Journal of Microscopy, Vol. 195, Pt 1, July 1999, pp. 58–63. Received 15 May 1998 ; accepted 8 December 1998
Quantitative electron spectroscopic diffraction analyses of the crystal formation in dentine S. ARNOLD, U. PLATE,* H. P. WIESMANN,† U. STRATMANN,‡ H. KOHL & H. J. HO¨ HLING* Physikalisches Institut, Wilhelm-Klemm Straße 10; *Institut fu¨r Medizinische Physik und Biophysik, Robert-Koch Str. 31; †Klinik und Poliklinik fu¨r Mund-und Kiefer-Gesichtschirurgie, Waldeyerstr. 30; and ‡Anatomisches Institut, Robert-Koch Straße 26; Westfa¨lische Wilhelms-Universita¨t Mu¨nster, D-48149 Mu¨nster, Germany
Key words. Apatite, biomineralization, crystal nuclei, dentine, electron spectroscopic diffraction, energy-filtering transmission electron microscopy, paracrystal, zero-loss filtering.
Summary Newly formed apatitic crystallites of different hard tissues consist, according to our investigations, of chains composed of nanometre-sized particles (islands, dots) arising at nucleating sites of the collagenous and noncollagenous matrix macromolecules. In dentine these islands coalesce rapidly in longitudinal direction to form needle-like crystallites which further coalesce to ribbon-like crystallites. We have concluded that the centre-to-centre distances between these islands represent the distances between the nucleating sites of the matrix macromolecules. We have applied energy-filtering transmission electron microscopy in the selected area electron diffraction mode at different stages of crystal formation in dentine and have obtained quantitative information of the degree of crystal disorder on the basis of the paracrystal theory. The fluctuation of the lattice plane distances in c-axis direction decreases, proceeding from the region near the dentine/predentine border to the dentine/enamel border.
Introduction Hard tissue formation is a multistep process, in which crystal formation and crystal growth are the final steps. The idea prevailed and still exists that the earliest crystal formations in collagen-rich hard tissues as well as in enamel are thin platelets with a thickness of about 2 nm, but with much longer distances in length (Steve-Bocciarelli, 1970; Glimcher, 1981, 1992; Landis & Glimcher, 1982; Weiner & Traub, 1986; Traub et al., 1989; Bonar et al., Correspondence to: Siegfried Arnold, Physikalisches Institut, Westfa¨lische Wilhelms-Universita¨t
Mu¨nster,
Wilhelm-Klemm
Str.
10,
48149
Mu¨nster, Germany. Tel: þ 49 251 83 33636; fax: þ49 251 83 33602; e-mail:
[email protected]
58
1991; Kim et al., 1995, 1996). Alternatively the opinion exists that the primary crystal formations in collagen-rich hard tissues are needles (Boyde, 1974; Hayashi, 1984; Bonucci, 1985; Katz et al., 1989; Weiner et al., 1991; Fratzl et al., 1992, 1996a,b; Traub et al., 1992a,b). We have observed at high magnification chains composed of nanometre sized calcium–phosphate islands (e.g. Ho¨hling et al., 1990). Since these chains are orientated parallel to the bipolar c-axis of apatite, we have concluded that the centreto-centre distances between these particles represent the distances between the nucleating sites of the fibrous matrix macromolecules (Ho¨hling et al., 1990, 1995; Plate et al., 1992, 1994, 1998; Arnold, 1994; Arnold et al., 1996, 1997). X-ray and electron diffraction analyses of developing crystals in hard tissues and calcium–phosphate pharmaceuticals have shown that these calcium–phosphate crystallites describe structurally an intermediate state between amorphous and fully crystalline (Ho¨hling, 1966, 1989; Wheeler & Lewis, 1977; Grynpas et al., 1984; Bonar et al., 1985; Glimcher, 1992; Fukuoka et al., 1995). The early crystallites in dentine already describe structurally an apatitic character with lattice distortions. The degree of crystallinity of the calcium–phosphate crystallites increases during mineralization, but even in the mature stage lattice distortions of the apatite crystallites exist. The lattice distortions of the calcium–phosphate crystallites of developing bone and of synthetic crystalline pharmaceuticals, analysed by X-ray diffraction, were resolved approximately by the theory of paracrystallinity of Hosemann & Bagchi (1962) (Wheeler & Lewis, 1977; Fukuoka et al., 1995; Yamamura et al., 1996). In this study the degree of lattice distortion of the apatitic crystallites in circumpulpal dentine of rat incisors will be discussed on the basis of the paracrystal theory. The q 1999 The Royal Microscopical Society
E SD A NA LYS E S O F C RYS TA L F O R M AT I O N S
paracrystal size, i.e. the size of coherent diffraction domains, and the lattice distortion of apatite were determined from linescans of energy-filtered selected-area electron diffraction (SAED) patterns. Newly formed and more mature dentine crystallite structures were compared with regard to the biomineralization processes in dentine.
Materials and methods Young Sprague–Dawley rats (60–70 g) and young Wistar rats (50–90 g) were anaesthetized, and the upper and lower incisors were rapidly dissected and transferred to liquidnitrogen-cooled propane. The time for tissue preparation was about 1–3 min. These shock-frozen specimens were freeze-dried at ¹ 808C. After slowly warming to room temperature they were vacuum embedded in epoxy resin. Ultrathin sections were cut transversal to the longitudinal axis of the incisor with an ultramicrotome to a thickness of less than 80 nm, thereby ensuring that the time of water contact in the microtome trough was minimized to avoid crystal dissolution (Plate et al., 1992). Elastically (zero-loss) filtered SAED patterns of about 30 different specimens were taken by an energy-filtering transmission electron microscope (EFTEM) Zeiss EM902 at an acceleration voltage of 80 kV for different stages of mineralization. Using zero-loss filtering the background of the inelastic scattered electrons decreases and faint Debye– Scherrer rings, which cannot be recognized in unfiltered patterns appear (Reimer et al., 1992). Linescans of these patterns were obtained by a scintillator–photomultiplier combination under the final screen of the EFTEM, so that the zero-loss filtered SAED patterns were shifted sequentially
over the detector by scanning coils (Arnold, 1994; Hu¨lk, 1994; Plate et al., 1994; Arnold et al., 1997). According to the theory of Hosemann et al. (Hosemann & Bagchi, 1962; Hosemann & Hindeleh, 1995) the lattice of a paracrystal is characterized by the fact that the periodically arranged atoms fluctuate around their mean atomic distance in all directions. If lattice bricks have different sizes and shapes and are mixed statistically, their mutual distances fluctuate and they build a paracrystalline lattice (Hosemann et al., 1985; Hindeleh & Hosemann, 1991). This fluctuation can be characterized by the fluctuation factor ghkl with hkl as the Miller indices. This fluctuation factor and the paracrystal size L, which correspond to the size of coherent diffraction domains, are proportional to the integral breadth db of the corresponding diffraction profile. These parameters can be determined by the equation db¼1/ L+(pghklh)2/d¯hkl (Hosemann & Bagchi, 1962; Hosemann et al., 1985; Hosemann & Hindeleh, 1995), where d¯hkl is the mean lattice plane distance of the hkl lattice plane and h2 is the successive order of the Bragg reflections obtained. The integral breadth is the integral under the recorded diffraction profile above the background divided by the amplitude of the profile without background and is corrected by the Wagner method db¼dbobs¹(dbinst)2/dbobs (Wagner, 1966), with dbobs as the observed and dbinst as the integral breadth of the instrumental broadening profile.
Results Zero-loss filtered SAED patterns (Fig. 1a,b) and their corresponding linescans (Fig. 2a,b) of about 30 different specimens were taken in the mineralized circumpulpal
Fig. 1. Two examples of zero-loss filtered SAED patterns of (a) early formed and (b) mature crystallites in dentine. q 1999 The Royal Microscopical Society, Journal of Microscopy, 195, 58–63
59
60
S. A R N O L D ET AL .
Fig. 2. Corresponding linescans of the zero-loss filtered SAED patterns of Figure 1; of (a) early formed and (b) mature crystallites in dentine.
dentine near the mineralization front (about 10 mm behind the predentine/dentine border) and in a region about 5 mm before the dentine/enamel border (Fig. 3a,b). In crystallites, without any lattice distortion, the integral breadth db of the reflection is proportional to the reciprocal crystal size L, but in paracrystalline materials db is also a function of the lattice fluctuation g (Hosemann & Hindeleh, 1995). Figure 4 shows two plots of db plotted against the successive orders h2 of Bragg reflections in the early and the mature biomineralization stage corresponding to the linescans shown in Fig. 2. All plots of the integral breadth increase linearly, and this indicates the paracrystalline character of the apatitic crystallites for these mineralization stages in the circumpulpal dentine. The paracrystalline fluctuation factor g00l of early formed and mature crystallites (Table 1) shows that the apatitic crystallites in the circumpulpal dentine of rat incisors have
lattice fluctuations during the whole mineralization process. In the early stage of mineralization a fluctuation of about 3·4% and for the mature stage a fluctuation of about 2·3% of the mean atom lattice distances exist. The paracrystal size L of the crystallites along the c-axis of apatite is in the region of 4·8 nm for the early mineralization stage, about 10 mm behind the predentine/dentine border, and of 21 nm for the mature mineralization stage in dentine, at a distance of about 5 mm from the enamel/dentine border (Table 1).
Discussion According to the morphological and structural analyses of the crystal formation of circumpulpal dentine, the primary crystallites are chains composed of nanometresized apatitic particles (islands) (Fig. 5). The centre-tocentre distances between these islands represent, according
Fig. 3. Zero-loss filtered electron spectroscopic images of the (a) predentine (Pr)/dentine (De) border (mineralization front) and (b) enamel (En)/dentine (De) border. q 1999 The Royal Microscopical Society, Journal of Microscopy, 195, 58–63
E SD A NA LYS E S O F C RYS TA L F O R M AT I O N S
61
Fig. 4. Integral breadth db plotted against the successive orders h2 of the reflections (002), (004), (006) derived from the linescans of Fig. 2.
to our conclusions, the distances between the nucleating sites along the fibrous collagenous and noncollagenous matrix macromolecules. These chains of islands coalesce rapidly in longitudinal direction to needle-like crystallites, which further coalesce laterally to ribbon-like crystallites (Ho¨hling et al., 1990, 1995; Arnold, 1994; Plate et al., 1994, 1998; Arnold et al., 1996, 1997). The limit of the number Nhkl of lattice planes and the restriction of crystal growth are explained in the paracrystalline theory by the fluctuation of lattice distances, √ described by the empirical a* relation expressed by Nhkl ghkl ¼ a* ¼ 0·15 6 0·05. This means that the surface lattice planes of paracrystals have a statistical distance fluctuation of about 15%; this implies a limit of the number Nhkl of lattice planes and restricts the growth of the paracrystal (Hosemann et al., 1985; Hindeleh & Hosemann, 1991; Hosemann & Hindeleh, 1995). The value of a* during the whole apatitic biomineralization in circumpulpal dentine lies between 0·10 and 0·20 (Table 1), and is in the same range as other calcium–phosphate formations (Wheeler & Lewis, 1977; Fukuoka et al., 1995) and other paracrystals, e.g. molten metals, SiO2 glasses and polymers (Hosemann et al., 1985; Hosemann & Hindeleh, 1995). For the calculation of the paracrystal size L and the degree of lattice distortion (g-factor, paracrystallinity) concerning the Bragg reflections of the c-axes (0 0 2l) of
Fluctuation factor g00l Crystal size L concerning c-axis (nm) a* ¼ L/d00l g00l Subdivision according to the g-factor (Hosemann & Hindeleh, 1995)
apatite, we have chosen the biomineralization process for early formed crystallites and for mature crystallites in the circumpulpal dentine. The crystal distortion was analysed quantitatively on the basis of the paracrystal theory with the lattice fluctuation factor ghkl (Hosemann & Bagchi, 1962). It was found that the lattice fluctuation for the lattice planes, the lattice distortion, in c-axis direction (0 0 2l) decreases with the maturation of the circumpulpal dentine, although the apatitic crystallites still have lattice distortions during the whole biomineralization process. The early formed crystallites near the mineralization front (predentine/dentine border) have a fluctuation factor g00l ¼ 3·4%, while the more mature apatitic crystallites in dentine near the enamel/dentine border have a fluctuation factor g00l ¼ 2·3%. These results can be correlated with the results of the radial distribution function (RDF) analyses in developing bone (Grynpas et al., 1984; Bonar et al., 1985; Glimcher, 1992). The RDF analyses have led to the conclusion that the apatitic crystal formations in bone are poorly crystalline hydroxyapatite, also with a decreasing disorder during maturation, whereby the crystal order would not reach the fully crystalline state. However, quantitative factors for the degree of lattice distortion, such as the g-factor in this study, have not yet been evaluated for the RDF analyses on developing hard tissues (to our knowledge). In our study the subdivision of the
Early formed dentine
Mature dentine
3·4 6 (j ¼ 0·4) 4·8 6 (j ¼ 0·9) 0·10 6 (j ¼ 0·02)
2·3 6 (j ¼ 0·3) 21·1 6 (j ¼ 1·4) 0·14 6 (j ¼ 0·03)
biopolymer
‘Single crystal’ of polymer
q 1999 The Royal Microscopical Society, Journal of Microscopy, 195, 58–63
Table 1. Paracrystalline parameters of early formed crystallites near the mineralization front (predentine/dentine border) and of mature crystallites near the dentine/ enamel border; dhkl ¼ interplanar spacing.
62
S. A R N O L D ET AL .
Fig. 5. Zero-loss filtered electron spectroscopic image of an early mineralized noncollagenous/collagenous fibre; arrows show dots of the mineral chains at the surface and inside the fibril.
apatitic crystallites of circumpulpal dentine at different mineralization stages is comparable to ‘biopolymers’ of the newly formed crystallites and to ‘single crystals of polymers’ of mature minerals on the basis of the Hosemann definition according to the fluctuation factor g00l (see Hosemann & Hindeleh, 1995; Table 1).
Acknowledgement Financial support by the Deutsche Forschungsgesellschaft is gratefully acknowledged.
References Arnold, S. (1994) Quantitative Untersuchungen der Dentin-Mineralisierung von Ratten-Incisoren mittels energie-dispersiver Elektronenbeugung. Diplom-thesis in Physics, Mu¨nster. Arnold, S., Plate, U., Kohl, H. & Ho¨hling, H.J. (1996) EFTEManalyses of biomineralisation in dentine of rat incisors. Proc. 11th Eur. Congr. on Electron Microscopy, Dublin. Eurem 96, Belfield, Dublin. Arnold, S., Plate, U., Wiesmann, H.P., Kohl, H. & Ho¨hling, H.J. (1997) Quantitative electron-spectroscopic diffraction (ESD) and electron spectroscopic imaging (ESI) analyses of dentine mineralisation in rat incisors. Cell Tissue Res, 288, 185–190. Bonar, L.C., Grynpas, M.D., Roberts, J.E., Griffin, R.G. & Glimcher, M.J. (1985) Physical and chemical characterisation of the development and maturation of bone mineral. The Chemistry and Biology of Mineralized Tissues (ed. W. T. Butler), pp. 226–233. Ebsca Medica, Birmingham. Bonar, L.C., Shimizu, M., Roberts, J.E., Griffin, R.G. & Glimcher, M.J. (1991) Structural and composition studies on the mineral of newly formed dental enamel: a chemical, X-ray diffraction, and 31 P and proton nuclear magnetic resonance study. J. Bone Miner. Res, 6, 1167–1176.
Bonucci, E. (1985) The structural basis of calcification. Ultrastructure of the Connective Tissue Matrix (eds A. Ruggeri and P. M. Motta), pp. 165–191. Matinius Nijhoff, Boston. Boyde, A. (1974) Transmission electron microscopy of ion beam thinned dentine. Cell Tissue Res, 152, 543–550. Fratzl, P., Groschner, M., Vogl, G., Plenk, H. Jr, Eschberger, J., Fratzl Zelman, N., Koller, K. & Klaushofer, K. (1992) Mineral crystals in calcified tissues: a comparative study by SAXS. J. Bone Miner. Res, 7, 329–334. Fratzl, P., Paris, O., Klaushofer, K. & Landis, W.J. (1996b) Bone mineralization in an osteogenesis imperfecta mouse model studied by small-angle x-ray scattering. J. Clin. Invest, 97, 396–402. Fratzl, P., Schreiber, S. & Boyde, A. (1996a) Characterization of bone mineral crystals in horse radius by small-angle x-ray scattering. Calcif. Tissue Int, 58, 341–346. Fukuoka, E., Terada, K., Makita, M. & Yamamura, S. (1995) Paracrystalline lattice distortion in crystalline pharmaceuticals: determination of paracrystalline lattice distortion by powder xray diffraction. Chem. Pharm. Bull, 43, 671–676. Glimcher, M.J. (1981) On the form and function of bone: from molecules to organs. The Chemistry and Biology of Mineralised Connective Tissues (ed. A. Veis), pp. 617–673. North-Holland, Amsterdam. Glimcher, M.J. (1992) The nature of the mineral component of bone and the mechanism of calcification. Disorders of Bone and Mineral Metabolism (eds F. L. Coe and M. J. Favus), pp. 265–286. Raven Press, New York. Grynpas, M.D., Bonar, L.C. & Glimsher, M.J. (1984) Failure to detect an amorphous calcium-phosphate solid phase in bone mineral: a radial distribution function study. Calcif. Tissue Int, 36, 291–301. Hayashi, Y. (1984) Crystal growth in calcifying fronts during dentinogenesis. Acta Anat, 118, 13–17. Hindeleh, A.M. & Hosemann, R. (1991) Microparacrystals: the intermediate state between crystalline and amorphous. J. Mater. Sci, 26, 5127–5133. q 1999 The Royal Microscopical Society, Journal of Microscopy, 195, 58–63
E SD A NA LYS E S O F C RYS TA L F O R M AT I O N S
Ho¨hling, H.J. (1966) Die Bauelemente von Zahnschmelz und Dentin aus Morphologischer, chemischer und struktureller Sicht. Carl Hanser, Mu¨nchen. Ho¨hling, H.J. (1989) Special aspects of biomineralisation of dental tissues. Teeth. Handbook of Microscopic Anatomy Vol. V/6 (eds A. Oksche and L. Vollrath), pp. 475–524. Springer-Verlag, Berlin. Ho¨hling, H.J., Arnold, S., Barckhaus, R.H., Plate, U. & Wiesmann, H.P. (1995) Structural relationship between the primary crystal formations and the matrix macromolecules in different hard tissues. Discussion of a general principle. Connect. Tissue Res, 33, 493–500. Ho¨hling, H.J., Barckhaus, R.H., Krefting, E.R., Althoff, J. & Quint, P. (1990) Collagen mineralization: aspects of the structural relationship between collagen and the apatitic crystallites. Ultrastructure of Skeletal Tissue (eds E. Bonucci and P. M. Motta), pp. 41–62. Kluwer, Dordrecht. Hosemann, R. & Bagchi, S.N. (1962) Direct Analysis of Diffraction by Matter. North-Holland, Amsterdam. Hosemann, R., Hentschel, M.P., Balta-Calleja, F.J. & Hindeleh, A.M. (1985) Evalution of the alpha-constant in polymers, biopolymers and catalysts. Exp. Tech. Phys, 33, 135–148. Hosemann, R. & Hindeleh, A.M. (1995) Structure of crystalline and paracrystalline condensed matter. J. Macromol. Sci. Phys. B34, 327–356. Hu¨lk, C. (1994) Energiefilterung der Debye-Scherrer-Ringe von polykristallinen Aufdampfschichten. Diplom-thesis in Physics, Mu¨nster. Katz, E.P., Wachtel, E., Yamauchi, M. & Mechanic, G. (1989) The structure of mineralized collagen fibrils. Connect. Tissue Res, 21, 149–158. Kim, H.M., Rey, C. & Glimcher, M.J. (1995) Isolation of calciumphosphate crystals of bone by non-aqueous methods at low temperature. J. Bone Miner. Res, 10, 1589–1601. Kim, H.M., Rey, C. & Glimcher, M.J. (1996) X-ray diffraction, electron microscopy, and Fourier transform infrared spectroscopy of apatite crystals isolated from chicken and bovine calcified cartilage. Calcif. Tissue Int, 59, 58–63. Landis, W.J. & Glimcher, M.J. (1982) Electron optical and analytical observations of rat growth plate cartilage prepared by ultracryomicrotomy. J. Ultrastruct. Res, 78, 227–268.
q 1999 The Royal Microscopical Society, Journal of Microscopy, 195, 58–63
63
Plate, U., Arnold, S., Reimer, L., Ho¨hling, H.J. & Boyde, A. (1994) Investigation of the early mineralization on collagen in dentine of rat incisors by quantitative electron spectroscopic diffraction (ESD). Cell Tissue Res, 278, 543–547. Plate, U., Arnold, S., Stratmann, U., Wiesmann, H.P. & Ho¨hling, H.J. (1998) General principle of ordered apatitic crystal formation in enamel and collagen rich hard tissues. Connect. Tissue Res. 38, 149–157. Plate, U., Ho¨hling, H.J., Reimer, L., Barckhkaus, R.H., Wienecke, R., Wiesmann, H.P. & Boyde, A. (1992) Analyses of the calcium distribution in predentine by EELS and the early crystal formation in dentine by ESI and ESD. J. Microsc, 166, 329–341. Reimer, L., Fromm, I., Hirsch, P., Plate, U. & Rennekamp, R. (1992) Combination of EELS modes and electron spectroscopic imaging and diffraction in an energy-filtering electron microscope. Ultramicroscopy, 46, 335–347. Steve-Bocciarelli, D. (1970) Morphology of crystallites in bone. Calcif. Tissue Res, 5, 261–269. Traub, W., Arad, T. & Weiner, S. (1989) Three-dimensional ordered distribution of crystals in turkey tendon collagen fibers. Proc. Natl. Acad. Sci. USA, 86, 9822–9826. Traub, W., Arad, T. & Weiner, S. (1992a) Growth of mineral crystals in turkey tendon collagen fibers. Connect. Tissue Res, 28, 99–111. Traub, W., Arad, T. & Weiner, S. (1992b) Origin of mineral crystal growth in collagen fibrils. Matrix, 12, 251–255. Wagner, C.N.J. (1966) Analyses of the broadening and changes in position of peaks in an X-ray powder pattern. Local Atomic Arrangements Studied by X-Ray Diffraction. (eds J. B. Cohen and J. E. Hilliard), pp. 219–268, Gordon and Breach, New York. Weiner, S., Arad, T. & Traub, W. (1991) The structure of bone lamellae. FEBS Lett, 285, 49–54. Weiner, S. & Traub, W. (1986) Organization of hydoxyapatite crystals within collagen fibrils. FEBS Lett, 206, 262–266. Wheeler, E.J. & Lewis, D. (1977) An x-ray study of the paracrystalline nature of bone apatite. Calcif. Tissue Res, 24, 243–248. Yamamura, S., Momose, Y., Terada, K. & Fukuoka, E. (1996) Evaluation of paracrystalline lattice distortion by profile analysis using a single x-ray diffraction peak. Int. J. Pharm, 133, 117–125.