Mmp-20 and Klk4 Cleavage Site Preferences for Amelogenin Sequences

Abstract

Mmp-20 and Klk4 are the two key enamel proteases. Can both enzymes process amelogenin to generate the major cleavage products that accumulate during the secretory stage of amelogenesis? We isolated Mmp-20 and Klk4 from developing pig teeth and used them to digest the tyrosine-rich amelogenin polypeptide (TRAP), the leucine-rich amelogenin protein (LRAP), and 5 fluorescence peptides. We characterized the digestion products by LC-MSMS, SDS-PAGE, and C18 RP-HPLC monitored with fluorescence and UV detectors. Mmp-20 cleaves amelogenin sequences after Pro¹⁶², Ser¹⁴⁸, His⁶², Ala⁶³, and Trp⁴⁵. These cleavages generate all of the major cleavage products that accumulate in porcine secretory-stage enamel: the 23-kDa, 20-kDa, 13-kDa, 11-kDa, and 6-kDa (TRAP) amelogenins. Mmp-20 cleaves LRAP after Pro⁴⁵ and Pro⁴⁰, producing the two LRAP products previously identified in tooth extracts. Among these key cleavage sites, Klk4 was able to cleave only after His⁶². We propose that Mmp-20 alone processes amelogenin during the secretory stage.

Introduction

Because of the large size of their developing teeth and their availability from slaughterhouses, pigs have been the predominant model for the study of amelogenin, the most abundant enamel protein (Fincham et al., 1999). Porcine amelogenins and their cleavage products have been isolated and extensively characterized by protein sequencing (Fukae et al., 1979, 1980; Fukae and Shimizu, 1983; Yamakoshi et al., 1989, 1994, 2003). In the pig, there are separate amelogenin genes on the X and Y chromosomes, but these genes secrete proteins having identical amino acid sequences (Ikawa et al., 2005). As in other mammals, the complicated assortment of amelogenins in porcine secretory-stage enamel is due to alternative splicing (Hu et al., 1996) and to proteolysis (Bartlett and Simmer, 1999).

The major porcine amelogenin (P173; Met¹-Asp¹⁷³) has 173 amino acids and an apparent molecular weight of 25 kDa. Most of the secreted P173 is cleaved after Ser¹⁴⁸ to generate the 20-kDa amelogenin (P148; Met¹-Ser¹⁴⁸), which in turn is cleaved after Trp⁴⁵ to yield the tyrosine-rich amelogenin polypeptide (TRAP; P45; Met¹-Trp⁴⁵) and the 13-kDa amelogenin (Leu⁴⁶-Ser¹⁴⁸) (Fincham et al., 1981). The TRAP cleavage site is conserved in humans and other mammals (Fincham et al., 1989; Fincham and Moradian-Oldak, 1993). Point mutations in AMELX that alter the human amelogenin amino acid sequence near Trp⁴⁵ cause amelogenesis imperfecta (AI) and also alter the susceptibility of the protein to proteolysis (Li et al., 2001, 2003; Tanimoto et al., 2008).

Another important amelogenin isoform is the leucine-rich amelogenin protein (LRAP) (Fincham et al., 1981; Fincham and Moradian-Oldak, 1993). Cloning of the LRAP mRNA was the first evidence of amelogenin alternative splicing (Gibson et al., 1991). The LRAP mRNA is about 10% as abundant as the message for the major amelogenin isoform (Yuan et al., 1996). In the pig, LRAP (P56) is cleaved after Pro⁴⁵ or Pro⁴⁰, and these cleavage products accumulate in the matrix (Fincham and Moradian-Oldak, 1993).

The only known extracellular proteases in developing enamel are Mmp-20 (Bartlett et al., 1996) and kallikrein 4 (Klk4) (Simmer et al., 1998), and both of these enzymes are critical for proper dental enamel formation (Hart et al., 2004; Kim et al., 2005). Enamel proteases degrade enamel proteins, which facilitates their removal from the matrix, allowing for thicker individual crystallites and a higher degree of mineralization (Smith, 1998; Lu et al., 2008). Enamel proteases may also play a role in processing enamel proteins to generate products that serve separate functions than the parent proteins. In this study, we tested the ability of Mmp-20 and Klk4 to catalyze the cleavages that generate the P173 and LRAP cleavage products that accumulate in secretory-stage enamel.

Materials & Methods

All experimental procedures involving the use of animals were reviewed and approved by the University of Michigan Institutional Animal Care and Use Committee.

Preparation of Soft and Hard Enamel Samples

Tooth germs of permanent molars were surgically extracted from the maxillae and mandibles of six-month-old pigs at the Michigan State University Meat Laboratory (East Lansing, MI, USA). The enamel organ epithelia and dental pulp tissue were removed by means of tissue forceps. The soft, cheese-like enamel was separated from the crown with a spatula. We obtained hard enamel samples by scraping the remaining hard, chalky enamel.

Separation of Soft Enamel and Hard Enamel Extracts

We sequentially extracted soft or hard enamel scrapings to obtain neutral and alkaline extracts. The neutral extract (supernatant) was obtained by homogenization of the enamel samples in Sörensen buffer (made by mixing Na₂HPO₄ and KH₂PO₄ to achieve a final phosphate concentration of 50 mM and a pH of 7.4), followed by centrifugation. The pellet was re-suspended in carbonate buffer (made by mixing NaHCO₃ and Na₂CO₃ to a final concentration of 50 mM and a pH of 10.8), homogenized, and then centrifuged. The supernatant was designated as alkaline extract.

Purification of Porcine Klk4

The hard enamel-neutral extract was raised to 40% saturation by the addition of ammonium sulfate, then centrifuged. The supernatant was raised to 65% saturation and centrifuged. The 40-65% saturation pellet was re-suspended in 0.5 M acetic acid. Acid-insoluble material was pelleted by centrifugation. The acid-soluble supernatant was applied to a C18 RP-HPLC column (Discovery^® C18, 1.0 cm x 25 cm, Sigma-Aldrich/Supelco, Bellefonte, PA, USA) equilibrated with 0.05% trifluoroacetic acid (TFA) and eluted with a linear gradient (20-80%) of acetonitrile containing 0.05% TFA at a flow rate of 1.0 mL/min. The fraction containing Klk4 was buffer-exchanged with 50 mM Tris-HCl (pH 7.4) with the use of a regenerated cellulose membrane (Ultracel; NMWL: 3000; Millipore, Billerica, MA, USA).

Purification of Porcine Mmp-20

The soft enamel-alkaline extract was fractionated by ion exchange chromatography in a Q-Sepharose^TM Fast Flow column (1.6 cm x 20 cm, GE Healthcare Biosciences, Little Chalfont, UK) equilibrated with 50 mM Tris-HCl/6 M urea (pH 7.4) and eluted with a linear NaCl gradient (0-0.5 M). Mmp-20 eluted in the fourth peak was de-salted and concentrated by ultrafiltration (Ultracel; Millipore Corporation) and lyophilized. The lyophilized sample was dissolved with 20 mM sodium citrate/6 M urea (pH 7.4) and applied to a Heparin Sepharose^TM 6 Fast Flow column (1.6 cm x 20 cm, GE Health-care) equilibrated with the same buffer and eluted with a step NaCl gradient (0.2, 0.5, and 2 M at a flow rate of 0.2 mL/min). Mmp-20 eluted in the second peak, which was buffer-exchanged with 50 mM Tris-HCl (pH 7.4) with the use of a regenerated cellulose membrane (Ultracel; NMWL: 3000; Millipore).

Isolation of LRAP and TRAP and Recombinant Pig Amelogenin (rP172)

LRAP was commercially synthesized (YenZym, Burlingame, CA, USA). TRAP (in the soft enamel, alkaline extract) was fractionated by ion exchange chromatography (2nd peak) and then by size exclusion chromatography on a Superdex G200 column (1.6 cm x 100 cm, GE Healthcare) equilibrated with 50 mM Tris-HCl/4 M guanidine (pH 7.4). TRAP (again in the 2nd peak) was de-salted and concentrated with an ultrafiltration membrane and lyophilized. The LRAP and TRAP samples were further purified by C18 RP-HPLC. The expression and isolation of rP172 have been described previously (Ryu et al., 1999).

Digestion of LRAP and TRAP by Mmp-20 or Klk4

LRAP or TRAP (100 µg) was incubated in 200-µL reactions with 21 µg of Mmp-20 or 1.25 µg of Klk4 extract for 1, 2, 4, 8, 12, and 24 hrs at 37°C in 50 mM Tris-HCl/5 mM CaCl₂ (pH 7.4) and analyzed by SDS-PAGE. The 24-hour digests were analyzed by LC-MS/MS.

Digestion of FRET-Amelogenin Peptide by Mmp-20 or Klk4

Five porcine amelogenin fluorescence resonance energy transfer (FRET) peptides were synthesized and purified by JPT Peptide Technologies (Berlin, Germany). Each was labeled with 2-aminobenzoyl (Abz) and 2,4-dinitrophenyl (Dnp): Amel1 (Tyr³⁹-Ile⁵⁰) Abz-YEPMGGWLHHQI-KDnp; Amel2 (Gln⁵⁷-Ile⁷⁰) Abz-QTPQSHALQPHHHI-KDnp; Amel3 (Gln¹⁰¹-Gln¹¹³) Abz-QPNLPLPAQQPFQ-KDnp; Amel4 (Leu¹⁴³-Asp¹⁵⁵) Abz-LPPMFSMQSLLPD-KDnp; and Amel5 (Leu¹⁵³-Lys¹⁶⁶) Abz-LPDLPLEAWPATDK-KDnp. Each FRET peptide (100 µg) was incubated in 200-µL reactions with 21 µg of Mmp-20 or 1.25 µg of Klk4 extract in 50 mM Tris-HCl/10 mM CaCl₂ (pH 7.4) at 37°C for 24 or 48 hrs, and aliquots were analyzed by LC-MS/MS or fractionated by RP-HPLC and monitored with UV (230 nm) and fluorescence detectors (excitation λ 320 nm/emission λ 420 nm).

SDS-PAGE, Western Blotting, and Zymography

SDS-PAGE was performed with Novex^® 18% or 4-20% Tris-Glycine Gel (Invitrogen, Carlsbad, CA, USA). The gels were stained with SimplyBlue^TM SafeStain (Invitrogen), and apparent molecular weights estimated by comparison with SeeBlue^® Plus2 PreStained Standards (Invitrogen). Duplicate SDS-PAGE gels were transblotted onto nitrocellulose membranes (Hybond^TM-P; 0.45 µm, GE Healthcare) and immunostained with a polyclonal antibody against recombinant pig Klk4 (Hu et al., 2000) or an anti-peptide antibody specific for the Mmp-20 hemopexin-like domain (Fukae et al., 1998). Blocking with 5% non-fat milk was followed by overnight incubation with a 1:10,000 dilution of antibody in Tween Tris-buffered saline (TTBS) containing 5% milk. The membranes were washed 3x in TTBS and incubated with anti-rabbit IgG secondary antibody (GE Healthcare) at a dilution of 1:10,000 for 1 hr. Immunostaining was performed by chemiluminescent detection (for Mmp-20) with the ECL Plus Western Blotting Detection System (GE Healthcare) or by colorimetric detection (for Klk4) with diaminobenzidine (DAB) (Sigma, St. Louis, MO, USA). Zymograms were performed on SDS polyacrylamide gels containing 10% gelatin or 12% casein. Reactions were performed in 50 mM Tris buffer/10 mM CaCl₂ (pH 7.4) at 37°C for 24 hrs.

Mass Spectrometry (LC-MSMS)

LC-MSMS was performed by Nextgen Sciences (Ann Arbor, MI, USA) on purified LRAP and TRAP, their digests, and on the digests of the 5 FRET peptides after 24 or 48 hrs.

Results

TRAP from developing pig teeth and synthetic LRAP were purified until they displayed a single chromatographic peak on a C18 RP-HPLC column and a single CBB-stained band on SDS-PAGE (Fig. 1A). LC-MSMS analysis confirmed the identities of the proteins and that both were phosphorylated (Fig. 1B). Klk4 and Mmp-20 were isolated from the hard and soft enamel, respectively. Klk4 was evident as two bands migrating above 30 kDa on SDS-PAGE, gelatin zymograms, and on Western blots (Fig. 1C) (Tanabe, 1984). We purified Mmp-20 taking advantage of its heparin-binding affinity. Intact Mmp-20 was evident on casein zymograms and on Western blots as a doublet at 45 and 41 kDa. An Mmp-20 catalytic domain was also evident on the zymograms (Fig. 1C). Both enzymes were able to process recombinant pig amelogenin into multiple cleavage products by 24 hrs, but the patterns of cleavage by Mmp-20 and Klk4 differed (Figs. 1D, 1E). Mmp-20 processed LRAP into slightly smaller products, but did not process TRAP (Fig. 1D). Klk4 degraded both LRAP and TRAP (Fig. 1E).

Figure 1.

Purification and characterization of porcine LRAP, TRAP, Klk4, and Mmp-20. (A) The leucine-rich amelogenin protein (LRAP) and the tyrosine-rich amelogenin peptide (TRAP) were purified until they represented single chromatographic peaks on a C18 RP-HPLC column and a single band on SDS-PAGE. The numbers under the chromatograms indicate time (in min). (B) Mass spectrometry of LRAP gave a value of 6618.77 Da (calculated mass, 6617.50), and TRAP gave a value of 5404.28 (calculated mass, 5404.08), which closely matched the calculated masses of the single phosphosphorylated proteins. (C) The final purification of Klk4 was on a C18 column, where Klk4 was collected from a well-defined chromatographic peak and appeared as a doublet above 30 kDa on SDS-PAGE stained with Coomassie Brilliant Blue (CBB), on a gelatin zymogram (Z), and on a Western blot (W). The final purification of Mmp-20 was on a heparin affinity column. Mmp-20 could not be detected by SDS-PAGE, but was evident on a casein zymogram (Z) and on a Western blot stained with an anti-peptide antibody raised against its C-terminal hemopexin-like domain. The Mmp-20 catalytic domain appears as lower bands on the zymogram, but is not detected by the antibody raised against the hemopexin-like domain.(D) Standard aliquots of Mmp-20 and (E) Klk4 were used to digest recombinant pig amelogenin (rP172), LRAP, and TRAP for 0, 1, 2, 4, 8, 12, and 24 hrs. The protein without enzyme was run as a control (C). Both enzymes digested rP172, but produced different patterns of digestion products. Mmp-20 showed little activity against LRAP, but was able to produce a slightly smaller product by 24 hrs. Klk4 degraded LRAP by 8 hrs and generated multiple products. Mmp-20 showed no appreciable activity against TRAP, while Klk4 degraded TRAP.

Single aliquots of Klk4 and Mmp-20 did not exhibit detectable chromatographic peaks at 230 nm (Appendix Fig. 2) and were used to digest 5 fluorescent peptides covering all of the major amelogenin (P173) cleavage sites that were previously identified by characterization of amelogenin cleavage products isolated from porcine secretory-stage enamel (Fukae and Shimizu, 1983; Yamakoshi et al., 1994, 2003). After 24 hrs, the digestions were characterized by LC-MSMS and by RP-HPLC monitored by UV and fluorescence detectors (Figs. 2, 3). Mmp-20 cleaved amelogenin sequences after Trp⁴⁵ (in Amel1) and Ser¹⁴⁸ (in Amel4). In vivo, these cleavages generate the most abundant amelogenin cleavage products in porcine secretory-stage enamel: the 20-kDa (Met¹-Ser¹⁴⁸), TRAP, and the 13-kDa (Leu⁴⁶-Ser¹⁴⁸) amelogenins. Klk4 was not able to catalyze either of these cleavages. Mmp-20 was also able to catalyze all of the cleavages that generate the less-abundant amelogenin cleavage products in secretory-stage enamel. Mmp-20 cleaved the Amel5 sequence after Pro¹⁶² that in vivo generates the 23-kDa (Met¹-Pro¹⁶²) amelogenin and cleaved the Amel2 sequence (in the 48-hour incubation) after His⁶² and Ala⁶³ that in vivo yields “extended TRAP” (Met¹-His⁶² and Met¹-Ala⁶³) and the 11-kDa (Ala⁶³-Ser¹⁴⁸ and Leu⁶⁴-Ser¹⁴⁸) amelogenin. Klk4 was not able to catalyze the cleavage of amelogenin sequences after Pro¹⁶² or Ala⁶³, but was able to catalyze the cleavage after His⁶². However, cleavage after both His⁶² and Ala⁶³ is used to generate the 11-kDa amelogenin in vivo (Appendix Fig. 1).

Figure 2.

Digestion of fluorescent peptides Amel1 (Tyr³⁹-Ile⁵⁰), Amel2 (Gln⁵⁷-Ile⁷⁰), Amel3 (Gln¹⁰¹-Gln¹¹³), Amel4 (Leu¹⁴³-Asp¹⁵⁵), and Amel5 (Leu¹⁵³-Lys¹⁶⁶). The numbers under the chromatograms indicate time (in min). The first row shows the UV (blue) and fluorescent (red) chromatograms for each uncut peptide. The second and third rows show each peptide after a 24-hour digestion with a standard aliquot of Klk4 or Mmp-20, respectively. Red arrows point to peaks containing a fluorescent (Abz-containing) cleavage product. Yellow arrows point to cleavage products that do not fluoresce (lack the Abz). Arrowheads point to minor fluorescent (Abz-containing) cleavage products that were not characterized in these 24-hour digests, but became more prominent in 48-hour digests (data not shown). Column 1 (Amel1) shows that Mmp-20 catalyzed the TRAP cleavage (after Trp⁴⁵), but Klk4 did not. Column 2 (Amel2) shows that Klk4 can cleave after His⁶². Mmp-20 cleaved Amel2 after His⁶⁸ near the end of the peptide. This cleavage has not been detected in enamel extracts (in vivo) and might be an artifact of the use of a peptide substrate. Mmp-20 generated multiple cleavages of Amel2 after 48 hrs (Appendix). Neither enzyme showed much activity for Amel3, which contained minor cleavage sites previously shown to be susceptible to cleavage by recombinant Mmp-20 (Ryu et al., 1999). Column 4 (Amel4) shows that Mmp-20, but not Klk4, catalyzes the cleavage after Ser¹⁴⁸ that generates the 20-kDa amelogenin. Column 5 (Amel5) shows that Mmp-20, but not Klk4, catalyzes the cleavage after Pro¹⁶² that generates the 23-kDa amelogenin.

Figure 3.

Sites cleaved by Klk4 and Mmp-20 in the 5 fluorescent peptides. Arrows indicate sites cleaved after 24 hrs. Arrowheads indicate sites in Amel2 cleaved by Mmp-20 in a 48-hour digest (after Gln⁶⁰, Ser⁶¹, His⁶², and Ala⁶³). The His⁶² and Ala⁶³ sites are used in a minor processing pathway that generates extended TRAP and the 11-kDa amelogenin (Appendix).

The cleavage of fluorescent peptides and characterization of the sites cleaved by mass spectrometry clearly demonstrated that Mmp-20 is capable of cleaving amelogenin sequences at all of the sites that are cleaved during the secretory stage to generate the main cleavage products of the major amelogenin isoform (P173). Because of its smaller size (56 amino acids), LRAP was synthesized in its entirety without fluorescent tags, digested with Mmp-20 and Klk4, and the sites cleaved were identified by LC-MSMS analysis. Mmp-20 cleaved LRAP after Pro⁴⁵ and Pro⁴⁰, whereas Klk4 did not cleave at these sites, but instead degraded LRAP by cleaving at other sites (Figs. 4A, 4B). Finally, Mmp-20 showed little to no activity against TRAP, whereas KLK4 degraded TRAP (Fig. 4C).

Figure 4.

Summary of the digestion of porcine LRAP, TRAP, and the major amelogenin isoform (P173) by Mmp-20 and Klk4. (A) The porcine LRAP sequence. Downward pointing arrows mark the two processing sites (after Pro⁴⁰ and Pro⁴⁵) catalyzed by Mmp-20. Upward pointing arrows are degradation sites catalyzed by Klk4, as determined by LC-MSMS analysis of a 24-hour digest of LRAP. (B) Diagram showing the processing of LRAP by Mmp-20. (C) The porcine TRAP sequence. Mmp-20 has very low catalytic activity against TRAP. Upward pointing arrows are degradation sites catalyzed by Klk4, as determined by LC-MSMS analysis of a 24-hour digest of TRAP. (D) Porcine 173-amino-acid major amelogenin (P173) sequence. Downward arrows show the major Mmp-20 processing sites; arrowheads show the minor Mmp-20 processing sites. (E) Diagram showing the major and minor processing sites that generate the most abundant amelogenin cleavage products in porcine secretory-stage enamel. The secreted protein (P173) is usually processed in two steps. The initial cleavage is after Ser¹⁴⁸ and generates P148, the most abundant amelogenin cleavage product in the matrix. The second cleavage is after Trp⁴⁵ and generates TRAP and the 13-kDa amelogenins. A less-used processing pathway involves an initial cleavage after Pro¹⁶², generating the 23-kDa amelogenin, which is cleaved again to generate P148. P148 is also cleaved in a less-often-used alternative pathway that generates extended TRAP and the 11-kDa amelogenin.

Discussion

In situ hybridization studies have demonstrated that Mmp-20 is expressed early in enamel formation, while Klk4 is not expressed until the transition stage (Bègue-Kirn et al., 1998; Hu et al., 2000, 2002; Simmer et al., 2004). Digestion of recombinant amelogenin (rP172; Pro²-Asp¹⁷³) with recombinant Mmp-20 yielded all of the major porcine amelogenin cleavage products that accumulate in secretory-stage enamel (Ryu et al., 1999), while digestion of rP172 by Klk4 produced an entirely different set of cleavages (Ryu et al., 2002). Here we demonstrated, through the digestion of fluorescent peptides, that Mmp-20 is able to digest amelogenin sequences at the exact sites that yield all of the major amelogenin cleavage products: the 23-kDa (Met¹-Pro¹⁶²), 20-kDa (Met¹-Ser¹⁴⁸), 13-kDa (Leu⁴⁶-Ser¹⁴⁸), 11-kDa (Ala⁶³-Ser¹⁴⁸/Leu⁶⁴-Ser¹⁴⁸) extended TRAP (Met¹-His⁶²/Met¹-Ala⁶³), and TRAP (Met¹-Trp⁴⁵). Klk4 cannot cleave amelogenin sequences at any of the sites that generate the major amelogenin cleavage products (except after His⁶²), but freely cleaves at many other sites. Only Mmp-20 processes LRAP after Pro⁴⁰ and Pro⁴⁵, and Mmp-20 does not degrade LRAP further. Bovine Mmp-20 also catalyzes the Pro⁴⁵ cleavage of LRAP (Li et al., 1999). Klk4, in contrast, degrades both TRAP and LRAP. Besides cleaving at a host of other sites, Klk4 efficiently cleaves phosphorylated TRAP and LRAP after Asn¹⁴ and Phe¹⁵, near the phosphorylated Ser¹⁶.

Our findings support the conclusion that, in developing pig teeth, Mmp-20 is the only significant proteolytic activity in the enamel extracellular space during the secretory stage. It processes amelogenins into a group of cleavage products that accumulate and are only slowly degraded further by Mmp-20. The tyrosine- and leucine-rich amelogenin peptides accumulate because Mmp-20 has little activity to degrade them. Klk4 cannot be active during the secretory stage, since it would prevent their accumulation. The widening of secretory-stage enamel crystals with depth (Daculsi and Kerebel, 1978) is likely due to the slow degradation of amelogenins by Mmp-20. The importance of the processing function of Mmp-20 is suggested by the loss of prism structure and failure of the enamel layer to achieve full thickness in the Mmp20 null mice, while its degradative function is evident by the retention of enamel proteins and the tendency of the enamel to chip off in these mice (Caterina et al., 2002). Klk4, in contrast, does not contribute to the processing of amelogenin, but is dedicated to its degradation during the transition and maturation stages.