Amelogenin-Cytokeratin 14 Interaction in Ameloblasts duringEnamel Formation*

Received for publication, May 22, 2001, and in revised form, June 18, 2001Published, JBC Papers in Press, June 25, 2001, DOI 10.1074/jbc.M104656200

Rajeswari M. H. Ravindranath‡, Wai-Yin Tam, Pablo Bringas, Jr., Valentino Santos,and Alan G. Fincham

From the Center for Craniofacial Molecular Biology, School of Dentistry, University of Southern California,Los Angeles, California 90033

The enamel protein amelogenin binds to the GlcNAc-mimicking peptide (GMp) (Ravindranath, R. M. H.,Tam, W., Nguyen, P., and Fincham, A. G. (2000) J. Biol.Chem. 275, 39654–39661). The GMp motif is found in theN-terminal region of CK14, a differentiation markerfor ameloblasts. The binding affinity of CK14 andamelogenin was confirmed by dosimetric binding ofCK14 to recombinant amelogenin (rM179), and to thetyrosine-rich amelogenin polypeptide. The specificbinding site for CK14 was identified in the amelogenintrityrosyl motif peptide (ATMP) of tyrosine-richamelogenin polypeptide and specific interaction be-tween CK14 and [3H]ATMP was confirmed by Scat-chard analysis. Blocking rM179 with GlcNAc, GMp, orCK14 with ATMP abrogates the CK14-amelogenin in-teraction. CK14 failed to bind to ATMP when the thirdproline was substituted with threonine, as in somecases of human X-linked amelogenesis imperfecta orwhen tyrosyl residues were substituted with phenyl-alanine. Morphometry of developing teeth distin-guished three phases of enamel formation; growth ini-tiation phase (days 0–1), prolific growth phase (days1–7), and growth cessation phase (post-day 7). Confocalmicroscopy revealed co-assembly of CK14/amelogeninin the perinuclear region of ameloblasts on day 0, mi-gration of the co-assembled CK14/amelogenin to theapical region of the ameloblasts from day 1, reaching apeak on days 3–5, and a collapse of the co-assembly.Autoradiography with [3H]ATMP and [3H]GMp corrob-orated the dissociation of the co-assembly at theameloblast Tomes’ process. It is proposed that CK14play a chaperon role for nascent amelogenin polypep-tide during amelogenesis.

Dental enamel is formed within a protein matrix secreted byameloblast cells of ectodermal origin (1). Ameloblasts synthe-size several matrix proteins (2) and also cytokeratins 5 and 14(3–5). Ninety percent of the enamel matrix protein constitutesamelogenins (6–8). Amelogenins are tissue-specific, non-glyco-sylated proteins, rich in proline, glutamine, leucine, and histi-dine. Post-secretory processing of amelogenins involves a seriesof discrete steps including supramolecular self-assembly and

progressive proteolytic reduction in molecular size, facilitatingenamel biomineralization and maturation (9). Little is knownabout the presecretory and secretory stages of amelogenesis.Understanding of these events may shed light on the signifi-cance of the interaction of amelogenins with other ameloblastproteins, and the functional roles of the different domains ofamelogenin polypeptide structures, including the possible sig-nificance of the highly conserved phosphorylation locus at ser-ine 16, which remains enigmatic.

While seeking to define the functional role of different do-mains of amelogenins, we have observed that the conservedtri-tyrosyl motif (amelogenin tri-tyrosyl motif peptide(ATMP):1 PYPSYGYEPMGGW) of the N-terminal region ofamelogenins binds specifically to N-acetylglucosamine (Glc-NAc) of glycoconjugates (10). Furthermore, we have demon-strated that the ATMP also recognizes peptide mimics ofGlcNAc (11). Most importantly, one of the GlcNAc mimickingpeptides (GMp: SFGSGFGGGY) binds avidly to ATMP. Modi-fications of the ATMP motif, including substitution of proline 3by threonine as observed in a case of human X-linked amelo-genesis imperfecta (12), resulted in the loss of binding to bothGlcNAc and GMp (10, 11). Interestingly, the GMp sequence islocalized in the highly conserved N-terminal domain of cyto-keratins 14, 16, and 17. Since CK14 is a known marker forameloblasts in a developing tooth prior to synthesis ofamelogenins (3–5) and it contains the GMp that binds specifi-cally to the ATMP sequence of amelogenins, we hypothesizethat interactions between CK14 and amelogenins may play animportant role in amelogenesis, enamel development, anddisease.

In the present investigation, we demonstrate that CK14binds specifically to amelogenins through the ATMP. Further-more, we show that putative loss of function mutations ofATMP (e.g. substitution of a proline residue with threonine, asnoted above) abrogates binding of amelogenins to CK14. Usingconfocal laser microscopy, we demonstrate co-assembly ofamelogenin-CK14, its migration to the apical region of amelo-blasts, and subsequent dissociation at Tomes’ process. Ourfindings suggest that CK14 functions as a chaperon for nascentamelogenin polypeptides during amelogenesis.

* This work was supported by National Institutes for Health NIDRGrant DE-03660. The costs of publication of this article were defrayedin part by the payment of page charges. This article must therefore behereby marked “advertisement” in accordance with 18 U.S.C. Section1734 solely to indicate this fact.

‡ To whom correspondence should be addressed: Center for Cranio-facial Molecular Biology, School of Dentistry, University of SouthernCalifornia, 2250 Alcazar St., Los Angeles, CA 90033. Tel.: 323-442-3171; Fax: 323-442-2981; E-mail: rravindr@hsc.usc.edu.

1 The abbreviations used are: ATMP, amelogenin tri-tyrosyl motifpeptide; CK, cytokeratin; TRAP, tyrosine-rich amelogenin polypeptide;LRAP, leucine-rich amelogenin polypeptide; T-ATMP, where proline issubstituted by threonine; F-ATMP, where all three tyrosyl residues arereplaced by phenylalanine; HPLC, high performance liquid chromatog-raphy; NB, newborn; PN, post-natal; GlcNAc, N-acetyl-D-glucosamine;GMp, GlcNAc mimicking peptide; PVDF, polyvinylidene difluoride;HSA, human serum albumin; FITC, fluorescein isothiocyanate; TRITC,tetramethylrhodamin isothiocyanate; ELISA, enzyme-linked immu-nosorbent assay; PBS, phosphate-buffered saline.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 276, No. 39, Issue of September 28, pp. 36586–36597, 2001© 2001 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org36586

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from

EXPERIMENTAL PROCEDURES

Animals—Twenty-five normal, healthy, female Swiss Webster preg-nant mice (Charles River Breeding) were used to obtain a sufficientnumber of litters for this study. All protocols involving mice wereapproved by the Institutional Animal Care and Use Committee (LosAngeles, CA). Mandibles were obtained from Swiss Webster mice atdifferent developmental stages ranging from E18-day pregnant micethrough newborn (NB) day “0,” postnatal days (PN) 1, 3, 5, 7, and 9.Mandibular incisors were used for morphometric studies.

Amelogenin Proteins—The recombinant mouse amelogenin rM179(20.16 kDa) was prepared by expression in Escherichia coli, isolated,and purified by high performance liquid chromatography as previouslydescribed (13). The protein was further purified by analytical reversedphase HPLC and homogeneity was assessed by SDS-PAGE (10, 11). Theamino acid sequence of this protein (ATMP in bold) is as follows: rM179,PLPPHPGSPGYINLSYEVLTPLKWYQSMIRQPYPSYGYEPMGGW-LHHQIIPVLSQQHPPSHTLQPHHHLPVVPAQQPVAPQQPMMPVP-GHHSMTPTQHHQPNIPPSAQQPFQQPFQPQAIPPQSHQPMQP-QSPLHPMQPLAPQPPLPPLFSMQPLSPILPELPLEAWPATDKTKR-EEVD179

..The polypeptides used in this investigation include: TRAP (5.20

kDa), a synthetic polypeptide representing the N-terminal conserved 45amino acid residues of amelogenins; LRAP (6.82 kDa), a syntheticleucine-rich amelogenin polypeptide (a polypeptide sharing 33 aminoacid residues of the N terminus and 26 amino acid residues of the Cterminus of full-length amelogenins); ATMP (1.45 kDa) (see sequenceabove), a synthetic polypeptide, representing the 13 C-terminal resi-dues of the TRAP amelogenin and the two altered ATMP peptides;T-ATMP (PYPSYGYETMGGW) and F-ATMP (PFPSFGFEP-MGGW). P162 (23 kDa) and P148 (20 kDa) (Porcine amelogenins) wereisolated from unerupted mandibular pig molars and purified as previ-ously described (10).

Synthesis and Purification of Polypeptides—All the polypeptidesused (TRAP, LRAP, ATMP, T-ATMP, F-ATMP, and GMp (SFGSGF-GGGY located at the highly conserved N-terminal region of CK14) inthis study were synthesized by the University of Southern CaliforniaMicrochemical Core Laboratory using an Applied Biosystems model430A single column peptide synthesizer with the modified Merrifieldprocedure (14). Peptides were purified by reversed phase HPLC (C4–214TP54 column or C18–291HS54 column (Vydac, The SeparationsGroup, Hesperia, CA) with a gradient of 35–50% B in 60 min (Bcontained 60% (v/v) aqueous acetonitrile in 0.1% (v/v) trifluoroaceticacid) at a flow rate of 1.0 ml/min (10, 11).

Specificity of the Antibodies—The murine monoclonal antibody forCK14 (clone LL002) is affinity purified and specifically recognizes the14 C-terminal residues of CK14 (15). Anti-amelogenin antibody is de-veloped in rabbits after immunizing with recombinant amelogeninrM179 (13). The specificity of the antibody was assessed using differentfragments of amelogenin in ELISA (16–18). The polyclonal antibody (at1/6000) binds better to rM166 (a construct that lacks the hydrophilicC-terminal 13-residue segment), than to rM179 at the same concentra-tion (50 ng/well), suggesting that the C-terminal acidic residues are notessential for binding. The antibody did not recognize the N-terminalTRAP region in ELISA or in Western blots suggesting that it recognizesthe hydrophobic central core of the amelogenin.

3H Labeling of ATMP and GMp—The 13-residue ATMP(P[3H]YPSYGYEPMGGW) was prepared from tritium gas by NycomedAmersham Plc. (Amersham Pharmacia Biotech) and after labeling, thematerial co-chromatographed with the ATMP synthesized by the Uni-versity of Southern California Microchemical Core Laboratory. A massspectrum is consistent with the proposed structure. The material wassupplied as a water:ethanol (1:1, v/v) solution in a silanized borosilicatemultidose vial with additional screw cap under nitrogen. GMp(SFGSGFGGG[3H]Y) was also labeled as noted previously (11).Polypeptides were purified by HPLC on a Vydac C18 300-Å (protein orpeptide: 250 � 4.6 mm) column with a gradient of solution A (0.01 M

aqueous trifluoroacetic acid) and solution B (0.01 M trifluoroacetic acidin acetonitrile), 0–100% B over 30 min, at a flow rate of 1.0 ml/min. Thepeptide was supplied in an aqueous solution. The peptides were storedin the absence of light and air at �20 or 4 °C, respectively.

Binding of rM179 and TRAP to CK14 by Enzyme-linked Immunosor-bent Assay (ELISA)—ELISA was performed using CK14 (RDI ResearchDiagnostics, Inc.) or TRAP as antigen following the protocol previouslydescribed (16–18). Antigen coating was done by adding 100 �l of asolution containing varying amounts of CK14 (Research Diagnostics) inPBS (pH 7.2), TRAP in carbonate and bicarbonate buffer (pH 9.6) wasadded to wells (Falcon 3915, Fisher Scientific, Pitsburgh, PA) and

incubated at room temperature overnight. Wells were blocked with PBScontaining 1% HSA at 37 °C for 1 h. One-hundred microliters of aknown amount of rM179 (5 pmol/100 �l)/CK14 (10 pmol/100 �l) wasadded to wells and incubated for 1 h at 37 °C. After washing the platesfive times, primary antibody against the recombinant M179 protein (12)(at a dilution of 1:6000) or anti-CK14 affinity purified murine mono-clonal (1:1000) (LL002 Neomarkers) (15) was added and incubated for1 h at 37 °C and then incubated with the secondary antibody (goatanti-rabbit IgG; Jackson ImmunoResearch, West Grove, PA, rabbitanti-mouse IgG) for 1 h. After washing, substrate (o-phenylenediaminedihydrochloride; Life Technologies, Inc., Gaithersburg, MD) in citrate-phosphate buffer and hydrogen peroxidase) was added to the plates forcolor development. The enzymatic oxidation was arrested after 30 minin the dark, with 6 N H2SO4. The absorbence difference at 490–650 nmwas measured after automix in a Bio-Tek microplate reader (Bio-TekInstruments). The values were corrected for background (wells withoutantigen). BSA and LRAP were used as negative controls.

Assessment of Binding of CK14 with Amelogenins by Western BlotAnalysis—The proteins were resolved via SDS-PAGE using 12 or 15%resolving and 3.5% stacking gels (19) and electrotransfered to polyvi-nylidene difluoride (PVDF) membranes (Millipore Corp., Immunolon-PTransfer Membrane) at 100 mA for 1 h using a semidry transblotapparatus (Hoefer Scientific Instruments, San Francisco, CA) (11, 20).Protein transfer was assessed by staining the PVDF strips with 0.1%Fast Green (Sigma) in 40% methanol and 10% acetic acid, and the stripswere compared with Coomassie Blue-stained protein bands (11, 21).Replicas were treated with ligands (GlcNAc, GMp) after blocking themembrane with phosphate-buffered saline, 1% HSA for 1 h at 37 °C.The membranes were washed five times with phosphate-buffered salinecontaining 0.1% HSA (11, 22). After washing, the membranes wereoverlaid with CK14 alone or CK14 preincubated with ATMP for 1 h. Thestrips were washed (5 times) and immunostained with anti-CK14monoclonal antibody for CK14 (clone LL002, 1/1000).

Dosimetric Binding of CK14 to [3H]ATMP—100 �l of [3H]ATMP(30 � 104 dpm in Tris-buffered saline, pH 7.2) was added to 1.5-mlpolypropylene microcentrifuge tubes containing increasing amounts ofCK14 in Tris-buffered saline (pH 7.2) and the mixture was gentlyshaken every 20 min for 2 h at 37 °C. The proteins were precipitatedwith 1 ml of cold ethanol (200 proof; Gold Shield Chemical Co., Hay-ward, CA) at 4 °C for 20 min, centrifuged for 15 min at 12,000 � g, andthe supernatant was removed. The unbound [3H]ATMP was removedcompletely by repeated vortex mixing and washing four times withethanol. The final pellets were dissolved in 50 �l of 1 N NaOH, andbound radioactivity was measured 15 min after adding 4 ml of scintil-lation fluid (Amersham Pharmacia Biotech) in a �-counter, as describedelsewhere (10, 11).

Specific Binding of CK14 to [3H]ATMP as a Function of IncreasingConcentration of ATMP—The total binding of labeled ATMP to CK14(350 pmol) was determined using increasing concentrations of[3H]ATMP (2–600 pmol). The nonspecific binding of labeled ATMP wasdetermined in the presence of 40 nmol of unlabeled ATMP at 37 °C for2 h and was subtracted from the total binding to obtain the specificbinding. The specific binding was further analyzed by the Scatchardplot.

Loss of Function “Mutations” of ATMP Results in Loss of Binding toCK14 as Assessed by Western Blot Analysis—Recombinant M179 onSDS-PAGE were electrotransfered to PVDF membranes at 100 mA for1 h using a semidry transblot apparatus. Protein transfer was assessedas described in the legend to Fig. 3. After blocking the membrane with1% HSA in PBS for 1 h at 37 °C and washing, the membranes wereoverlaid with CK14 alone or CK14 preincubated (for 1 h) with ATMP orCK14 preincubated with T-ATMP or F-ATMP. The strips were immu-nostained with murine monoclonal antibody for CK14 (clone LL002,1/1000).

Morphological Analyses—Mouse mandibular molar tissues for im-munohistochemistry were fixed immediately in 10% neutral bufferedformalin for 12 h at 4 °C. The fixed tissues were embedded in paraffinand 6-�m saggital sections were mounted on Histostik-coated slides(Accurate Chemical and Scientific Corp., Westbury, NY). Tissues forenamel morphometric studies were fixed in formalin for 24 h, decalci-fied with 10% EDTA. Day “0” mandibles were decalcified for 2 h (PN),day 1 for 4 h, day 3 for 30 h, day 5 for 66 h, day 7 for 108 h, and day 9samples for 156 h, washed, and then processed for paraffin embedding.

Morphometry of Enamel during Tooth Development—The serialcross-sections (6 �M) obtained from the whole length of the mousemandibular incisor (PN 0, 1, 3, 5, 7, and 9) was taken as 100%. Thesections were stained with Mallory’s triple stain for 5 min and the widthof enamel was measured at the 40 and 60% level from the base of the

Amelogenins-CK14 Interaction 36587

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from

incisor using the software program Image-Pro Plus 4.0 (Media Cyber-netics, L. P.) on 9 sections (three incisors from three mice) per day.

Immunochemical Localization of CK14 in Ameloblasts—Saggital sec-tions (6-�m) of mouse mandibular molar of day 0 (NB), day 1, 3, 5, 7,and 9 (PN) were deparaffinized and immunostained. Endogenous per-oxidase activity was blocked with 3% H2O2. Tissue sections were treatedwith 1% HSA in PBS (pH 6.0) and then incubated with affinity purifiedmouse monoclonal antibody for CK14 (LL002, 1/500) at 37 °C for 1 h.Negative controls were performed by replacing the primary antibodywith Tris-buffered saline and also IgG isotypes. Biotinylated anti-mouse secondary antibody was added to sections, incubated for 30 minat room temperature, and rinsed (3 times). The sections were treatedwith streptavidin-peroxidase conjugate for 30 min at room temperatureand stained with hematoxylin:eosin. The images were photographed.

Co-localization of CK14 and Amelogenins with Confocal Laser Scan-ning Microscopy in Ameloblasts during Enamel Formation—To exam-ine the spatial distribution and co-localization of CK14 and amelogenin,saggital sections of mouse postnatal mandibular molars at differentdevelopmental stages were prepared as described earlier (23, 24). Thesections were deparaffinized, rehydrated, and endogenous peroxidaseactivity was blocked with 3% H2O2. The sections were stained withprimary antibody against CK14 and then incubated with fluoresceinisothiocyanate (FITC)-conjugated secondary antibody (goat anti-mouseIgG, Jackson Immuno Research, West Grove, PA), 1:40 dilution, for 30min at room temperature. The sections were sequentially stained withthe primary antibody against recombinant mouse amelogenin for 1 hand then incubated with the secondary antibody coupled with tetra-methylrhodamin isothiocyanate (TRITC) to goat anti-rabbit IgG (Jack-son ImmunoResearch, West Grove, pA), 1:40 dilution for 30 min at roomtemperature. Replacing the primary antibody with Tris-buffered salineand also IgG isotypes performed as negative controls. Sections werethen washed, mounted immediately with glycerol (95% glycerol with 5%phosphate buffer as mounting media), and examined with a Zeiss LaserScan Microscope (LSM 510) (Carl Zeiss, Oberkochen, Germany)equipped with a 514� argon and a 543� helium-neon laser.

To determine coassembly of CK14 and amelogenin in vivo, co-local-ization was studied on sections stained sequentially with FITC-conju-gated antibody for CK14 (green signal at 514) and TRITC-conjugatedantibody for amelogenin (red signal at 543) (25–28). The 514- and543-nm lines of an argon-helium-neon laser exited FITC and TRITC,respectively. Both lines pass through a dichroic mirror and a neutraldensity filter before passing through the cell. The emitted light reflectedfrom the sample and fluorescence was collected by an oil immersion lensand imaged on to a photomultiplier tube after passing through a con-focal aperture at an optical filter. The cells were scanned at 4 s/frameand averaged twice to reduce background noise. The laser power wasadjusted to ensure any photodynamic effect on the cells was negligible.

Lectin Histochemistry—Sections from mouse mandibular tooth or-gans (PN day 5) were treated with lectins before and after treatmentwith sialidase. The following lectins were used; Triticum vulgaris(wheat germ agglutinin) that binds to sialic acid and GlcNAc (29, 30)and Datura stramomium specific for GlcNAc (purchased from EY Lab-oratories) (31). Sections were treated with neuraminidase to rule outthe staining due to sialic acids. Clostridium perfringens neuraminidase(type X) (purchased from Sigma) was used in all the major experiments.The enzyme treatments were done as described elsewhere (10, 22).Briefly, the sections were overlaid with 200 milliunits of enzyme in 200�l of PBS/section and were incubated for 1 h at 37 °C. After washing,the sections were treated with lectin peroxidase-coupled wheat germagglutinin or biotinylated D. stramomium (1 mg/ml, 1:10).

Autoradiography of [3H]ATMP to CK14 and 3H-Labeled GMp toAmelogenins—Sections from mouse mandibular tooth organs (saggitalsections, 6 �m) were used to assess the localization of free CK14 or freeamelogenins in ameloblasts during enamel formation. Sections weretreated with 50 milliunits of N-acetylglucosaminidase (Sigma) (22) toassess the location of free CK14 in ameloblasts. The sections wereblocked with PBS (pH 6.0) with 1.0% HSA at 37 °C for 1 h and thenincubated with [3H]ATMP or 3H-labeled GMp for 2 h at 37 °C to identifyfree CK14 or amelogenins. After washing the slides 3 times with PBS(0.1% HSA, pH 6.0), the slides were dried for 30 min at room temper-ature. Each slide was dipped separately in emulsion fluid diluted inwater at 1:1 dilution (Kodak Autoradiography Emulsions, Type NTB2for � emitters, International Biotech Inc. Eastman Kodak Co., NewHaven, CT) in the dark for 1 min, dried, and stored at 4 °C. Theautoradiographs were developed in Kodak DEKTOL Developer (East-man Kodak Co., Rochester, NY) according to the manufacturers recom-mendations and counterstained with hematoxylin. Digital and phasecontrast microscopy was used to identify the grains on sections.

Quantitative Analysis of Images using LSM 510—The scatter dia-gram (co-localization) is created and analyzed for pixel distribution (27).All pixels having the same positions in both images are considered apair. Of every pair of pixels (P1, P2) from the two images, the brightnesslevel of pixel P1 is interpreted as the X coordinate, and that of pixel P2as the Y coordinate of the scatter diagram. Non-co-localized P1 and P2pixels and the background values were excluded to distinguish thepixels that are co-localized. The results were expressed as percentage ofco-localization of amelogenins and CK14 at different developmentalstages from day 0 to 9 from initiation, growth, and cessation (double-headed arrow) of enamel formation. The quantity of non-co-localized P1and P2 pixels representing free amelogenin and free CK14, respec-tively, are also plotted for comparison.

RESULTS

Recombinant amelogenin (rM179) is a 20.16-kDa polypep-tide with 179 amino acids. It differs from native amelogeninonly in the absence of the N-terminal methionine and thatserine at position 16 is not phosphorylated. The ATMP se-quence that interacts with CK14 is localized in the C-terminalregion of TRAP. It binds to GlcNAc as well as the GMp localizedin the N-terminal head region of CK14. The specific binding ofATMP-GMp was confirmed earlier (11).

rM179 and TRAP Interact with CK14—Solid matrix immu-noassay analysis of the interaction between CK14 andamelogenins showed dosimetric binding of recombinant M179to CK14 (Fig. 1A). Low (�0.15) but no dose-dependent bindingwas observed with BSA. Fig. 1B indicates a similar dose-dependent binding of CK14 to TRAP. The titration does notreach saturation unlike in Fig. 1A. No binding was observedwith LRAP.

CK14- and Amelogenin Interaction on Western Blot—Theresults (Fig. 2) show that CK14 binds to rM179 (lane 5) nativeporcine amelogenins (lanes 9 and 10) P148 and P162 and TRAP(lane 12). The binding of CK14 to amelogenin was abrogatedwhen the Western blot of rM179 was pretreated (overlaid andwashed) with GlcNAc or GMp (lanes 6 and 7 and Table I).Similarly when CK14 was preincubated with ATMP it failed tobind to rM179 (lane 8 and Table I), suggesting that the CK14-rM179 interaction involves ATMP binding with the GMp motifof CK14.

Specific Binding of [3H]ATMP to CK14—To select the opti-mal concentration of CK14 for assessing the specific interactionbetween CK14 and ATMP, the dosimetry of [3H]ATMP bindingto CK14 was determined (Fig. 3A). The purity and homogeneityof ATMP was assessed by reversed phase HPLC and a typicalprofile of the purified fraction is illustrated in the inset of Fig.3A. Fig. 3B shows the specific binding of [3H]ATMP to CK14 asa function of increasing concentration of ATMP. The nonspe-cific binding was measured with unlabeled ATMP andsubtracted from the total binding to obtain the specific ATMP-CK14 interaction. A Scatchard plot of the binding of [3H]ATMPto CK14 indicates that the peptide-binding site is homogenouswith respect to the association constant. The significance of theslope (p � 0.001) and r2 (0.96) are indicated in the Fig. 3B,inset.

ATMP but Not Mutations of ATMP Bind to CK14—The bind-ing of CK14 to rM179 was assessed in Western blot before andafter preincubating CK14 with ATMP or loss of function mu-tations of ATMP. In one of the two mutant ATMP peptides(T-ATMP), the third proline is substituted with threonine, andin the other (F-ATMP) all three tyrosyl residues are replaced byphenylalanine (the T-ATMP mutation has been found in somecases of human X-linked amelogenesis imperfecta). The bind-ing of CK14 to rM179 is abrogated after pretreatment withATMP but not after pretreatment of CK14 with T-ATMP orF-ATMP suggesting that mutated ATMP is not capable ofbinding to CK14 as does ATMP (Fig. 4).

Amelogenins-CK14 Interaction36588

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from

Morphometry and CK14 Distribution in Ameloblasts duringEnamel Formation—To determine CK14-amelogenin interac-tion in vivo during development, morphometry of enamelgrowth and the distribution and migration of CK14 was stud-ied. Enamel growth was quantified by measuring the width ofenamel in cross-sections of mouse incisors from day 0 to 9. Theheight of the tooth is distinguished into four zones fromthe base of the tooth as 20, 40, 60, and 80%. Fig. 5 shows thatthe width of enamel increases from day 1 to 5 at 60% level. Itreaches a maximum on day 5 without much change thereafteron days 7 and 9. A similar pattern was observed at the 40%level. The enamel width steadily increased until day 7 and isconstant thereafter. These morphometric observations indi-cated that tooth development is initiated on day 0 and prolificon day 3 and reaches a maximum between days 5 and 7 andceases to increase after day 7. During the corresponding period,immunolocalization of CK14 in ameloblasts was carried outusing a affinity purified monoclonal antibody specific for theC-terminal region of CK14 (15). Fig. 6 shows that CK14 is

distributed throughout the cytoplasm on day 0. The immuno-staining is increased in intensity at the apical region adjacentto the extracellular matrix on day 1. Staining is intense in theTomes’ process. CK14 immunostaining is maximum in theapical region on day 3. Such staining persisted until day 5 butis lost on day 7 and thereafter, when the enamel has reached itsmaximum thickness.

Co-localization of CK14 and Amelogenin—Three completeindependent experiments were done using confocal microscopy.One for CK14-FITC, another for amelogenin-TRITC, and athird for the combination of the two dyes. Fig. 7 illustratesthree fluorescent images of ameloblasts on day 0 with FITC(green) signal only (Fig. 7A) or TRITC (red) signal only (Fig. 7B)or with combination of signals (Fig. 7C). Coassembly of CK14and amelogenin is indicated by the yellow signal (Fig. 7C). Theyellow granules (pixel) are distributed all around the nucleuson day 0. Arrows indicate circular distribution or aggregation ofthe granules in the cytoplasm. Green signal (CK14 �/ameloge-nin �) is seen at the distal end of the ameloblasts (Fig. 7C). Fig.

FIG. 1. A, amelogenin (rM179) bindingto varying concentrations of CK14 asmeasured by a solid matrix immunoas-say. B, cytokeratin-14 (CK14) binding tovarying concentrations of TRAP by a solidmatrix immunoassay.

Amelogenins-CK14 Interaction 36589

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from

8 demonstrates co-localization of CK14 and amelogenin inameloblasts during different stages of enamel growth. Accumu-lation of co-localized CK14-amelogenin granules is observed inthe apical region of ameloblasts on day 1 (Fig. 8C). Presence ofyellow granules in the cytoplasm and increasing density of thegranules toward the apical region suggests migration of co-assembled CK14 and amelogenin from the perinuclear regionto the apical end of the cell. The sections showed a distinctgreen signal in the cytoplasm as well as in the distal region ofameloblasts indicative of free CK14 on day 1. Fig. 8D shows

co-localization of CK14 and amelogenin on day 3. The accumu-lation of co-localized CK14-amelogenin in the apical region ofameloblasts is very prominent (yellow arrow). However, greensignals were also observed at the apical end (green arrows).Further magnification (inset in Fig. 8D) revealed a mixture ofsignals in the apical region of the ameloblasts (Fig. 8E). Thesignals were predominantly yellow (yellow arrow). Green sig-nals (dark green or green arrows) are observed toward theinterior of the apical zone of ameloblast. Red signals (red ar-rows) are abundant toward the exterior of the apical zone.While the yellow signal showed the co-localization of CK14 andamelogenin, the red and green signals signify the presence of(non-co-localized) free CK14 and amelogenin in the apical zone.Free amelogenin (red signal) moves toward exterior while freeCK14 (green signal) remains or moves toward interior of theameloblasts. Fig. 8, F, G, and H, represent co-localization andsecretion of amelogenin on days 5, 7, and 9, respectively. Onday 9 (Fig. 8H) amelogenin forms a distinct layer at the form-ative zone of enamel. The yellow granules in the cytoplasm(yellow arrow) are very few. Red and green granules are alsofound in the disintegrating ameloblasts at days 7 and 9 (Fig. 8,G and H).

Collapse of Co-assembled CK14 and Amelogenin duringAmelogenesis—Fig. 8E (magnification of inset in 8D) providesevidence to suggest collapse of coassembled CK14 and ameloge-nin after accumulation at the apical region of ameloblasts byday 3. To verify whether such dissociation occurs at the apicalzone, free CK14 and amelogenin were localized with [3H]ATMPand [3H]GMp. The rational for this investigation is that thecoassembly of CK14 and amelogenin involves binding of ATMP(PYPSYGYEPMGGW) with the GMp of CK14. Dissociationwould expose these peptide sequences in CK14 and ameloge-nin, which can be monitored using [3H]ATMP and [3H]GMp.[3H]ATMP may bind to both GlcNAc and GMp of CK14. Thepresence of GlcNAc is identified with D. stromonium lectin and

TABLE IAmelogenin-CK14 interaction in Western Blots in the presence and

absence of specific inhibitorsRecombinant M179 on SDS-PAGE were electrotransfered to PVDF

membranes at 100 mA for 1 h using a semidry transblot apparatus.After assessment of protein transfer, the replicas were treated withGlcNAc or GMp1, after blocking the membrane with 1% HSA in PBS for1 h at 37 °C. After washing, the membranes were overlaid with CK14alone or CK14 preincubated (for 1 h) with ATMP or T-ATMP or F-ATMP. The strips were immunostained with affinity purified, murinemonoclonal antibody for CK14 (clone LL002, L1000).

Fractions on Westernblots

Inhibitors ofamelogenin

CK14 with or withoutinhibitors Resultsa

CK14 None ���rM179 None None �rM179 None CK14 ����rM179 GlcNAc CK14 �rM179 GMp1 CK14 �rM179 None CK14 (with ATMP) �rM179 None CK14 (with T-ATMP) ���rM179 None CK14 (with F-ATMP) ���TRAP None CK14 ����LRAP None CK14 �BSA None CK14 �Native porcine

amelogeninsNone CK14 ����

a �, negative; ���, moderate intensity; ����, high intensity.

FIG. 2. Western blot analysis of binding of CK14 with rM179 and TRAP in the presence and absence of inhibitors. RecombinantM179, TRAP, and BSA (negative control) on SDS-PAGE were electrotransfered to PVDF membranes at 100 mA for 1 h using a semidry transblotapparatus. Lanes 1–10, transblot from 12% gel. Lane 1, standards; Lane 2, CK14 plus anti-CK and 2nd antibody; Lane 3, CK14 plus 2nd antibodyonly; Lane 4, immunostaining rM179 without CK14 with anti-CK14 antibody and second antibody, no staining observed; Lane 5, immunostainingof CK14 bound to rM179; Lane 6, immunostaining of CK14 added to rM179 after overlay with GlcNAc, no staining observed; Lane 7, immuno-staining of CK14 added to rM179 after overlay with GMp, no staining observed; Lane 8, immunostaining of CK14 after rM179 overlaid with CK14preincubated with ATMP, no staining observed; Lane 9, immunostaining of CK14 on native (porcine) amelogenin P148; Lane 10, immunostainingof CK14 on native (porcine) P162; Lanes 11–13, transblot from 15% gel; Lane 11, standards; Lane 12, immunostaining of CK14 bound to TRAP;Lane 13, immunostaining of CK14 on BSA (negative control), no staining observed.

Amelogenins-CK14 Interaction36590

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from

FIG. 3. A, [3H]ATMP binding to varying concentrations of CK14. The mean values of duplicate analyses at each concentration were plotted. Theposition of 350 pmol of CK14, selected to study the specific binding of CK14 with ATMP is indicated on the graph as a vertical dotted line. BSAis used as a negative control. Inset provides the profile of synthetic ATMP, with the amino acid sequence “PYPSYGYEPMGGW” purified byC18-reversed phase HPLC. B, specific binding of [3H]ATMP to CK14 as a function of increasing concentration of ATMP-Scatchard plot analysis.The Scatchard analysis of this specific binding is shown as an inset. Each point represents the mean of duplicate determinations. r2 and p valuesof the slope are indicated in the graph.

Amelogenins-CK14 Interaction 36591

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from

wheat germ agglutinin (data not shown). The staining indi-cated diffused distribution of GlcNAc containing glycoconju-gates in the cytoplasm of ameoloblasts toward the distal zone.To eliminate the interaction of GlcNAc-glycoconjugates with[3H]ATMP, autoradiography was performed after removing theGlcNAc residues with N-acetylglucosaminidase (Fig. 9). Thefigure shows the localization of free CK14 with [3H]ATMP atdays 0, 1, and 3. [3H]ATMP was observed throughout thecytoplasm at all days. On day 0, the free CK14 is abundant inthe cytoplasm (Fig. 9A). Accumulation of free CK14 in theapical end is observed on day 1 (Fig. 9, B and C). Concomitantwith an increase in enamel growth on day 3, free CK14 accu-mulated in the apical region of ameloblasts, as evidenced by theincreased accumulation of [3H]ATMP. Fig. 10, C and D, indi-cate accumulation of [3H]ATMP within the Tomes’ process onday 5. In contrast, [3H]GMp failed to show such accumulationin the apical region of the cytoplasm either on day 5 or 3 (Fig.10, A and E). However, magnification of the apical region of theameloblasts on day 5 revealed accumulation of [3H]GMp withinthe Tomes’ process and at the exterior of the ameloblasts (Fig.10B). These observations suggest that the coassembled CK14-amelogenin may dissociate in the Tomes’ process. The freeamelogenin moves out of ameloblasts, whereas free CK14 mayremain within the ameloblasts.

Quantitation of Co-localized and Free Amelogenin andCK14—To evaluate co-assembly and dissociation of CK14 andamelogenin the percentage of green, yellow, and red signals(pixels) were measured during the growth of enamel at differ-ent days of development of teeth (Fig. 11).

The level of free cytokeratin within ameloblast as evidencedby the green signal appeared constant throughout and re-mained �20%. Free amelogenin (red signal) in ameloblast in-creased on day 3. This finding suggests that the dissociation ofthe CK14-amelogenin assembly may commence on day 3. Thedecline and the steady-state level of free amelogenin (red sig-nals) between days 3 and 7 could be due to continuous and

prolific secretion of amelogenin. The quantitative variation ofco-assembled amelogenin-CK14 within ameloblasts showed apattern concomitant with enamel growth. The yellow granulesalthough abundant on day 0 tend to decline on day 3 concom-itant with increase in free amelogenin (red signals). The de-crease in the percent of pixels on day 3 could be due to theaccumulation of the granules. Increase in co-localized granuleson days 5 and 7 are consistent with enamel growth during thisperiod. It also suggests active synthesis and transport of nas-cent amelogenin polypeptides to the sites of secretion.

DISCUSSION

Properties of the N-terminal “Head” Region of CK14—Ex-pression of cytokeratin is specific for each epithelial cell typeand its state of differentiation (32). CK14, a member of thefamily of acidic type I cytokeratins, consists of a conserved roddomain with four �-helical regions separated by short non-helical linker sequences and flanked by non-helical globularsequences commonly referred to as head (N-terminal) and tail(C-terminal) domains (33). Self-assembly of CK14 leads to for-mation of intermediate filaments of 10 nm diameter. The self-assembly is facilitated by the �-helical region (33). Removal ofthe head domain does not affect self-assembly (34) suggestingthat the head region may have a role different from that ofCK14 self-assembly, which is necessary for formation of thecytoskeleton or intermediate filaments. The presence of theGMp sequence in the head region of CK14 favors a role inbinding to GlcNAc-binding lectins. Indeed, the ATMP motif ofthe N-terminal domain of amelogenins, that specifically bindsto GlcNAc residues (10), also recognizes the GMp sequence(11). The binding specificity of the ATMP sequence(PYPSYGYEPMGGW) to the GMp motif (SFGSGFGGGY) wasconfirmed by dosimetric binding of amelogenin or TRAP with[3H]GMp, specific binding in varying concentrations of labeledGMp, Scatchard plot analysis, and competitive inhibition withunlabeled GMp (11). These findings lend support to the bindingcapability of the ATMP motif of amelogenins with the GMpsequence in the N-terminal head region of CK14.

Specificity of Amelogenin-CK14 Interaction and the Biologi-cal Significance—The dosimetric interaction between CK14and amelogenin (Fig. 1A), CK14 and TRAP (Fig. 1B), and CK14and the ATMP motif (Fig. 2) further favor the sequence-sequence interaction between the two proteins. Inability ofamelogenin to bind to CK14, if blocked by GlcNAc or GMpsequence and inhibition of CK14 binding to amelogenin byATMP (Table I), confirms the sequence-sequence interactionbetween the ATMP motif of amelogenins and the GMp motif ofCK14. The possible biological significance of amelogenin-CK14interaction is revealed by the loss of function mutations ofATMP (T-ATMP and F-ATMP) failing to interact with CK14.One of these mutations involves substitution of the third pro-line with threonine as has been reported in the inheritedenamel defect of human amelogenesis imperfecta (12). Simi-larly, substitution of tri-tyrosyl residues with phenylalanineaffected the binding of ATMP with CK14, suggesting that mu-tational alterations of the ATMP sequence that cause enameldefects could be due to loss of binding of the mutant amelogeninto CK14 and a consequent inability of the amelogenin to reachthe site of secretion. Thus, CK14 may play an important role inamelogenesis by interacting with the ATMP motif of ameloge-nin. The immunofluorescence studies made in this investiga-tion suggest a possible role of CK14 in amelogenesis.

CK14 and Amelogenin as Ameloblast-differentiation Mark-ers—While amelogenin is the postnatal differentiation markerof ameloblasts, the presence of CK14 in ameloblasts 48–96 hearlier than the expression of amelogenin indicates that CK14is a stage-specific differentiation marker for ameloblasts (3).

FIG. 4. Loss of function mutations of ATMP results in loss ofbinding to CK14 as assessed by Western blot analysis. Lanes 1–7,transblot from 12% gel. Lane 1, standards; Lane 2, rM179 stained withFast Green; Lanes 3–7 were immunostained with affinity purified,murine monoclonal antibody for CK14 (clone LL002, 1/1000). Lane 3,CK14 after immunostaining. Lane 4, rM179 overlaid with CK14 alonestained. Lane 5, CK14 preincubated (for 1 h) with ATMP did not stain.Lane 6, CK14 preincubated with T-ATMP stained. Lane 7, CK14 pre-incubated with F-ATMP, stained.

Amelogenins-CK14 Interaction36592

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from

Morphometric measurements of developing enamel define thestage-specific secretion of the enamel matrix by ameloblasts.Three phases of enamel formation can be distinguished atdifferent levels of a growing murine incisor tooth. Enamelsecretion commences (stage I) on day 0 and remains slow untilday 1. Phase II commences after day 1 when the enamel growthis prolific by day 3 and reaches a maximum between days 5 or7 (Fig. 5). In phase III enamel growth ceases and ameloblastslose their distinctive morphology. Correlated with sequentialdevelopment of enamel formation, a change in the pattern ofdistribution of CK14 is observed in ameloblasts. While CK14immunostaining is highly prevalent in the perinuclear region,

the migration of CK14 to the apical region of the cell from day1 is evident in both light and confocal microscopy. Intensefluorescent green signal, at dual wavelengths in confocal mi-croscopy, at the distal region of ameloblasts provides evidenceof self-assembly of CK14. Amelogenin is also found around theperinuclear region of ameloblasts on days 1 to 3, but ameloge-nin is always co-assembled with CK14.

Co-assembly of Amelogenin and CK14 during Enamel For-mation—Co-localization of CK14-amelogenin was evident bythe yellow signal or pixels in confocal microscopy (Fig. 7).Absence of red signals at dual wavelengths on day 0 in thecytoplasm suggests either the absence of free amelogenin or

FIG. 5. Morphometry of enamel dur-ing tooth development. The serialcross-sections (6 �m) obtained from thewhole length of mouse mandibular incisor(PN 0, 1, 3, 5, 7, and 9) was taken as100%. The sections were stained withMallory’s triple stain and the width ofenamel was measured at the 40 and 60%level from the base of the incisor using thesoftware program Image-Pro Plus 4.0 onnine sections (three incisors from threemice) per day. Mean (n � 9) and standarddeviation (vertical bar) are indicated.

FIG. 6. Distribution and migration of CK14 in ameloblasts during different stages of enamel formation. Saggital sections (6 �m) ofmouse mandibular molar of day 0 (newborn), days 1, 3, 5, 7, and 9 (post-natal) were immunostained. A, negative control day 0 treated with IgGisotype shows no immunostaining; B, negative control on day 1 treated with IgG isotype shows no immunostaining; C, negative control, day 7treated with IgG isotype shows no immunostaining. D, day 0, horizontal arrows point out the distribution of CK14 throughout the cytoplasm ofameloblasts (am). Vertical arrows show increased staining observed toward the periphery. E, day 1. Horizontal arrows show CK14 stainingintensity at the apical region adjacent to the extracellular matrix. Long and vertical arrows indicate Tomes’ processes. Note that enamel secretionhas commenced. F, day 3. CK14 immunostaining is maximum at the apical region (horizontal arrows). Enamel secretion is evident. G, day 5 CK14immunostaining is still intense at the apical region (vertical arrows). Enamel matrix is thickened. H, day 7 CK14 immunostaining declined(arrows). The enamel has reached maximum thickness. Horizontal bar refers to measurement. A, 20 �m; B, 20 �m; D, 20 �m; E, 30 �m; F, 30 �m.

Amelogenins-CK14 Interaction 36593

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from

masking of amelogenin by CK14. The perinuclear distributionof yellow granules in the cytoplasm of ameloblasts on day 0indicates a co-localization of CK14 and amelogenin. Progres-

sive accumulation of the same granules at the apical region ofameloblasts from days 1 to 3 suggests co-migration of theCK14-amelogenin complex from the perinuclear region to the

FIG. 7. Co-localization of CK14 and amelogenin in ameloblasts on day 0 as revealed by Lazer scan double-label confocal fluores-cence microscopy. Saggital sections (6 �m) were sequentially stained and examined under confocal laser scan microscopy. A, green (FITC) signal(514) refers to CK14 immunostaining. Note the distribution of the granules throughout the cytoplasm of the ameloblasts. Arrows indicate thedistribution of the cytoplasmic granules around the nucleus. B, red (TRITC) signal (543) refers to amelogenin immunostaining. Note thedistribution of amelogenins in the cytoplasm. C, co-localization of CK14 and amelogenin near the nucleus (n) of ameloblasts is shown by a hybridyellow signal (CK14�/amelogenin�). Horizontal bar refers to measurement. A, 30 �M; B, 30 �M; C, 30 �M.

FIG. 8. Co-localization, migration and dissociation of CK14 and amelogenin in ameloblasts during enamel development. Saggitalsections of mouse postnatal mandibular molars at days 1, 3, 5, 7, and 9 were deparaffinized, rehydrated, and endogenous peroxidase activity wasblocked with 3% H2O2. The sections were stained and examined as described in the legend to Fig. 9. Green (FITC) signal (514) refers to CK14immunostaining (CK14�/amelogenin�). Red (TRITC) signal (543) refers to amelogenin immunostaining (CK14�/amelogenin�) co-localization ofCK14 and amelogenin is shown by a hybrid yellow signal (CK14�/amelogenin�). A-C, day 1. Note the distribution and co-localization of CK14 andamelogenin toward the secretary end of the ameloblast (am). There is a clear and distinct accumulation of CK14�/amelogenin� pixels at the apicalend. Note also the presence of CK14�/amelogenin� as well as CK14�/amelogenin� pixels in the cytoplasm. CK14�/amelogenin� pixels arerestricted to the site of enamel secretion. D, day 3, accumulation of CK14�/amelogenin� pixels at the apical end. Note the presence ofCK14�/amelogenin� pixels (green arrow) in the apical region. CK14�/amelogenin� pixels appear as a layer at the site of enamel (e) secretion. E,day 3, white rectangular insert seen in D is magnified to show co-localization as well as dissociation of CK14 and amelogenin at the zone of enamelsecretion. Yellow arrow refers to pixel indicative of co-localization (CK14�/amelogenin�). Green arrows indicate to pixels with green signal(CK14�/amelogenin�). Red arrows indicate pixels with red signal (CK14�/amelogenin�). Note the presence of red signal toward at the site ofenamel formation. F, G, and H, distribution of green, yellow, and red signals on days 5, 7, and 9, respectively. Note the gradual disruption anddisintegration of ameloblasts on these days. H, red signal is seen as a distinct layer at the formative zone of enamel on day 9. Arrows in F, G, andH indicate the presence of free CK14. Am, ameloblast; d, dentine; e, enamel; n, nucleus. Horizontal bar refers to measurement. A, 20 �m; B, 20�m; D, 20 �m; E, 30 �m; F, 30 �m.

Amelogenins-CK14 Interaction36594

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from

apical region. Co-localization of CK14-amelogenin is promi-nently seen at the apical region until day 5.

Co-assembly of CK14-Amelogenin Is Transient—During theprolific phase (phase II) of enamel matrix secretion betweendays 3 and 5 (Fig. 8E), three distinct signals are observed at theapical zone. They are yellow signals indicative of co-assembly ofCK14-amelogenin, emigrating or red signals indicative of freeamelogenin, and retention or green signals indicative of freeCK14. The augmentation of the red signal outside the periph-ery of ameloblasts on day 7 as a distinct layer of red signal at

the interphase between the apical region of ameloblasts andenamel matrix indicates newly secreted extracellular ameloge-nin that is yet to undergo postsecretory modifications. All theseobservations suggest that amelogenin may be translocatedfrom the site of synthesis in the perinuclear region to the apicalregion of the cell in association with CK14, and at the apicalregion the dissociation of the co-assembly complex of CK14-amelogenin may occur, prior to amelogenin secretion.

Evidence for Dissociation of Co-assembled CK14-Ameloge-nin—While confocal microscopy indicated co-distribution and

FIG. 9. Binding of [3H]ATMP to free CK14 in N-acetylglucosaminidase-treated ameloblasts. Sections from mouse mandibular toothorgan (6 �m, saggital sections) were treated with 50 milliunits of GlcNAcase to assess the localization CK14 in ameloblasts during the enamelformation. A, day 0. Arrows indicate binding of ATMP to CK14 in the cytoplasm of ameloblasts. Note the uniform distribution of [3H]ATMP in thecytoplasm. B, day 1 early. Arrows indicate distribution of [3H]ATMP (site of CK14) at the apical end of the ameloblasts, the site of secretion ofenamel. C, day 1, late. Arrows indicate accumulation of [3H]ATMP (site of CK14) at the apical end of the ameloblasts. D, day 3. Arrows indicateintensity in the accumulation of [3H]ATMP at the apical end of the ameloblasts. Am, ameloblasts; e, enamel; d, dentine; pd, predentine; si, stratumintermedium.

FIG. 10. Binding of [3H]GMp to freeamelogenins or [3H]ATMP to freeCK14 at the Tomes’ process in amelo-blasts. A, B, and E, binding [3H]GMp tofree amelogenins. On day 5 (A) and on day3 (E) no accumulation of [3H]GMp is ob-served. Inset in A is magnified in B. Ar-rows indicate [3H]GMp, suggestive of freeamelogenin, in Tomes’ process of amelo-blasts. C and D, binding [3H]ATMP onday 5 in GlcNAcase-treated sections. Insetis magnified in D and the arrows indicate[3H]ATMP, suggestive of free-CK14, inTomes’ process of ameloblasts.

Amelogenins-CK14 Interaction 36595

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from

co-migration of CK14-amelogenin, autoradiographic studieswith [3H]ATMP provided information on the dissociation of theamelogenin-CK14 complex. Free CK14 is restricted within theameloblast as evidenced by the distribution of [3H]ATMP. Ac-cumulation of free CK14 at the apical region of ameloblastscommenced early on day 1 and increased on day 3 and re-mained constant until day 5. On the other hand, no prominentaccumulation of radiolabeled GMp could be seen in the apicalregion of ameloblasts. However, upon magnification, a specificaccumulation of [3H]GMp is observed in the apical cytoplasmicprojections (Tomes’ process) of the ameloblasts. Further exam-ination revealed that [3H]ATMP is also seen in the same re-gion, suggesting that these apical cellular projections of theameloblasts are a specialized zone of the cytoplasm facilitatingthe dissociation of CK14-amelogenin complex.

Transient Co-expression between Cytokeratins and OtherProteins—Similar to co-expression of CK14 and amelogenin,transient coexpression of cytokeratins and vimentin was ob-served during ontogenic development (4), in the cornea (4),fetal tongue epithelia (35), and in an adenomatoid odontogenictumor (36). Interestingly such co-expression of cytokeratin andvimentin in epithelial cells is considered a typical feature of aproliferative situation and the secretory functions of vimentin(4). Robinson and co-workers (37) envisaged a similar interac-tion between tuft proteins and keratins in ameloblasts duringenamel formation. Although none of these investigations doc-

FIG. 11. Quantitation of CK14-amelogenin paired pixels during different stages of development using confocal laser scan micros-copy. The quantitation is based on the analyses of the scatter diagram (co-localization) obtained with confocal laser scan microscopy. All pixelshaving the same positions in both images are considered a pair. Of every pair of pixels (P1, P2) from the two images, the brightness level of pixelP1 is interpreted as X coordinate, and that of pixel P2 as Y coordinate of the scatter diagram. Non-co-localized P1 and P2 pixels and the backgroundvalues were excluded to distinguish the pixels that are co-localized. The results were expressed as percentage of co-localization of amelogenins andCK14 at different developmental stages from day 0 to 9 from initiation, growth, and cessation (double-head arrows) of enamel formation. Thequantity of non-co-localized P1 and P2 pixels representing free amelogenin and free CK14, respectively, are also plotted for comparison.

FIG. 12. Scheme of the putative functional role for CK14 inamelogenesis. CK14 binds to the ATMP motif of the N-terminal regionof the nascent amelogenin polypeptide in the perinuclear region. ATMPmotif serves as a signal peptide. The GMp motif at the N-terminalregion of CK14 acts as ligand to bind to the signal peptide and form aco-assembly. Co-assembled CK14-amelogenin migrates to the Tomes’process, where collapse of co-assembly of CK14-amelogenin occurs re-sulting in release of amelogenin. Phosphorylation (P) at the serineresidue of amelogenin suggests that it may facilitate the dissociation ofamelogenin from CK14.

TABLE IIProteins associated with cytokeratins and intermediate filaments

Cytokeratin-associated proteins Ref.

Plectin 424344

Filaggrin 45Desmoplakins 46Loricin 47�-Internexin 48Nestin 49Desmocalin 50Synemin 51Paranemin 51

Amelogenins-CK14 Interaction36596

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from

umented sequence-sequence interaction between cytokeratinand vimentin or tuft proteins, it appears such coexpressions arein many cases not random but are intrinsic and cell typerelated. Several proteins have been identified to be associatedwith cytokeratins (38). A list of such cytokeratin-associatedproteins is presented in Table II. Most of these proteins areassociated with the C-terminal end of cytokeratins but no spe-cific sequence-sequence interaction was established.

Significance of CK14-Amelogenin Interaction—The observa-tions that amelogenin binds specifically to GlcNAc (10) led to astudy of its interaction with the GMp (11). The finding that thesequence of GMp has 100% homology with CK14 but not withany other known proteins or ameloblasts (present study) andthat CK14 is a differentiation marker of ameloblasts beforesynthesis of amelogenin (3) led to the serendipitous findingsreported in this investigation. The biological significance of theCK14-amelogenin association may not only be significant, inpromoting translocation of amelogenin to the site of secretionbut may also serve as a repressor of co-assembly and migrationof CK14-amelogenin and amelogenesis, as in the cases of X-linked inherited diseases like amelogenesis imperfecta. Thisfinding contrasts to our earlier assumption that mutations atthe ATMP site may affect post-secretory interactions ofamelogenin. Similarly, mutations at the GMp site of CK14 mayalso suppress the secretion of amelogenin. In support of thisconcept Rugg et al. (39) have reported the presence of dis-colored and notched front teeth in a child with a functional“knockout” of CK14, displaying clinical symptoms of epidermol-ysis bullosa. However, amelogenins mutated at sites other thanthe ATMP and non-GMp mutations of CK14 may not affect theco-assembly and migration of CK14-amelogenin.

In conclusion, our investigation appears unique in document-ing a specific sequence-sequence interaction between CK14 andamelogenins. Based on this interaction between amelogeninsand CK14 and their co-assembly, distribution, migration, anddissociation of the co-assembled proteins, we envisage that thefunctional role of CK14 may be similar to heat shock proteins(40) in binding to nascent peptides and carrying them to thecell surface. Our observations reveal that the ATMP motif atthe N-terminal region of amelogenins function as a signalpeptide for GMp-ligand of CK14 and CK14 may perform achaperon role during amelogenesis. A schematic for the puta-tive role of CK14 during amelogenesis is presented in Fig. 12.An additional point of note is that it has been shown thatcytokeratin-associated proteins may become dissociated follow-ing phosphorylation of the associated protein (38). In this con-text it is of interest that serine 16 is phosphorylated in thenative amelogenins extracted from the enamel, although evi-dence for the presence of a non-phosphorylated entity has alsobeen reported (41). It remains unclear whether amelogeninphosphorylation contributes to the dissociation of the ame-lognein-CK14 complex prior to amelogenin secretion. A study ofthe enzymes involved in serine phosphorylation in the apicalregion of ameloblasts, particularly in the Tomes’ process isunder investigation.

Acknowledgments—We thank Dr. Charles F. Shuler, Chairman,Center for Craniofacial Molecular Biology, University of Southern Cal-ifornia, for encouragement and support, and Ernesto Barron at the EMcore facility at Doheny EYE institute for assistance in confocal laserscanning microscopy.

REFERENCES

1. Eastoe, J. E. (1979) J. Dent. Res. 58, 753–7642. Fincham, A. G. (1998) Connect. Tissue Res. 38, 1–304

3. Tabata, M. J., Matsumura, T., Liu, J. G., Wakisaka, S., and Kurisu, K. (1996)Arch. Oral Biol. 41, 1019–1027

4. Kasper, M., Karsten, U., Stosiek, P., and Moll, R. (1989) Differentiation 40,207–214

5. Domingues, M. G., Jaeger, M. M., Araujo, V. C., and Araujo, N. S. (2000) Eur.J. Oral Pathol. 108, 43–47

6. Termine, J. D., Belcourt, A. B., Miyamoto, M. S., and Conn, K. M. (1980)J. Biol. Chem. 255, 9769–9772

7. Fincham, A. G., Belcourt, A. B., Termine, J. D., Butler, W. T., and Cothran,W. C. (1983) Biochem. J. 211, 149–154

8. Akita, H., Fukae, M., Shimoda, S., and Aoba, T. (1992) Arch. Oral Biol. 37,953–962

9. Fincham, A. G., Moradian-Oldak, J., and Simmer, J. P. (1999) J. Struct. Biol.126, 270–299

10. Ravindranath, R. M. H., Moradian-Oldak, J., and Fincham, A. G. (1999)J. Biol. Chem. 274, 2464–2471

11. Ravindranath, R. M. H., Tam, W., Nguyen, P., and Fincham, A. G. (2000)J. Biol. Chem. 275, 39654–39661

12. Collier, P. M., Sank, J. J., Rosenbloom, J., Yuan, Z. A., and Gibson, C. W.(1997) Arch. Oral Biol. 42, 235–242

13. Simmer, J. P., Lau, E. C., Hu, C. C., Aoba, T., Lacey, M., Nelson, D.,Zeichner-David, M., Snead, M. L., Slavkin, H. C., and Fincham, A. G. (1994)Calcif. Tissue Int. 54, 312–319

14. Merrifield, B. (1996) Science 232, 341–34715. Perkins, W., Campbell, I., Leigh, I. M., Mackie, R. M. J. (1992) J. Cutaneous.

Pathol. 19, 476–48216. Ravindranath, M. H., Ravindranath, R. M. H., Morton, D. L., and Graves,

M. C. J. (1994) Immunol. Methods 169, 257–27217. Ravindranath, R. M. H., Mondino, B. J., Adamu, S. A., Halabi, H. P., Hasan,

S. A., and Glasgow, B. J. (1995) Invest. Ophthalmol. Vis. Sci. 36, 2482–249118. Ravindranath, R. M. H., Hasan, S. A., and Mondino, B. J. (1997) Curr. Eye Res.

16, 1036–104319. Laemmli, U. K. (1970) Nature 227, 680–68520. Towbin, H., Staehelin, T., and Garden, J. (1979) Proc. Natl. Acad. Sci. U. S. A.

76, 4350–435421. Ravindranath, R. M. H., and Graves, M. C. (1992) Virology 188, 143–15122. Ravindranath, R. M. H., and Graves, M. C. (1990) J. Virol. 64, 5430–544023. Nakamura, M., Bringas, P., Jr., Nanci, A., Zeichner-David, M., Ashdown, B.,

and Slavkin, H. C. (1994) Anat. Record. 238, 383–39624. Zeichner-David, M., Vo, H., Tan, H., Diekwisch, T., Berman, B., Thiemann, F.,

Alcocer, M. D., Hsu, P., Wang, T., Eyna, J., Caton, J., Slavkin, H. C., andMacDougall, M. (1997) Int. J. Dev. Biol. 41, 27–38

25. Leenders, G. V., Dijkman, H., Hulsbergen-van de Kaa, C., Ruiter, D., andSchalken, J. (2000) Lab. Invest. 80, 1251–1258

26. Yu, W. H., Yu, S. C., Meng, Q., Brew, K., and Woessner, J. F., Jr. (2000) J. Biol.Chem. 275, 31226–31232

27. Lacoste, T. D., Michalet, X., Pinaud, F., Chemla, D. S., Alivisatos, A. P., andWeiss, S. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 9461–9466

28. Trivedi, N. S., Wang, H. W., Nieminen, A. L., Oleinick, N. L., and Izatt, J. A.(2001) Photochem. Pathobiol. 71, 634–650

29. Peters, B. P., Ebisu, S., Goldstein, I. J., and Flashner, M. (1979) Biochemistry.18, 5505–5511

30. Shibuya, N., Goldstein, I. J., Broekaert, W. F., Nsimba-Lubaki, M., Peters, B.,and Peumans, W. J. (1987) J. Biol. Chem. 262, 1596–1601

31. Yamashita, K., Totani, K., Ohkura, S., Takasaki, S., Goldstein, I. J., andKobata, A. (1987) J. Biol. Chem. 262, 1602–1607

32. Moll, R., Franke, W. W., and Schiller, D. D. (1982) Cell. 31, 11–2433. Fuchs, E., and Weber, K. (1994) Annu. Rev. Biochem. 63, 345–38234. Coulombe, P., and Fuchs, E. (1990) J. Cell Biol. 111, 153–16935. Van Muijen, G. N. P., Ruiter, D. J., Franke, W. W., Achtstatter, T., Haasnoot,

W. H. B., Ponec, M., and Warnaar, S. O. (1986) Exp. Cell. Res. 162, 97–11336. Tatemoto, Y., Tanaka, T., Okada, Y., and Mori, M. (1988) Virchows Arch. A

Pathol. Anat. 413, 341–34737. Robinson, C., Lowe, N. R., and Weatherell, J. A. (1977) Calcif. Tissue Res. 23,

19–3138. Foisner, R., and Wiche, G. (1991) Curr. Opin. Cell Biol. 3, 75–8139. Rugg, E. L., McLean, W. H. I., Lane, E. B., Pitera, R., McMillan, J. R.,

Dopping-Hepenstal, P. J., Navasaria, H. A., Leagh, I. M., and Eady, R. A. J.(1994) Genes Dev. 8, 2563–2573

40. Georgopoulos, C., and Welch, W. J. (1993) Annu. Rev. Cell Biol. 9, 601–63441. Fincham, A. G., Moradian-Oldak, J., and Sarte, P. E. (1994) Calif. Tissue Int.

55, 398–40042. Foisner, R., and Wiche, G. (1987) J. Mol. Biol. 198, 515–53143. Foisner, R., Leichtfried, F. E., Herrmann, H., Small, J. V., Lawson, D., and

Wiche, G. (1988) J. Cell Biol. 106, 723–73344. Wiche, G. (1989) Crit. Rev. Biochem. Mol. Biol. 24, 41–6745. Resing, K. A., Walsh, K. A., Haugen-Scofield, J., and Dale, B. A. (1989) J. Biol.

Chem. 264, 1837–184546. Green, K. J., Parry, D. A., Steinert, P. M., Virata, M. L., Wagner, R. M., Angst,

B. D., and Nilles, L. A. (1990) J. Biol. Chem. 265, 2603–261247. Mehrel, T., Hohl, D., Tothnagel, J. A., Longley, M. A., Bundman, D., Cheng, C.,

Lichti, U., Bisher, M. E., Steven, A. C., Steinert, P. M., and Yuspa, S. H.(1990) Cell 61, 1103–1112

48. Fliegner, K. H., Ching, G. Y., and Liem, R. K. (1990) EMBO J. 9, 749–75549. Lendahl, U., Zimmerman, L. B., and McKay, R. D. G. (1990) Cell 60, 585–59550. Cartand, A., Ludosky, M. A., Courvalin, J. C., and Carland, J. A. (1990) J. Cell

Biol. 111, 581–58851. Foisner, R., and Wiche, G. (1991) Curr. Opin. Cell Biol. 3, 75–81

Amelogenins-CK14 Interaction 36597

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from

Alan G. FinchamRajeswari M. H. Ravindranath, Wai-Yin Tam, Pablo Bringas, Jr., Valentino Santos andAmelogenin-Cytokeratin 14 Interaction in Ameloblasts during Enamel Formation

doi: 10.1074/jbc.M104656200 originally published online June 25, 20012001, 276:36586-36597.J. Biol. Chem. 

  10.1074/jbc.M104656200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/276/39/36586.full.html#ref-list-1

This article cites 51 references, 17 of which can be accessed free at

by guest on June 16, 2018http://w

ww

.jbc.org/D

ownloaded from