The TGF-? superfamily encompasses a group of structurally related growth and differentiation factors which have diverse activities in the regulation of cell growth, differentiation, embryonic induction and morphogenesis in a wide range of cells and tissues. They can be subdivided into (1) the TGF-? subfamily: TGF-?₁, -?₂, -?₃ and -?₅; (2) the activin subfamily: inhibin/activin-?_A, inhibin/activin-?_B and inhibin/activin-?_C; (3) the BMP family: BMP-2, BMP-4, dpp, BMP-5, Vgr-1/BMP-6, GDF-8, BMP-7/OP-1, BMP-8/OP-2, 60A, GDF-5, GDF-6, GDF-7, Vg-1, GDF-1, GDF-3/Vgr-2, dorsalin, BMP-3/osteogenin, GDF-10 and nodal; and (4) divergent genes: Müllerian inhibiting substance, GDF-9, inhibin-, GDNF, screw and lefty.

TGF-?₁ transcripts are first seen in the dental epithelium as the early bud stage extends into the condensing mesenchyme. BMP-4 transcripts are present in presumptive dental epithelium during the initiation of tooth development, and shift to the condensing dental mesenchyme during the bud stage. BMP-2 is expressed in dental epithelium from the early bud stage until the cap stage, when it shifts to the mesenchyme. In human fetal teeth, BMP-4 mRNA appears during the compaction of the dental mesenchyme, and BMP-2 mRNA appears later during tooth development, located in differentiated odontoblasts. BMP-6 is distributed in the early dental epithelium, and later in preodontoblasts and odontoblasts. During the differentiation of pulp cells in vitro, a dynamic temporal expression of BMP-2 and BMP-4 is observed. BMP-2 and BMP-4 are implicated in the induction of tertiary dentine during pulpal wound healing following the surgical amputation of adult teeth in the dog. Treatment with TGF-?₂ antisense produces enlargement of cultured tooth organs and advances their development, suggesting a role for TGF-?₂ in the regulation of tooth size and the stage of development. Activin-?_A mRNA is present in mesenchyme in the tooth bud, and activin-?_A-deficient mice lack lower incisors, hinting at the presumptive role in tooth morphogenesis. BMP-7 transcripts are detected in inner dental epithelium during the cap and bell stage, and in odontoblasts and the stratum intermedium at the bell stage. Three distinct cDNA clones encoding human BMP-2, -4, and -6 have been isolated from human pulp-cell cultures. GDNF is expressed in the tooth bud only, limited to the mesenchyme at the bud stage (embryonic day 14.5). In view of the emerging information on the TGF-? superfamily in tooth development and in pulp, we hypothesized that there might be novel members yet to be discovered. We have employed degenerative oligonucleotide primers and RT–PCR to investigate the presence of novel as well as known members. Although no novel members were detected, we identified several transcripts for BMPs, GDFs and GDNF in rat incisor dental pulp. The study described here is an extension of observations on BMP receptors reported in Toyono et al., 1997.

The apical portion and the entire pulp of the maxillary and mandibular incisors from 24 Long–Evans rats, 3 weeks of age, were isolated as described by Matsuki et al. (1995). The lungs and calvarium were also isolated from three-week-old Long–Evans rats. RNA preparation, RT–PCR and cloning were as described in Toyono et al. (1997). In brief, 10 degenerate primers whose design was based on the conserved amino-acid alignment of activins, BMPs, GDFs and GDNF in mature regions were provided by Dr S. J. Lee, except the primers for GDFs and GDNF. PCR was done with seven sets of primers at 94°C for 1 min, 50°C for 1 min, and 72°C for 2 min for 40 cycles and then by extension at 72°C for 10 min. PCR products of approx. 280 bp were gel-purified and cloned into the pCRII cloning vector (TA Cloning Kit, Invitrogen, San Diego, CA). Bacterial colonies carrying individual clones were picked into 96-well microtitre plates, and a replica was prepared by plating the cells on to nitrocellulose. The replica filters were hybridized to a mixture of probes for inhibin/activin-?_B, BMP-2, -3, -4, -5, -6, -7, and -8 (gift from Dr S. J. Lee, Johns Hopkins University). Hybridizing colonies were used to identify known members of inhibin/activin and BMPs when the RT/PCR amplifications were done with the primers SJL158, SJL159, SJL188, SJL189 and SJL195, each combined with SJL160. Non-hybridizing colonies were used to identify GDFs when the primer combination was YKL51/YKL32, and to identify GDNF when the primer combination was MN21/MN22. The plasmid DNAs were isolated from the overnight bacterial culture and sequenced.

A quantity (2 ?g) of poly(A)⁺RNA fractionated in denaturing agarose gels was transferred to Nytran (Schleicher & Schuell, Keene, NH) for Northern blot analysis. Each blot was probed individually but never reprobed. Probes used for hybridization were for rat TGF-?₁ (gift from Dr Su Wen Qian, NIH, Bethesda, MD), murine TGF-?₂ (Genentech Inc., San Francisco, CA), mouse TGF-?₃ (gift from Dr A. Roberts and Dr M. Sporn, NIH), human inhibin/activin-?_A, inhibin/activin-?_B, human BMP-2 and -4 (Suntory Limited, Osaka, Japan), murine BMP-6 (gift from Dr S.E. Gitelman, University of California, San Francisco), human BMP-7, human BMP-8, mouse GDF-1, GDF-5, GDF-6, GDF-7, and GDF-10 (gift from Dr S. J. Lee), human BMP-3 (Genentech) and mouse GDNF (gift from Dr K. Watabe, Jikei University). The hybridization was as described previously except that the temperature was 60°C.

We have cloned the cDNAs encoding TGF-? superfamily members, activins, BMPs, GDFs and GDNF by the colony hybridization method followed by RT/PCR in rat incisor pulp tissue. Nucleotide-sequence analyses of the isolated plasmids from the hybridizing colonies with the probe mixture of inhibin/activin-?_B and the known BMPs showed that activin-?_B, BMP-2, BMP-4, BMP-7 and BMP-8 had been amplified using the sets of primers SJL189/SJL160, SJL188/SJL160, SJL158/SJL160, SJL159/SJL160 and SJL195/SJL160, respectively. The cDNAs encoding GDF-1, GDF-5, GDF-6 and GDNF were also cloned from the non-hybridizing colonies following PCR amplifications with the primer sets YKL51/YKL32 for GDFs and MN21/MN22 for GDNF.

The predicted amino-acid sequences of inhibin/activin-?_B, BMP-2, BMP-4 and BMP-7, excluding the primer portions in the rat, were completely identified with those previously reported in both rat and mouse. The predicted amino-acid sequences of BMP-8, GDF-1, GDF-5, GDF-6 and GDNF, excluding the primer portions of the rat, had 96%, 97%, 100%, 97% and 94% identity with those of the mouse, showing extensive homology between rat and mouse.

Fig. 1. Comparison of the deduced amino-acid sequences of rat (R) inhibin/activin-?_B, BMP-2, BMP-4, BMP-7, BMP-8, GDF-1, GDF-5, GDF-6 and GNDF with those of mouse (M). The mouse sequences are referenced as follows: inhibin/activin-?_B, BMP-2, BMP-4, BMP-7, BMP-8, GDF-1, GDF-5, GDF-6 and GDNF, in which only amino acids different from those of rat are shown. 3? and 5? primers are underlined.

Northern analysis showed that mRNAs for TGF-?₁, -?₂ and -?₃; inhibin/activin-?_A and -?_B; BMP-2, -4,- 6, -7, -8; GDF-1, -5, -7; and GDNF were expressed in rat incisor dental pulp. However, BMP-3 and GDF-10 mRNAs were not detected in the pulp, although they were expressed in rat lung and calvarial bone, respectively. We do not have an explanation for the lack of transcripts for BMP-3 and GDF-10 in the pulp. The cDNA probes detected an approx. 2.5-kb transcript for TGF-?₁; 4.0-, 4.9-, 5.9- and 6.9-kb transcripts for TGF-?₂, 3.2kb for TGF-?₃; 8.0kb for inhibin/activin-?_A; 4.9kb for inhibin/activin-?_B, 4.0kb for BMP-2, 2.0kb for BMP-4, 3.1 and 5.2kb for BMP-6, 2.7 and 4.6kb for BMP-7, 4.9kb for BMP-8, 3.0kb for GDF-1, 1.9kb for GDF-5, 2.9kb for GDF-7 and 6.6kb for GDNF in the pulp. The cDNA probes for BMP-3 and GDF-10 detected transcripts of 6.4 and 2.8 kb, respectively, in rat lung and calvaria.