THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 264, No. 5, Issue of February 15, pp. 2840-2845, 1989 Printed in U.S.A.

Differences between Enamel-related and Cementum-related Dentin in the Rat Incisor with Special Emphasis on the Phosphoproteins*

(Received for publication, May 24, 1988)

Joop Steinfort, The0 van den Bos, and Wouter BeertsenS From the Experimental Oral Biology Group, Department of Periodontology, Academic Centre for Dentistry Amsterdam (ACTA) and the Laboratory of Cell Biology and Histology, Faculty of Medicine, University of Amsterdam, Amsterdam, The Netherlands

In order to determine whether qualitative and quan- titative differences exist between the non-collagenous proteins of crown and root dentin, rat incisors were separated into their enamel- and cementum-related dentin portions (ERD and CRD, respectively). Isolation of the mineral-bound proteins was performed under nondegradative conditions. Analytical procedures in- cluded DEAE-chromatography on high pressure liquid chromatography, sodium dodecyl sulfate-polyacryl- amide gel electrophoresis, and determination of phos- phate, protein, and hydroxyproline. The results have shown that considerable differences exist among the two dentins with respect to the quantity of the various phosphoproteins. For this group of proteins as a whole, the ERD contains about 2 times the amount of organic phosphate found in the CRD and about 1.4 times the amount of protein. The content of higher phosphoryl- ated phosphoproteins was about 4 times higher in the ERD than in the CRD, whereas the reverse was shown for the lower phosphorylated phosphoproteins. All dif- ferences were found to be statistically significant. So- dium dodecyl sulfate-polyacrylamide gel electropho- resis revealed that while the ERD contains phospho- proteins with an apparent molecular mass of 98 kDa, the CRD contains two classes of phosphoproteins one of 98 and one of 88 kDa. The relevance of the observed differences in phosphoprotein distribution is discussed in relation to their possible role in mineralization.

Coronal (enamel-related) and root (cementum-related) den- tin are usually not considered as distinct entities. Recently, however, we have shown that in the mouse incisor the enamel- and cementum-related dentin (ERD’ and CRD, respectively) react differently to 1-hydroxyethylidene-1,l-bisphosphonate (1). For the same animal it has been demonstrated that the conversion rate of predentin into dentin is considerably higher in the CRD than in the ERD (2).

As to the inorganic component of the dentin, differences have been reported between the ERD and the CRD of the rat incisor with respect to the Ca, S, and Mg contents and the

* Part of this work has been presented at the FECTS (Federation of European Connective Tissue Societies) meeting in Manchester (September 1986) and at the Continental European Division of IADR (International Association of Dental Research) in Regensburg, West Germany (September 1987). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should be addressed. The abbreviations used are: ERD, enamel-related dentin; CRD,

cementum-related dentin; SDS-PAGE, sodium dodecyl sulfate-poly- acrylamide gel electrophoresis; PP, phosphoproteins.

orientation of the mineral crystallites (3-5). It was the aim of the present study to determine whether

these differences are accompanied by differences in organic matrix composition. Particular attention was given to the phosphoproteins (PP).

EXPERIMENTAL PROCEDURES

Materials For dialysis Spectra/Por 1 dialysis tubing was used (molecular

mass cut off, 6-8 kDa). For DEAE-chromatography on high pressure liquid chromatography, a Beckman Gradient Liquid Chromatograph model 322, 112 Solvent Delivery Module, 412 Controler, and 160 Absorbance Detector (detection at 227 nm) were used in combination with a Bio-Gel TSK-DEAE-5-PW column (75 X 7.5 mm). Peak integration was performed with a Shimadzu C-R1B Chromatopak Integrator. SDS-PAGE was carried out using the PhastSystem (Phar- macia, Uppsala, Sweden). Low molecular weight markers were also from Pharmacia.

Analytical Methods

Colorimetric Methods Determination of hydroxyproline, general protein, organic phos-

phate, and DNA were performed according to methods previously described (6-9). Hydroxyproline and general protein were determined after preceding hydrolysis of samples in 6 N HCl at 135 “C for 3.5 h. Stainability of eluates with Alcian Blue was also tested (10).

D E A E Chromatography The TSK-DEAE column was eluted with a NaCl gradient in 4 M

urea, 50 mM Tris, pH 8.0. Optimal resolution was obtained using a flow rate of 0.15 ml/min and a gradient of 0-0.55 M NaCl in 1400 min. The column was loaded at a flow rate of 1 ml/min. After each run the column was washed with 1 M NaC1, but virtually no additional protein was detached. The eluates were scanned at 227 nm, tested for organic phosphate and Alcian Blue stainability, and run over SDS- PAGE. Peak area was expressed as the integrated absorbance detector output in Volt. seconds (V. s).

SDS-PAGE Samples treated with 5% P-mercaptoethanol were applied to 8-

25% gradient gels (Pharmacia) and stained with Coomassie Brilliant Blue and/or Alcian Blue. Coomassie staining was carried out accord- ing to the instructions of the manufacturer, except that the destaining temperature was kept at 27 “C in order to improve contrast. Alcian Blue staining and destaining were performed in 30% methanol, 10% acetic acid at 37 “C. In case gels were stained with Alcian Blue only, staining had to be preceded by two destaining steps in order to remove SDS. Before use, samples were dialyzed against 5 mM NHaHCOa and freeze-dried.

Analysis of the Organic Matrix of the ERD and CRD

Dissection and Preparation of Dentin Twenty female Wistar rats (about 3 months old; body weight +.

S.D., 151 +. 5 g) were distributed over four groups of five animals each. An older group comprising 120 animals, approximately 1 year

2840

Phosphoproteins in Dentin 284 1

of age (body weight f S.D., 244 & 29 g) was divided into six groups of 20 animals each. All 10 groups were processed separately.

The animals were anesthetized with ether and killed by intracardial injection of Nembutal. Their upper and lower incisors were dissected out and frozen in liquid nitrogen. For each animal this was carried out within less than 5 min.

Thawing of the incisors occurred at 4 'C in phosphate-buffered saline, pH 7.4, containing a mixture of proteinase inhibitors (13) (10 mM EDTA, 1 mM iodine acetate, 100 mM norleucine, 5 mM benzam- idine HCl, 1 mM phenylmethylsulfonyl fluoride, 1 mg/liter soybean trypsin inhibitor, and 5 mg/liter pepstatin). After removal of the pulp, periodontal ligament, cementum, enamel organ, and immature enamel the ERD and CRD (Fig. 1) were separated from each other by carefully cracking the teeth in a small bench vice. (The separation was estimated visually to be at least go%.) Pulp, immature enamel plus enamel organ, and cementum plus periodontal ligament were pooled separately. All procedures were performed at 4 "C.

Extraction of Dentin Organic Matrix Extraction of dentin was carried out by means of a two-step

extraction method as described previously (13, 14) and modified by Rahemtulla et al. (15). The dentin pieces were sonicated 2 X 1 min in a 50 mM Tris solution, pH 7.4, containing proteinase inhibitors (see above). The solution was then discarded and the dentin particles were extracted under constant stirring in a solution of 4 M guanidine HCl, 50 mM Tris, pH 7.4 (+ proteinase inhibitors). After 24 h the solution was centrifuged at 1,000 X g during 20 min and the pellets exhaustively dialyzed against a solution containing 0.4 M EDTA, 4 M guanidine HC1,50 mM Tris (+ proteinase inhibitors; pepstatin 1 mg/ liter). After decalcification the contents of the dialysis bags were centrifuged at 30,000 X g during 30 min and the pellets containing the dentin collagen fraction separated from the EDTA extract con- taining the mineral bound non-collagenous protein matrix of dentin. All procedures were performed at 4 "C.

Purification of Dentin Organic Matrix The EDTA-insoluble fractions were repeatedly washed with water

followed by centrifugation and then freeze-dried and stored until further use.

The guanidine extracts were exhaustively dialyzed against 50 mM Tris, pH 8.0, at 4 "C and then freeze-dried and stored until further use.

The EDTA extracts were dialyzed against 4 M urea, 50 mM Tris, pH 8.0, a t 4 "C and frozen until further use. Alternatively, portions of the EDTA extracts were dialyzed against 50 mM Tris, pH 8.0, for determination of general protein and hydroxyproline content.

Analysis of Dentin Organic Matrix Fractions The EDTA-insoluble fractions were digested with proteinase K and

the collagen content determined via hydroxyproline measurement. The guanidine extracts were analyzed by means of SDS-PAGE. In

addition the contents of general protein, DNA, and collagen were determined. Determination of phosphate was performed after recen- trifugation at 20,000 X g during 15 min.

The EDTA extracts were analyzed by means of SDS-PAGE and by determination of the contents of general protein, hydroxyproline, and

CEMENTUM \

ENAMEL

FIG. 1. Transverse section of rat incisor showing the en- amel- and cementum-related dentin portions (ERD and CRD, respectively).

organic phosphate. Further analysis was carried out by quantitative DEAE-chromatography followed by SDS-PAGE. Samples containing 3 mg of protein were taken from the six groups of older animals and applied to the column. Using a CaCL precipitation step (16), the EDTA extracts were divided into two fractions one being enriched in higher phosphorylated PP, the other in lower phosphorylated PP (17). To portions of the EDTA extract, which had been dialyzed against 4 M urea, CaC12 was added to a final concentration of 1 M. (Portions were taken in such a way that they all contained the same amount of organic phosphate.) After 12 h of constant stirring at 4 "C, the solutions were centrifuged at 25,000 X g during 15 min. The sediments were redissolved in 4 M urea containing 0.4 M EDTA. The CaClp-insoluble and -soluble fractions were then dialyzed against 4 M urea as mentioned before. The CaC12-precipitable and -soluble PP were analyzed by means of SDS-PAGE and determination of phos- phate content. Samples of the CaClz-insoluble and -soluble fractions of the ERD and CRD of all 10 groups of animals were further analyzed quantitatively by DEAE-chromatography followed by SDS-PAGE.

Comparison of the Non-collagenous Proteins of Dentin with Those of Dentin Surrounding Tissues

Pulp, periodontal ligament (plus cementum), immature enamel (plus enamel organ), and mature enamel were extracted in the same way as dentin. The first three tissues were collected when cleaning the teeth used in the above analyses.

Mature enamel was obtained from the incisors of an additional group of 20 female Wistar rats (150 g each). Using a dental bur which was kept slowly rotating in water at 4 "C, ground enamel was col- lected. The guanidine and EDTA extracts were dialyzed against 50 mM Tris, pH 7.4, containing proteinase inhibitors, freeze-dried and analyzed with SDS-PAGE.

Determination of the Collagen Content of ERD and CRD The upper and lower incisors of 20 Wistar rats (body weight about

180 g; age approximately 4 months) were dissected out and cleaned as described above. The enamel was ground away and the ERD and CRD were separated from each other by using a dental airotor cooled with tap water. The separated dentin portions were examined under a binocular, and remnants of enamel were taken away if necessary. After determination of the weight under water and the wet weight (in air), the volume of each portion of the dentin cylinder was calculated by subtraction of both weights, using the Archimedean principle. The dentin portions were then decalcified and hydrolyzed in 6 N HC1 at 135 'C during 3.5 h after which the hydroxyproline content was determined. The experiment was repeated once.

Statistical Method Statistical analysis was performed according to Student's t test (2a

= 0.05).

RESULTS

The Use of Collagen As an Internal Standard (EDTA Sedi- ment Fraction)-The amount of collagen/mg wet weight was about 17% higher in the older than in the younger rats (both for the ERD and the CRD: p < 0.05 andp < 0.01, respectively) and about 35% higher in the CRD than in the ERD fraction. The latter difference was probably due to the fact that the ERD did not only contain dentin but also enamel. In a separate experiment this explanation was verified by grinding away the enamel before determination of wet weight, volume, and collagen content. I t was found that whereas the amount of collagen/wet weight differed only slightly between the two dentin fractions the collagen content/volume appeared to be the same: 0.17 f 0.02 mg/mm3.

In the following paragraphs the amount of non-collagenous organic matrix substances was related to the amount of col- lagen present in the same dentin, thus using collagen as an internal standard.

Guanidine Extract-No statistically significant differences were found between ERD and CRD with respect t o the amount of soluble collagen (1% of total collagen) and general protein (25% as compared to the general protein content of

2842 Phosphoproteins in Dentin TABLE I

Relative amount of general protein and organic phosphate in the EDTA extract as function of the collngen content in the EDTA insoluble fraction

Values are means f S. D.

General protein/collagen Organic phosphate/collagen Organic phosphate/ general protein

flglg/mg Younger rats ( n = 4) ERD

CRD Older rats ( n = 6) ERD

CRD 33.4 19.6 f f 5.11 4.7 ] a 148 f 17 273 f 241 .]

a N.S. p < 0.005.

' p < 0.001. dp < 0.01.

m o r n o n n o n

r n O

0 0 0 0 I I I I l l I V V I 1 0 0 n n

m o n o 0 0

-

94 k

43 67

30

2 0 1 4 4 .. ~

a b c d e f g h i j k I m n

FIG. 2. SDS-PAGE (8-25% gradient gels). Gels were stained with Coomassie Brilliant Blue (a-c, f-h), Alcian Blue (d, e, and k) or both ( i , j , I-n). a, low molecular weight markers. b-e, EDTA extracts of ERD (b , d ) and CRD (c, e) . Lanes b and c were loaded with 2.5 pg of protein, lanes d and e with 0.5 pg. f-k, peak fractions of the EDTA extracts (CRD) after DEAE-chromatography. Differences between ERD and CRD were only found in the phosphoprotein-containing peak IV ( i and j , respectively). 1-n, phosphoprotein-containing frac- tions after CaC12-precipitation and DEAE-chromatography: 1 and rn, CaClz-precipitable fraction of ERD and CRD, respectively; n, CaC12- soluble fraction of CRD. Lanes rn and n were loaded with protein related to equal amounts of collagen.

the EDTA extract). The amount of phosphate was signifi- cantly higher ( p < 0.005; n = 6) in the ERD (1.5 & 0.3 pg of phosphate/mg of collagen) than in the CRD (1.0 +. 0.1 pg of phosphate/mg of collagen) which could not be attributed to the presence of DNA. SDS-PAGE (results not shown) dem- onstrated similar patterns for the ERD and CRD fractions. No Alcian Blue positive bands resembling those of dentinal PP could be discerned.

EDTA Extract-Using collagen as an internal standard, the general protein content of the EDTA extract was found to be about the same in the ERD and the CRD (Table I). In the younger animals general protein/collagen was slightly higher than in the older ones. This was probably due to the lower amount of collagen/wet weight in the younger rats. Virtually no hydroxyproline was found in the EDTA extract.

Considerable differences among the ERD and CRD were found as to the amount of nondialyzable phosphate and the phosphate to general protein ratio (Table I). This was the case for both the younger and the older age groups thus pointing to a difference in PP content between ERD and CRD. The difference in phosphate content between the younger and the older rats again could be explained by the observation that the dentin of the younger rats contained less collagen.

SDS-PAGE (Fig. 2, b e ) showed similar banding patterns for the ERD and CRD. The overall pattern resembled the results obtained by other authors (18). A major difference among the two dentins, however, concerned the Coomassie

I

I

E R D .

.0.4 G IV

0.5 ; 0

/

II / / 10.3

J L v V I VI1

{ 0.1

, CR D

500 1000 t (min) .0.5 / D ,

0 /

IV / . 0.4 CI

II

" - I t - - Ill/ 0 lo.l 0.2 A. 0 1

, . . . . . . . . .

k 0 - E227nrn - - - [NaCI]

FIG. 3. DEAE-chromatographs of the EDTA extract of the ERD and CRD. The differences between the peak areas of the two fractions are shown in Table 111. Peak IV appeared to be the only phosphate containing peak. Portions of 3 mg of protein were applied to the column, which was then eluted at 0.15 ml/min.

staining band at 100 kDa. Whereas the CRD contained an intensely staining band in this region, a faint one was seen in the ERD. A second difference appeared following Alcian Blue staining, the ERD containing one prominent band at 98 kDa, and the CRD two with one at 98 kDa and one at 88 kDa.

We attempted to investigate these differences in some more detail by DEAE-chromatography (Fig. 3) followed by SDS- PAGE. It was found that the E 2 2 7 n m lines of the chromato- graphs of the ERD and CRD exhibited a similar pattern of seven peaks or peak complexes. Peak complexes I and I1 appeared to be not fully separated from each other. Complex I (Fig. 2f) contained bands at 135,68,58,47,39,34, and 11.5 kDa and complex I1 (Fig. 2g) at 160, 147,58,47, 14.5, and 12 kDa. The 100-kDa band (Fig. 2h) which was particularly abundant in the CRD appeared to be associated exclusively with peak 111. Peak IV, a single sharp peak, both with respect to E 2 2 7 n m and phosphate, appeared on SDS-PAGE (Fig. 2, i and j ) as Alcian Blue, but not Coomassie stainable material with molecular mass 98 kDa (ERD) or 98 and 88 kDa (CRD). Based on these characteristics peak IV is considered to consist of phosphoproteins. Peaks V, VI, and VI1 (Fig. 2n) corre- sponded with Alcian Blue stainable material of about 130 to 150 kDa. Their proteoglycan character was confirmed by means of sugar analysis.'

2T. van den Bos, J. Steinfort, and W. Beertsen, manuscript in preparation.

Phosphoproteins in Dentin 2843

TABLE I1 Areas of peak (complex) I to VII as calculated from the E227 ",,, curve of the DEAE-eluate of the EDTA extracts (the older rats), expressed

as percentage of the total area under the curve (n = 6) Values are means f S. D.

Peak no. [NaCl]" Relative area

ERD CRD mM %

I 40 13.9 f 6.3 14.8 f 3.6 I1 95 26.3 f 4.1 33.1 f 4.2'

I11 160 7.8 f 2.4 15.3 f 1.5' IV 230 44.9 f 8.9 V 295 2.8 f 0.8 2.4 f 1.0

30.4 f 5.4d

VI 360 4.1 f 0.7 2.8 f O.Sb VI1 420 0.7 f 0.4 1.2 f 0.6

' In the calculation of [NaCl], a correction for the dead volume of

b p < 0.025. the system (6.75 ml) was made.

' p < 0.001. d p < 0.01.

TABLE 111 Relative amount of organic phosphate and relative peak area (E227 ,J of PP in the DEAE-eluate of samples of the EDTA extract expressed as function of the collagen content in the EDTA insoluble fraction;

CaC12 precipitability of organic phosphate in the EDTA extract (older rats n = 6)

Values are means f S. D.

@ / w V . s/mg P d V . S % ERD

21.2 f 4.7 ] LI 0.44 f 0 . 1 7 b 49.1 f 4 . i 0

91.9 & 3.0 CRD 10.2 f 3.5 0.33 f 0.12 31.4 f 2. 39.6 f 3.7 1 " p < 0.001. ' N.S.

Comparison of the area under the E227nm curve for each peak or peak complex showed a number of differences between ERD and CRD which are summarized in Table 11. Peaks IV and VI were found to be more abundant in the ERD, the CRD being richer in peaks I1 and 111. The total area under the E227nm curve relative to collagen was about the same for the ERD and CRD fractions (1.00 f 0.34 and 1.11 f 0.46 V. s/mg collagen, respectively) which was in agreement with the gen- eral protein content presented in Table I.

Phosphoproteins-More information about the PP in the ERD and CRD fractions was obtained by determination of phosphate in the DEAE-eluate of peak IV. The amount of phosphate/collagen was found to be significantly higher in the ERD than in the CRD (Table 111). This phosphate was bound to organic substances (K2HPO4 applied to the column did not coelute with peak IV).

The higher amount of phosphate/peak area in the ERD indicates that PP in the ERD fraction have on the average a higher degree of phosphorylation than those in the CRD.

CaClz Treatment of Phosphoprotein Fractions-The differ- ence between the ERD and CRD with respect to the degree of PP phosphorylation became much more pronounced when the EDTA extracts were treated with CaClZ (Table 111). In the ERD about 92% of the phosphate appeared to be precip- itable with CaC12 in both age groups, whereas in the CRD only 58 +. 3% (younger rats) and 40 f 4% (older rats) of the phosphate was recovered in the CaClz-sediment (difference between the two age groups: p < 0.001). A control experiment showed that the degree of precipitation with CaClz was inde- pendent of the protein concentrations that were used (0.25- 1.00 mg/ml).

Further studies on the degree of PP phosphorylation were carried out by means of DEAE-chromatography. The results are presented in Table IV.

For the CaClZ-soluble fraction, a great difference was seen between ERD and CRD with respect to the amount of phos- phate and the PP peak area. PP not precipitable with CaCh were much more abundant in the CRD than in the ERD while the amount of phosphate/peak area was higher.

For the CaClz-insoluble fraction, the reverse was found. PP precipitable with CaC12 were more frequent in the ERD than in the CRD (in both age groups). In this fraction a slight difference between the ERD and the CRD occurred with respect to the phosphate/peak area.

Comparing the ratio of phosphate to peak area between the CaClz-soluble and -insoluble fractions showed that the precip- itable PP had a much higher degree of phosphorylation. This was seen in both ERD and CRD although the contrast be- tween both fractions was greatest in the ERD.

SDS-PAGE (Fig. 2, I-n) of the PP peaks after DEAE- chromatography showed that in the ERD the CaClz-precipi- table PP consisted of a single band (98 kDa). After concen- trating the CaClz-soluble fraction 15 times, two bands were seen, a major one at 98 kDa and a minor second one at 88 kDa.

With respect to the CRD, two bands were found, one at 98 kDa and one at 88 kDa, both in the CaClz-precipitable and in the CaClz-soluble fraction. In additional experiments we were unsuccessful in our attempts to separate the two bands fully from each other by means of a CaCL precipitation step. The amount of PP precipitating with CaClz could be increased by lowering the urea concentration and by increasing the cen- trifugation time and speed. With increasing amounts of PP precipitating, the 88-kDa band started dominating the CaClZ- soluble fraction but at the same time a 88-kDa band was seen in the precipitated material.

Phosphoproteins in Dentin-associated Tissues-Samples of the EDTA and guanidine extracts of pulp, remnants of peri- odontal ligament (plus cementum), immature enamel (plus enamel organ), and mature enamel were run on SDS-PAGE. With the exception of the guanidine extract of immature enamel, only faint bands could be detected on the gels, none of them staining with Alcian Blue.

DISCUSSION

The results of the present study have shown that enamel- and cementum-related dentin of the rodent incisor exhibit distinct biochemical differences. Differences were particularly evident with respect to the PP, the prevailing component of the non-collagenous organic matrix (19). The content of the higher phosphorylated PP fraction was about 4 times larger in the ERD than in the CRD, whereas the reverse was seen for the lower phosphorylated PP fraction. SDS-PAGE re- vealed that the ERD contains only one major PP band (98 kDa), the CRD contains two (98 and 88 kDa). Another major difference between the two dentins concerned a 100-kDa Coomassie staining band. Although the exact nature of this protein has not yet been established, sugar analysis (data not shown) revealed that it is a glycoprotein. We therefore con- jecture that it is the alleged 95-kDa glycoprotein (20). The CRD contained about twice the amount found in the ERD.

Similarities between ERD and CRD were found with re- spect to the contents of collagen and total crystal bound non- collagenous protein. Also the basic pattern of non-collagenous proteins on SDS-PAGE and DEAE-chromatography were quite similar for the two dentin portions.

At first sight our finding in the rat that the ERD is richer

Phosphoproteins in Dentin

TABLE IV Amount of phosphate and peak area (EzZ7 .J of PP in the DEAE-eluate of samples of the CaCh-soluble and -insoluble fraction of the EDTA extract as function of the amount of collagen in the EDTA insoluble fraction

Values are means f S. D.

Phosphate/collagen Peak area/collagen Phosphate/ peak area

I.'. d m g M/ v. s CaClz-soluble Younger rats ( n = 4) ERD

CRD 19f 7 Older rats (n = 6) ERD

CRD CaClz-insoluble

Younger rats ( n = 4) ERD CRD

Older rats ( n = 6) ERD 22.8 f 4 . 7 d ] 0.46 f 0.061 .] a

CRD 4.8 f 1.8 0.11 f 0.03 " p < 0.005. b p < 0.025. e N.S. d p < 0.001. ' p < 0.01.

in PP than the CRD, would seem to be in line with recent work of Takagi and co-workers (21) who studied PP in the crown and root fractions of bovine teeth. A problem with the latter work, however, is that in bovine teeth crown and root dentin are not formed at the same age. Differences between the two dentins may therefore be considered to be age-related (see also below). In the rat incisor the crown (ERD) and root (CRD) portions of the dentin cylinder are formed at the same time and at about the same rate (2).

In the present study much attention was given to prevent possible loss of material during the extraction and purification procedures. In order to be able to use the area under the 227 nm peaks as a measure of the amount of protein, DEAE- chromatography was performed using a combination of low flow rates and long run times. This resulted in optimal reso- lution (see also 22).

Despite our attempts to keep the procedures as simple as possible, some material was lost during the experimental procedures. A gradual reduction of the relative amount of phosphate was seen at each step of the separation procedure (compare Tables I, 111, and IV). This could have been due to the initial presence of some remnants of inorganic phosphate or a gradual loss of some organic phosphate. Because the contribution of each of these factors is not known, the exact PP content in the ERD and CRD fractions remains uncertain. On the other hand ERD and CRD were treated simultaneously and identically, and the relative differences found between the two throughout the experiment were quite consistent.

The Polydispersity of Phosphoproteins-The composition of the PP fraction, especially that of the CRD, seems to be rather complex. Both the CaClz-precipitable and the CaC12- soluble PP showed up as the same two bands on SDS-PAGE. This may be explained by assuming that during CaClz treat- ment not only complexes were formed between higher phos- phorylated PP (presumably 98 kDa), but also between higher and lower phosphorylated PP (88 kDa), resulting in two bands in both the CaC12- precipitable and -soluble fraction. This explanation seems to be in line with the observed difference in degree of phosphorylation between the two fractions, in that the CRD shows less variation than the ERD (Table IV).

An alternative explanation would presume that both the 98- and 88-kDa PP occur in the CRD with a varying degree of phosphorylation. It seems hard to conceive, however, that

such a difference in degree of phosphorylation would not result in a discernible shift in molecular mass.

A number of studies (23-25) would seem to suggest that in the rodent incisor there is some turnover of highly phospho- rylated PP following their deposition at the mineralization front. The partial degradation of highly phosphorylated PP resulting in the loss of a PSer- and Asp-rich part (26) could lead to an increase in lower phosphorylated PP. If so, the relatively high concentration of lower phosphorylated PP in the CRD could have been the result of a higher turnover. Indirect support for this is given by the work of Beertsen and Niehof (2) who demonstrated that, in the mouse incisor, the conversion rate of predentin into dentin is higher in the CRD than in the ERD. The difference in CaCL precipitability between the CRD of the younger and older rats could then possibly be explained by assuming that in older rats the incisors are replaced at a lower rate than in younger rats.

The work of Butler et al. (12, 17) and Linde (20), however, does not support this view. These authors reported the pres- ence of at least three different types of PP in rat dentin, which seem to express small but discernible differences in molecular mass and have a different composition, high PP, moderate PP, and low PP. While high PP contains more PSer than moderate PP and low PP, the latter contain significant but low numbers of amino acid residues which do not occur in high PP. These observations make it unlikely that moder- ate PP and low PP are degradation products of high PP.

Are the different PP in rat dentin different gene products or are they the result of differences in breakdown of a precur- sor PP during dentin mineralization and maturation? This issue has not been settled, yet, and requires further investi- gation. In the rat the slow degradation process, as it occurs in bovine (27, 28) and human dentin (29), is probably a factor of minor importance, because of the high turnover-rate of the incisors (about 4 weeks) and also because no signs of Alcian Blue staining low molecular weight fractions were seen on

The question remains what is the function of the various PP as they can be extracted from rat dentin.

Differences in Function between the Phosphoproteins of the ERD and CRD with Respect to Mineralization-The role of PP in dentin mineralization is subject to some controversy (20, 30, 31). Both initiation and retardation of crystal growth

SDS-PAGE.

Phosphoprote

by these components have been reported from in vitro studies. In addition these substances are thought to promote hydrox- ylapatite formation as well as retard the conversion of amor- phous calcium phosphate into apatite. It has been shown that PP form ternary complexes with calcium and phosphate ions (32) may facilitate diffusion of Ca toward mineral surfaces (33) and inhibit the formation of amorphous calcium phos- phate in favor of the formation of hydroxylapatite (34). On the other hand several studies suggest that both the amount of precipitate formed from supersaturated calcium and phos- phate-containing solutions as well as the conversion-rate of amorphous calcium phosphate to apatite are decreased by PP (35-38).

All functions mentioned appear to depend on the degree of phosphorylation. While the initiating role in mineralization is thought to be related to a pattern of (PSer-Asp), (26, 39), the inhibitory effect of PP on mineral precipitation and the stabilization of pre-existing mineral can be explained by as- suming that PP, once adsorbed to the crystal surface, forms a barrier between solution and mineral (40). It has been shown that the effect of the adsorbed protein is related to the number of negative charges, especially PSer (35, 41, 42).

Since the degree of phosphorylation seems to be an impor- tant aspect in the functioning of PP, we suggest that the observed differences in phosphoprotein composition among the ERD and CRD are somehow related to differences in the composition of the inorganic dentin matrix. Recent work in our laboratory has shown that the ERD is harder and denser than the CRD and contains less Mg.3

Combining the present results with those of Beertsen and Niehof (2) who found that the processing of predentin into dentin in the mouse incisor is considerably slower in the ERD than in the CRD, it is suggested that the higher content of highly phosphorylated PP in the ERD may be the cause of this delay.

Finally, our work has shown that the dentin of one tooth may exhibit considerable differences, depending on the site of sampling. This implies that in future studies on dentin composition care must be taken to indicate which part of the tooth the dentin is taken from.

Acknowledgment-We thankfully acknowledge the technical as- sistance of Annemarie Reurings.

REFERENCES

1. Beertsen, W., Niehof, A., and Everts, V. (1985) Am. J. Anat.

2. Beertsen, W., and Niehof, A. (1986) Anat. Rec. 2 1 5 , 106-118 3. Rosenberg, G. D., and Simmons, D. J . (1980) Calc. Tiss. Int. 32 ,

4. Driessens, F. C. M., and Goldberg, M. (1989) J. SOC. Chem. Belge,

5. Mishima, H., and Sakae, T. (1986) J. Dent. Res. 6 5 , 932-934 6. Guis, M. B., Slootweg, R. N., and Tonino, G. J. M. (1973) Arch.

7. Rosen, H. (1957) Arch. Biochem. Biophys. 6 7 , 10-15 8. Kirkpatrick, D. S., and Bishop, S. H. (1971) Anal. Chem. 4 3 ,

174,83-103

45-53

in press

Oral Biol. 18,253-263

1707-1709

J. Steinfort, B. Deblauwe, T. van den Bos, and W . Beertsen, manuscript in preparation.

ins in Dentin 2845

9. Burton, K. (1956) Biochem. J. 62,315-323 10. Bartold, P. M., and Page, R. C. (1985) Anal. Biochem. 150,320-

324 11. Anderson, C. W., Baum, P. R., and Gesteland, R. F. (1973) J.

Virol. 12 , 241-252 12. Butler, W . T., Bhown, M., Dimuzio, M. T., and Linde, A. (1981)

Collagen Rel. Res. 1,187-199 13. Linde, A., Bhown, M., and Butler, W . T. (1980) J. Biol. Chem.

255,5931-5942 14. Termine, J. D., Belcourt, A. B., Christner, P. J., Conn, K. M.,

and Nylen, M. U. (1980) J. Biol. Chem. 255,9760-9768 15. Rahemtulla, F., Prince, C. W., and Butler, W. T. (1984) Biochem.

16. Kuboki, Y., Fujisawa, R., Aoyama, K., and Sasaki, S. (1979) J. Dent. Res. 58,1926-1932

17. Butler, W. T., Bhown, M., DiMuzio, M. T., Cothran, W . C., and Linde, A. (1983) Arch. Biochem. Biophys. 225 , 178-186

18. Butler, W. T. (1987) Methods Enzymol. 145 , Part E, 290-303 19. Linde, A. (1984) in Dentin and Dentinogenesis (Linde, A., ed)

Vol. 11, pp. 55-92, CRC Press, Boca Raton, FL 20. Butler, W . T., Sato, S., Rahemtulla, F., Prince, C. W., Tomana,

M., Bhown, M., Dimuzio, M. T., and Bronckers, A. L. J. J. (1985) in The Chemistry and Biology of Mineralised Tissues (Butler, W. T., ed) pp. 107-112, Ebsco Media, Birmingham, AL

21. Takagi, Y., Nagai, H., and Sasaki, S. (1988) Calcif. Tissue Znt.

22. Rahima, M., and Veis, A. (1988) Calcif. Tissue Int. 4 2 , 104-112 23. Dimuzio, M. T., and Veis, A. (1978) J. Biol. Chem. 2 5 3 , 6845-

24. Dimuzio, M. T., Bhown, M., and Butler, W. F. (1985) Calcif.

25. Maier, G. D., Lechner, J. H., and Veis, A. (1983) J. Biol. Chem.

26. Lechner, J. H., Veis, A., and Sabsay, B. (1981) in The Chemistry and Biology of Mineralised Connective Tissues (Veis, A., ed) pp. 395-398, Elsevier Scientific Publishing Co., Amsterdam

J. 218,877-885

42,97-103

6852

Tissue Int. 3 7 , 242-249

2 5 8 , 1450-1455

27. Lee, S. L., Kossiva, D., and Glimcher, M . J. (1983) Biochemistry 22,2596-2601

28. Fujisawa, R., Kuboki, Y., and Sasaki, S. (1985) Arch. Biochem.

29. Masters, P. M. (1985) Calcif. Tissue Int. 37,236-241 30. Veis, A. (1985) in The Chemistry and Biology of Mineralised

Tissues (Butler, W . T., ed) pp. 170-176, Ebsco Media, Bir- mingham, AL

31. Linde, A. (1985) in The Chemistry and Biology of Mineralised Tissues (Butler, W . T., ed) pp. 344-355, Ebsco Media, Bir- mingham, AL

32. Lee, S. L., Glonek, T., and Glimcher, M. J. (1983) Calcif. Tissue Int. 35 , 815-818

33. Cookson, D. J., Levine, B. A., Williams, R. J. P., Jontell, M., Linde, A., and de Bernard, B. (1980) Eur. J . Biochem. 110 ,

34. Nawrot, C. F., Campbell, D. J., Schroeder, J. K., and Van Val-

35. Termine, J. D., and Conn, K. M. (1976) Calcif. Tissue Res. 2 2 ,

36. Termine, J. D., Eanes, E. D., and Conn, K. M. (1980) Calcif.

37. Udich, H. J., Hoft, H. D., and Bornig, H. (1986) Biomed. Biochim.

38. Fujisawa, R., Kuboki, Y., and Sasaki, S. (1987) Calcif. Tissue Int.

39. Lee, S. L., and Veis, A. (1980) Znt. J. Pept. Protein Res. 16 , 231-

40. Addadi, L., and Weiner, S. (1985) Proc. Natl. Acad. Sci. U. S. A.

41. Reynolds, E. C., Riley, P. F., and Storey, E. (1982) Calcif. Tissue

42. Fujisawa, R., Kuboki, Y., and Sasaki, S. (1986) Calcif. Tissue Int.

Biophys. 243,619-623

273-278

kenburg, M. (1976) Biochemistry 15,3445-3449

149-157

Tissue Znt. 3 1 , 247-251

45 , 703-711

4 1,44-47

240

82,4110-4114

Znt. 34 , S52-S56

39,248-251