The structure and distribution of proteochondroitin sulphate during the formation of chick embryo feather germs

The feather germ in chick embryo skin is formed through epidermal-dermal interactions (Sengel, 1976), which cause elongation of epidermal cells (epidermal placode) and condensation of dermal cells (dermal condensation). The distribution patterns of various components of the extracellular matrix during feather germ morphogenesis have been immunohistochemically elucidated by our group and that of Sengel (Kitamura, 1981; Mauger el al. 1982b). Collagen types I and III decreased in the region of condensation and increased in the surrounding region of condensation. Fibronectin, however, was found in the former region. Furthermore, basement membrane components, such as laminin and collagen type IV, showed uniform distribution during feather germ formation (Mauger et al. 1982b).

Proteoglycan, a molecule classified according to the type and distribution of glycosaminoglycans attached to the protein core, is also a major extracellular matrix component (Hascall & Hascall, 1981; Heinegard & Paulsson, 1984). The type, amount and distribution of glycosaminoglycan or proteoglycan are known to change during tissue and organ development. Proteochondroitin sulphate changes in these respects with the formation of cartilage (Kitamura & Yamagata, 1976; Okayama, Pacifici & Holtzer, 1976; Shiomura et al. 1984; Matsui, Oohira, Shoji & Nogami, 1986; Kimata et al. 1986). Regional changes in glycosaminoglycan in the basal lamina during the development of mouse embryonic salivary and sub-mandibular glands have been reported (Bernfield & Banerjee, 1972, 1982). Furthermore, the distribution profile of proteoheparan sulphate in the basement membrane changes during mouse kidney tubule and tooth development (Ekblom, 1981; Thesleff et al. 1981).

The present study was conducted so as to gain some understanding of the structure and distribution of proteoglycan during feather germ formation. In this research, the chick embryo dorsal skin was found to synthesize three types of proteoglycans and, with advancement of morphogenesis, significant changes in the distribution pattern of the major type of proteoglycans take place.

Fertilized eggs of White Leghorn were obtained commercially and incubated at 37°C until use.

The dorsal skin of an 8-day-old embryo, placed on a Millipore filter (HABP04700, Millipore), was labelled for 8h with [³⁵S]sulphate (carrier free, New England Nuclear) at 50pCiml^-1 in Eagle’s MEM containing 10% horse serum and 5 % chick embryonic extracts. The dorsal skin was detached from the filter and proteoglycan extracted overnight at 4 °C with 5vol. of 20mM-Tris-HCl buffer, pH 7·0 (containing 4M-guanidium chloride, 0·1 M-6-amino-hexanoic acid, 10mM-EDTA, 5 mM-benzaminidine, 10mM-N-ethylmaleimide and 1 mM-PMSF) by slow rotation. The suspension was centrifuged at 20 000 revs min^-1 for 30 min at 4°C and the residue reextracted with 2 vol. of the same buffer. The suspension was centrifuged as above. The supernatants were combined and macromolecules precipitated by the addition of 3 vol. of 95% ethanol containing 1·3 % potassium acetate. The precipitates were suspended in 2 vol. of water and precipitated with ethanol to remove guanidium chloride.

The precipitates from the embryonic skin extract were dissolved with 20 mM-Tris-HCl buffer, pH 7·3 (containing 7M-urea, 0·2% (w/v) Triton X-100, 0·1 M-aminohexanoic acid, 10 mM-EDTA, 5mM-benzaminidine, 10 mM-A-ethyl-maleimide and 1 mM-PMSF) and applied onto a DEAE-Sephacel column (l·4x10·0cm, Pharmacia) equilibrated with the same buffer. The column was eluted by a linear gradient of NaCl (0·05-0·80 M). The [³⁵S]sulphate-containing fractions were separately pooled and precipitated with ethanol. The precipitates were dissolved with 50mM-Tris-HCl buffer, pH 7·0 (containing 4 M-guanidium chloride and 0·2% (w/v) Triton X-100) and applied onto a Sepharose CL-2B or 6B column (l·0x94·0cm, Pharmacia) equilibrated with the same buffer. The [³⁵S]sulphate-containing fractions were pooled and precipitated with ethanol.

The precipitates of the fractions obtained from the column chromatography on Sepharose CL-2B and -6B were treated with 0·2M-sodium hydroxide for 20 h at 4°C. The solution was cooled in an ice bath and neutralized with 1 M-acetic acid. Following the addition of chondroitin-4-sulphate and chondroitin-6-sulphate as carriers, the sample solution was precipitated with ethanol and applied onto a Sepharose CL-6B column.

The precipitates of the fractions obtained from alkali treatment followed by gel chromatography were dissolved in 50 mM-Tris-HCl buffer, pH 8·0, and digested with chon-droitinase ABC (final concentration 0·2i.u. ml^-1, Seikagaku Kogyo) for 20 h at 37°C. The digest was applied onto a Sepharose CL-6B column.

Fractions not digested by chondroitinase ABC were pooled and precipitated with ethanol. The precipitate was dissolved in a solution made of equal volumes of 3·6 M-acetic acid and 0·48M-sodium nitrite and allowed to react for 100 min at room temperature (Stow, Glasgow, Handley & Hascall, 1982). 0·5 M-ammonium sulphamate was added to the reaction mixture and the sample was then precipitated with ethanol and applied onto a Sepharose CL-6B column.

The relative amounts of 4- and 6-sulphated disaccharides in the chondroitinase ABC digest of the released glycos-aminoglycans were estimated by the procedure of Saito, Yamagata & Suzuki, 1968.

The dorsal skin of an 8-day-old embryo was treated with 0·25 % EDTA in phosphate-buffered saline (PBS) for 10 min at 0°C and the epidermis was dissociated from the dermis with a needle. Epidermis and dermis were labelled separately with [³⁵S]Jsulphate as above. Proteoglycan was extracted from the labelled epidermis and dermis. DEAE-Sephacel column chromatography, gel chromatography on Sepharose CL-2B and -6B and chondroitinase ABC-digestion were carried out as above.

Monoclonal antibodies were prepared according to Galfre & Milstein, 1981. Female Balb/c mice were immunized with PCS-I (see Results) purified from the dorsal skin. The first injection (150μg of PCS-I in complete Freund’s adju-vant) were administered subcutaneously at several sites along the flank. Three booster shots (100μg of PCS-I in incomplete Freund’s adjuvant) were administered intraperitoneally at one month intervals. 4 days prior to its being subjected to hybridoma fusion, one mouse was administered an intravenous injection of 100pg of PCS-I without an adjuvant. This mouse was then sacrificed under a CO₂ atmosphere and the spleen cells (4x10⁸ cells) were fused with mouse myeloma SP2/0 cells (8x10⁷ cells) by the addition of 1ml of 50% polyethylene glycol 4000 (Merk). The fused cells were suspended in the DME-medium supplemented with 15 % (v/v) precolostrum newborn calf serum, 100μM-hypoxanthine, 16pM-thymidine, 50i.u. ml^-1 penicillin and 100μg ml^-1 streptomycin, and plated on six 96-well Coaster plates. On the following day, the medium was changed to HAT-medium prepared by the adding of 0·4μM-aminopterin to the medium described above. After 10 days, screening was conducted for PCS-I by an enzyme-linked immunosorbent assay (ELISA). For the assay, 0·2 ug of PCS-I was coated on each well of a 96-well microtitre plate (Dynatech, Immulon II). After blocking the plate with DME-medium (containing 15 % precolostrum new-born calf serum and 15% rabbit serum), the culture supernatant was introduced into each well and incubated at 37°C for 2h, followed by washing in five changes of 10 mM-phosphate buffer, pH 7·4 (containing 0·15M-NaCl, 0·05% Tween 20 and 0·1% BSA). Peroxidase-conjugated anti-mouse IgG (Dako) was then added to each well and followed by incubation at 37°C for 1 h. Each well was washed as above. The peroxidase complex was reacted with ABTS (Zymed). Well absorbance at 420 nm was determined with an ELISA plate reader (Corona Co.). The cell line T3B9’ (see Results) has been cloned by the limiting dilution method. The subclass of the monoclonal antibody T3B9’ was determined using a mouse monoclonal typing kit (Serotec). The specificity of I3B9 was examined by inhibition ELISA under nonequilibrium conditions (Rennard et al. 1980). The proteoglycan, collagen and fibronectin to be examined were dissolved in 0·02M-sodium phosphate, 0·15M-NaCl, pH7-2 (200μgml^-1), and serial dilutions of these solutions were added to the culture supernatant obtained from the cloned I3B9 cells (diluted 1:5 with PBS) and reacted overnight at 4°C. Portions (50/21) of the mixture were added to the PCS-I-coated wells and incubated for 2h at 37 °C to allow the remaining free antibody to bind. Succeeding steps in the assay were performed as described above. Proteoglycan ‘PGH’ was purified from the epiphyseal cartilage of a 13-day-old embryo (Kimata et al. 1971). Proteoglycan ‘PGM’ was purified from the limb buds of a 4-day-old embryo (Kitamura & Yamagata, 1976). Type I collagen was purified from the cranial bones of a 17-day-old embryo (von der Mark, von der Mark & Gay, 1976). Type IV collagen was purified from chicken gizzard (Mayne & Zettergren, 1980).

The dorsal skin of a chick embryo was fixed with 3-5 % formaldehyde in PBS for 30 min at 0°C, impregnated with 30 % sucrose, embedded in Tissue Tek II compound (Miles ó Scientific), frozen and then cut into 10pm thick sections x with a cryostat. The frozen sections were placed on a ‘_u polylysine-coated glass slide and stored at —20°C. They were then washed with PBS at room temperature to remove 7 the compound. The epidermal sheet dissociated from the g basement membrane and dermis was fixed with 96 % g ethanol and placed on a gelatin-coated glass slide. Both the sections and epidermal sheets were treated with the culture ^s supernatant of I3B9 for 1 h at 37 °C, followed by washing in five changes of PBS. The secondary antibody consisting of FTTC-conjugated Ig fractions of goat anti-mouse IgG (diluted 1:100 in PBS containing 1% BSA, Cappel) was made to react with each section and sheet for 30 min in the dark. The slides were rinsed as above and then mounted using buffered glycerol. The slides were viewed with a Zeiss standard microscope equipped with a IV FI epifluorescence condenser (Carl Zeiss). Photographs were taken using Kodak Tri-X film of ASA 400. Some sections were treated with 0·1 % (w/v) testicular hyaluronidase (Type IV, Sigma) in PBS for 1 h at 37°C prior to reaction with I3B9. The treated sections were washed three times with PBS and subjected to immunofluorescent staining as above.

The dorsal skin of various stages was labelled for 1 h with [³⁵S]sulphate at 50μCiml^-1 by the procedure described above. For pulse labelling, the labelled skin was washed and immediately fixed for 30 min in Bouin’s solution. In the pulse-chase labelling experiments, the labelled skin was washed and its organ culture continued in [³⁵S]sulphate-free medium. At a specified time, it was washed again and fixed as above. The fixed skin was dehydrated and embedded in Paraplast (Sherwood Medical) and sectioned at 5μm. The slides carrying the sections were dipped into an autoradio-graphic emulsion (Sakura NR-M2, Konishiroku) followed by exposure for 2 weeks in the dark at 4°C. Using a temperature-controlled water bath at 20°C, the slides were developed for 4 min with Konidol X (Konishiroku) and fixed for 8 min with Konifix X (Konishiroku). After washing with tap water, they were dehydrated and mounted with Permount (Fisher Scientific). Before dipping the slides in the above emulsion, some of the sections were digested with chondroitinase ABC for 1 h at 37 °C.

The dorsal skin of an 8-day-old embryo was metabolically labelled in a medium containing [³⁵S]sulphate. Approximately 95% of the incorporated radioactivity was solubilized by extracting two times. [³⁵S]Sulphate-labelled macromolecules were chromatographed on a DEAE-Sephacel column with a NaCl gradient (Fig. 1). The macromolecules were eluted as two peaks at NaCl concentrations of 0·40 M (pooled fractions 105-128) and 0·55 M (pooled fractions 135-155) and designated as fractions I and II, respectively, which were then purified by gel chromatography. Fraction I was chromatographed on a Sepharose CL-6B column and eluted as a single peak with a K_av of 0·38 (Fig. 2A). Fraction II was chromatographed on Sepharose CL-2B and eluted as a single broad peak with a K_av of 0-19 (Fig. 2B).

These fractions, following purification, were subjected to alkali treatment, chondroitinase ABC digestion and nitrous acid degradation to examine the glycosaminoglycans. After each treatment, glycos-aminoglycans were chromatographed on a Sepharose CL-6B column. The glycosaminoglycans of fraction I, liberated by alkali treatment, eluted as a single peak with a k_av of 0·55 (Fig. 3A). About 83 % of the released glycosaminoglycans of fraction I were digested by chondroitinase ABC (Fig. 3B) and the relative amounts of isomeric chondroitin sulphates obtained were as follows: chondroitin-4-sulphate, 62-0%; chondroitin-6-sulphate, 38-0%. The chon-droitinase-resistant fraction (pooled fractions 65-97 in Fig. 3B, 17%) were degraded by nitrous acid (Fig. 3C) and the chondroitinase ABC-resistant fraction has been shown to be proteoheparan sulphate. The glycosaminoglycans of fraction II, liberated by alkali treatment, were chromatographed on a Sepharose CL-6B column and eluted as a single peak with a K_av of 0·25 (Fig. 4A). The released glycosaminoglycans were digested for the most part by chondroitinase ABC (Fig. 4B) and the relative amounts of isomeric chondroitin sulphates in the glycosaminoglycans of fraction II were as follows: chondroitin-4-sulphate, 23·7 %; chondroitin-6-sulphate, 76·3 %.

The types and yields (of the solubilized radioactivity) of proteoglycans metabolically labelled in the dorsal skin were as follows: high molecular weight proteochondroitin sulphate (designated as PCS-I), 54-8%; low molecular weight proteochondroitin sulphate (designated as PCS-II), 37·5% and proteo-heparan sulphate (designated as PHS), 7·7 %.

The labelled epidermis and dermis were separately extracted and chromatographed on a DEAE-Sephacel column and both were found to contain fraction I and II (Fig. 5). Gel chromatography on Sepharose CL-2B and digestion with chondroitinase ABC indicated fraction II from the epidermis and dermis to be the same as PCS-I (Fig. 6A,B). No further analysis of fraction I in epidermis and dermis was made.

Monoclonal antibodies were prepared against PCS-I, the main proteoglycan in the dorsal skin of an 8-day-old embryo. Eighteen clones were found to be positive for PCS-I by ELISA. Among these, the monoclonal antibody, I3B9, which showed the highest activity in ELISA, was characterized and found to be of the IgG 2b subclass. To determine the specificity of I3B9, inhibition ELISA was carried out against various proteoglycans, such as PCS-I, chondroitinase ABC-digested PCS-I, PCS-II, PGH purified from embryonic cartilage and PGM purified from the limb bud of a stage-24 chick embryo. I3B9 reacted with PCS-I and chondroitinase ABC-digested PCS-I, but not with PCS-II or PGH (Fig. 7A). PGM weakly cross reacted with I3B9 (Fig. 7A). Neither did it react with collagen type I or IV, nor fibronectin (Fig. 7B). Thus, I3B9 appears highly specific for PCS-I and may possibly recognize the core protein of PCS-I as an epitope.

In the dorsal skin of a 6-day-old embryo, from which no feather rudiments had formed, PCS-I was present in both the epidermis and superficial dermis (Fig. 8A). Basement membrane also stained with I3B9 (Fig. 8A). In a 7-day-old embryo, feather rudiments composed of an epidermal placode and dermal condensation were noted. Epidermal placodes formed at the site of the presumptive feather were characterized by vertically elongated epidermal cells. Dermal cells beneath each epidermal placode were progressively condensed and stained with I3B9 (Figs8B, 9A). The epidermal placode, however, only weakly stained (Fig. 8B). This was particularly evident in the centre of an epidermal placode of a 7-5-day-old embryo (Fig. 9A). The decrease in PCS-I in the epidermal placode was also observed in the epidermal sheet of a 7-5-day-old embryo, detached from the basement membrane and dermis by EDTA treatment (Fig. 10). That is, PCS-I decreased more in that particular part of the epidermal placode region where epidermal cells were packed together, than in the region surrounding the epidermal placode (Fig. 10). The basement membrane situated beneath the epidermal placode stained with I3B9 (Figs8B, 9A). ‘

In an 8-day-old embryo, feather rudiments could be seen to start bulging out and develop into feather buds, in which dermal condensation was quite apparent. Feather bud formation brought about changes in the distribution of PCS-I. The apex region of the dermal condensation and basement membrane situated above the apex region of the dermal condensation stained only very weakly with I3B9, as did also the epidermis of the feather buds (Figs 8C, 9B). With the growth of feather buds in a 9-day-old embryo, this area of weak staining expanded so as to encompass all the dermal condensation (Fig. 8D). However, the basement membrane and dermis in the basal region of the feather buds stained with I3B9 (Fig. 8C,D). The epidermis, basement membrane and dermis of the interplumar region showed strong immunofluorescence (Fig. 8B-D).

By day 10·5, the feather buds elongated in a posterior direction and developed into feather filaments. Asymmetrical distribution of PCS-I was noted in these feather filaments. PCS-I was abundant in the epidermis, basement membrane and dermis of the anterior region of the feather filaments (Fig. 8E). The posterior region stained only very weakly with I3B9 (Fig. 8E).

Control sections treated with culture medium and the culture supernatant of I3B9, previously absorbed by PCS-I, showed no fluorescence (not shown). Treatment with testicular hyaluronidase had no significant effect on these staining patterns (not shown).

The turnover of PCS-I in the feather rudiments and buds was determined by pulse-chase experiments with [³⁵S]sulphate. Following the incorporation of [³⁵S]sulphate for lh, the basement membrane situated above the dermal condensation in the feather rudiments as well as the area of the dermal condensation itself were strongly labelled (Fig. HA). Most of the label was eliminated by the chondroitinase ABC digestion, thus indicating it to be virtually all proteochondroitin sulphate (Fig. 11B). After chase for 4h, the label in the basement membrane above the dermal condensation and the area of dermal condensation was found to be very weakly chased (Fig. 11C).

Following the [³⁵S]sulphate incorporation, the feather buds were found to have the same labelling patterns as those of the feather rudiments. That is, the area of dermal condensation and basement membrane above the dermal condensation were strongly labelled (Fig. 12A). However, label in the feather buds was rapidly chased for a period of 2 h (Fig. 12B). After chase for 4h, decrease in the amount of label was conspicuous in the apex region of the dermal condensation and basement membrane above the apex region of the dermal condensation (Fig. 12C). Most of the label disappeared following chase for 6h (Fig. 12D).

The extracellular matrix is considered to perform important functions in morphogenesis. During the formation of chick embryo feather germs composed of epidermal placode and condensed dermis, changes in the distribution profiles of collagen types 1 and III and fibronectin have been reported (Kitamura, 1981; Mauger et al. 1982b>). Sengel, Bescol-Liversac & Guillam (1962) have also demonstrated changes in the continuous labelling patterns of [³⁵S]sulphate during feather formation. The present study provides detailed clarification of the structure and distribution of proteoglycans during feather germ formation and confirms and extends the scope of the studies of Sengel et al. (1962).

The dorsal skin of an 8-day-old embryo synthesized three types of proteoglycans, PCS-I, PCS-II and PHS (Figs 1-4). Lever & Goetink (1976) have observed in the skin of chick embryos two types of proteoglycans differing in molecular weight. In the present study, a proteoglycan fraction with low molecular weight (fraction I) was found to consist of PCS-II and PHS. PCS-I and PGM are proteochondroitin sulphates synthesized by the noncartilageous mesenchymal tissues of chick embryos (Kitamura & Yamagata, 1976; Okayama et al. 1976; Kimata et al. 1986). No analysis was made of the protein core of PCS-I in this study and thus the present author is not in a position to say whether PCS-I is the same as PGM. However, since PGM showed weak cross reactivity with I3B9, it may possibly be a PCS-I similar to proteochondroitin sulphate (Kimata et al. 1986).

PCS-I was biochemically demonstrated to be present not only in dermis but epidermis as well (Figs 5, 6). Since the ectoderm of a chick embryo limb bud produces PGM (Kimata et al. 1986) and the cornea epithelium and epidermis of chick embryos synthesize collagen type I (Linsenmayer, Smith & Hay, 1977; Kitamura, unpublished data), it is possible that the embryonic epithelium and epidermis may also synthesize extracellular matrix components peculiar to mesenchyme at an early stage in chick development. Although the functions of PCS-I in the epidermis remain obscure, PCS-I may possibly contribute to stabilization of the dermis.

PCS-I was immunohistochemically detected in the epidermis of the skin of a 6-day-old embryo and the interplumar skin of 7- to 10·5-day-old embryos (Fig. 8). The ectoderm of a limb bud has also been reported to stain with anti-PGM antibody (Kimata et al. 1986). Although PCS-I in the epidermis is apparently intracellular at microscopical levels (Fig. 10, Vertel, Barkman & Morrell, 1985), the strict localization of PCS-I in the epidermis should be examined immunohistochemically by electron microscopy.

A marked decrease in PCS-I in the epidermal placode, exceeding that in the epidermis surrounding it, was noted following formation of the epidermal placode (Figs8B, 9A, 10). Experimental results of the hyaluronidase treatment of sections ruled out the possibility that this decrease detected by immunofluorescence results from the masking of PCS-I with other molecules such as hyaluronate. Similar local decrease in PCS-I was also noted in the epidermal placode of scale germs (Kitamura, unpublished data). Furthermore, collagen type I also decreased in the epidermal placode of feather and scale germs (Kitamura, unpublished data). It has recently been reported that epithelial morphology influences the amount of collagen produced (Sugrue & Hay, 1986). Although the actual mechanism for this decrease remains to be clarified, morphological changes in epidermal cells during epidermal placode formation may inhibit the synthesis of PCS-I. The decrease in PCS-I in the epidermis of the presumptive feather region may be one of the first indications of feather germ morphogenesis. Furthermore, a conspicuous decrease in PCS-I continued to persist in the epidermis of feather buds and filaments (Fig. 8C-E). Thus, a transition of epidermis appears to occur in the region of feather germ formation from the undifferentiated state. However, a transition would not necessarily indicate overt differentiation of epidermis as represented by β-keratinization (Haake, Kônig & Sawyer, 1984).

It is well known that the basement membrane contains proteoheparan sulphate synthesized by epithelial cells (Kanwar, Hascall & Farquhar, 1981). The distribution of PHS during feather germ formation has not been examined due to lack of an antibody specific for PHS from chick embryonic skin. The basal lamina of mouse salivary gland has been found to contain chondroitin sulphate from epithelium (Cohn, Banerjee & Bernfield, 1977). In this study, PCS-I was identified immunohistochemically in the basement membrane of chick embryo dorsal skin (Fig. 8A,B). PCS-I in the basement membranè of this skin may derive from the dermis since PCS-I was found along the basement membrane under the epidermal placode where PCS-I was noted to decrease (Figs 8B, 9A). Fibronectin in the basement membrane of developing tooth has also been reported to be produced exclusively by mesenchymal cells (Hurmerinta, Kuusela & Thesleff, 1986).

Although structural changes in the basement membrane have been observed during the morphogenesis of various organs (Ekblom et al. 1981; Bernfield & Banerjee, 1978; Thesleff et al. 1980), no temporo-spational changes in the distribution of laminin and collagen type IV have been detected during feather germ formation (Mauger et al. 1982a). PCS-I was also uniformly detected along the basement membrane during the formation of feather rudiments (Figs 8B, 9A). However, its distribution in the basement membrane of feather buds differed from that in the basement membrane of feather rudiments. The decrease in PCS-I was conspicuous in the basement membrane at the top of feather buds (Figs 8C,9B). This decrease was examined autoradiographically. Proteochondroitin sulphate in the basement membrane of the feather buds was less stable than that in the feather rudiments (Figs 11A,C, 12A-D). This lack of stability was similar to that of basal laminar glycosaminoglycans in submandibular morphogenesis (Bernfield & Banerjee, 1982). It should be noted that PCS-I in both the basement membrane and condensed dermis decreased at the same time.

A decrease in collagen type I in the condensed region of dermal cells has been noted to occur soon after the start of dermal condensation (Kitamura, 1981; Mauger et al. 19826). The decrease in PCS-I was different from that of collagen type I. It was immuno-histochemically detected in the area of dermal condensation of feather rudiments (Figs8B, 9A). Autoradiography also showed proteochondroitin sulphate to be synthesized in the area of dermal condensation of feather rudiments and to have a very slow turnover (Fig. 11 A,C). Thus, accumulated PCS-I may possibly be responsible in part for the condensation of dermal cells in the presumptive area of the feather germ. Fibronectin may perform the same function (Kitamura, 1981; Mauger et al. 1982).

Following rudiment protrusion, the decrease in PCS-I started from the top region of the dermal condensation (Figs 8C, 9B) and proceeded to its core region (Fig. 8D). In the region of dermal condensation, the decrease in PCS-I did not occur as a result of its being masked by other molecules, such as hyaluronate, from the hyaluronidase treatment. Pulse labelling with [³⁵S]sulphate showed strong incorporation of [³⁵S]sulphate into the area of dermal condensation and basement membrane situated above it (Fig. 12A). This indicates the high synthetic activity of proteochondroitin sulphate in the dermal cells of the condensed area. The incorporated label, however, subsequently underwent rapid turnover, which was particularly remarkable in the apex region of feather buds (Fig. 11B-D). From these findings along with the results of immunohistochemical and autoradiographic analysis, PCS-I is thus shown to undergo rapid degradation not only in the basement membrane but also in the condensed dermis during feather bud growth.

Mesenchymal cells have been reported to degrade epithelial basal lamina glycosaminoglycans (Smith & Bernfield, 1982). Thus, condensed dermis in feather buds likely retains a set of enzymes that degrade the core protein and glycosaminoglycans of PCS-I in the basement membrane and condensed dermis. Activation of these enzymes may be essential for the protrusion of feather rudiments. Furthermore, it should be emphasized that the decrease in PCS-I in the epidermal placode occurs prior to that in PCS-I in the basement membrane and condensed dermis. This decrease in the epidermal placode may cause the dermis to lose its stability and its subsequent morpho-genetic activation of dermis in the presumptive region of feather germs.

PCS-I was found in the basement membrane and dermis in the basal region of the feather buds (Fig. 8D). It was also present in a large amount in the basement membrane and dermis in the anterior region of feather filaments (Fig. 8E). As assumed for collagen types I and III, PCS-I in feather buds and filaments may also function to maintain the base structure necessary for feather germ protrusion and directional elongation (Mauger et al. 19826). The present data strongly indicate that PCS-I is quite likely involved in epidermal-dermal interactions in the morphogenetically active region of feather germ.

I acknowledge with deep gratitude the support of Drs Y. Kato and T. Higashinakagawa through this study. I wish to thank Dr S. Tanaka for many valuable discussions, Miss M. Sezaki for excellent technical assistance and Mrs Y. Murakami for typing my manuscript.