(3), 177–183. [DOI] [PubMed] [Google Scholar]
  34. Yamakoshi Y (2011). Porcine amelogenin : alternative splicing, proteolytic processing, protein - protein interactions, and possible functions. Journal of Oral Biosciences, 53(3), 275–283. [PMC free article] [PubMed] [Google Scholar]
  35. Yang X, Sun Z, Ma R, Fan D, & Moradian-Oldak J (2011). Amelogenin “nanorods” formation during proteolysis by Mmp-20. Journal of Structural Biology, 176(2), 220–228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Zhang X, Ramirez BE, Liao X, & Diekwisch TGH (2011). Amelogenin supramolecular assembly in nanospheres defined by a complex helix-coil-PPII helix 3D-structure. PLos ONE, 6(10), e24952. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES