- Email: [email protected]
Elevation of acidic fibroblast growth factor mRNA in lesioned rat brain
.~oleculrr
El\evier
and Crllulur
Scientific
Endocrrnolog~,
Publishers
Ireland,
58 (1988)
275-278
215
Ltd.
MC-E 01894
Short Communication
Elevation
of acidic fibroblast
growth factor mRNA in lesioned rat brain Ann L,ogan
Deprrrmwnt
of Ph~xolog~,
The Medrcul
(Received
Kn, words;
Fibroblast
growth
factor;
School.
C~nwrrsi~~~of R~rmr~t~hum. Blrmrtl~hum.
23 December
Oligodeoxynucleotide;
1987: accepted
mRN4;
RI.7 .?TJ. l .. K.
29 April 1988)
Brain lesion: (Rat)
Summary A 40-base oligodeoxynucleotide probe is described which has been prepared corresponding to the amino acid sequence 9-22 of acidic fibroblast growth factor. Following electrophoretic separation of rat brain mRNA under denaturing conditions the probe hybridizes to a major polyadenylated mRNA species of approximately 4.8 kb. This mRNA is the same size as that reported for acidic fibroblast growth factor mRNA. The relative abundance of the hybridizing 4.8 kb mRNA species increases in rat brain 3 days after cortical lesion indicating increased expression of the gene for acidic fibroblast growth factor. _
Introduction The structurally and functionally related acidic and basic fibroblast growth factors (FGFs) are now well characterized. Their complete amino acid and nucleotide sequences are known and recently cDNA and genomic clones for both basic and acidic FGF have been isolated (Abraham et al., 1986b). They are known to stimulate a wide variety of mesoderm- and neuro-ectoderm-derived cells in vitro and to be potent angiogenic agents in vivo (for review see Gospodarowicz et al., 1986; Senior et al., 1986: Gospodarowicz, 1987). Although the precise physiological roles of the FGFs have not been established. their spectrum of activity indicates that they could play a crucial role in the wound healing process in the CNS (Berry et al., 1983). Whilst basic FGF has been detected in a wide range of tissues, acidic FGF has been found
Address for correspondence: Dr. A. Logan, Department of Physiology. The Medical School. University of Birmingham, Birmingham. B15 2TJ. U.K. 0303.7207,‘88,‘$03.50
C 1988 Elsevier Scientific
Publishers
Ireland.
almost exclusively in those of neural origin (Baird et al.. 1985; Gospodarowicz. 1987). To date the precise cellular sites of synthesis of the FGFa remain to be established but there is evidence that the growth factor protein is localised in neurones (Pettmann et al.. 1986). This paper reports the synthesis of a 40-base oligodeoxynucleotide probe corresponding to amino acids 9-22 acidic FGF. Using a dried gel hybridization technique the radiolabelled probe hybridizes with a major band of approximately 4.8 kb of electrophoretically separated. polyadenylated mRNA from adult rat brain which corresponds exactly with the size of acidic FGF mRNA. The intensity of the hybridizing signal is shown to increase when mRNA is prepared from rat brain 3 days following cortical lesion. Materials
and methods
Probe prepurution. A 40-base oligodeoxynucleotide was prepared on a Biotech Instruments BT 8500 automatic synthesiser, using the cyanoethyl phosphoamidite protection strategy.
Ltd
Fig. I. Design of oligodeoxynucleotide probe Amino acids I%31 of basic FGF are aligned with amino acids 9-22 of acidic FGF. The nucleotide sequence known to code for this region of acidic FGF is shown. along with that of the synthesised oligodeoxynucleotide probe that 1s complementary to the coding sequence. Where the two amino acid sequences are Identical, the probe contains the same codon of choice as that found in the acidic FGF genomic fragment: at positions where the sequences differ, the probe incorporates the minimum possible number of nucleotide changes needed to allow the amino acid change.
Deprotection and cleavage from the solid phase was done with aqueous ammonia at 50 o C for 5 h. The oligodeoxynucleotide was designed using the homology between amino acids 18-31 of basic FGF and amino acids 9-22 of acidic FGF (Fig. 1). Where the two amino acid sequences are identical the probe contains the same codon choice as that found in the acidic FGF genomic fraction, at positions where the sequences differ the probe incorporates the minimum possible number of nucleotide changes needed to allow the amino acid change. The oligodeoxynucleotide was 5’ endlabelled with “P using T4 polynucleotide kinase (Maniatis et al.. 1982). Lesions ofudult rut brain. Groups of five adult Wistar rats were anaesthetised with halothane. Using a stereotactic instrument unilateral. cortical, knife cut lesions were made in the right cerebral hemisphere to a depth of approximately 3 mm along a line parallel with the sagittal suture. at least 3 mm lateral to the mid-line. Animals were allowed to recover for 3 days after which they were killed, their brains dissected on ice and the lesioned and contralateral unlesioned hemispheres frozen separately at - 80 o C for later RNA extracti0n.
RNA estruction from rut bruin. Total cellular RNA was extracted from brain tissue by the guanidine thiocyanate method (Maniatis et al., 1982). Polyadenylated RNA was isolated from this preparation using at least two cycles of HybondmAP chromatography (Amersham, U.K.).
RNA electrophoresis und dried gel hyhridirutlon. 10 pg per lane of mRNA was denatured with and dimethyl sulphoxide and size glyoxal fractionated by electrophoresis on a 0.5% agarose gel. 18s and 28s calf liver ribosomal RNA and HbrdIII-digested XDNA were used as size markers (Maniatis et al., 1982). Gels were dried under vacuum onto Whatman 3MM paper. The dried gels were then wetted with distilled water to remove the Whatman paper and sealed in plastic bags for direct hybridization with the radiolabelled oligodeoxynucleotide probe described previously. With no prehybridization, hybridization of 15 pmol of probe took place directly to the gel. in 5 ml SSPE (1 x SSPE = 10 mM sodium phosphate, pH 7.0; 0.X M sodium chloride and 1 mM EDTA) containing 0.3% SDS and 10 pg per ml of sonicated. denatured Escherichia coli carrier DNA, at 5O’C for 16 h. The gels were then washed twice for 30 min at room temperature in 2 x SSPE plus 0.1% SDS. once for 15 min at 40” C in 5 X SSPE plus 0.1% SDS and once for 15 min at 50” C in 5 X SSPE plus 0.1% SDS. After air drying the gels were covered with Saran wrap and exposed to Fuji RX X-ray film with DuPont Cronex II intensifying screens for 4 days at - 70°C. Results and discussion A 40-base oligodeoxynucleotide probe has been constructed from the nucleotide sequence of the amino terminal exon of bovine acidic FGF, taking into account the 55% amino acid sequence homology between basic and acidic FGF (Fig. 1) with the object of examining expression of both acidic and basic FGF genes in a variety of situations. Basic FGF is synthesised initially as a 155 amino acid protein which contains an aminoterminal extension of 9 amino acids not found in the sequenced 146 amino acid form of the protein (Esch et al.. 1985). Acidic FGF is remarkably similar comprising a protein of 140 amino acids with a 15 amino acid amino-terminal extension identical to the basic FGF sequence (Abraham et al., 1986a). In various culture cells and tissues the basic FGF gene gives rise to two polyadenylated mRNAs of 3.7 and 7.0 kb. whereas the acidic
FGF gene appears to encode a single mRNA species of 4.8 kb (Abraham et al., 1986b; Jaye et al.. 1986: Gospodarowicz, 1987; Neufeld et al., 1987; Schweigerer et al., 1987a, b. c; Winkles et al.. 1987). These mRNA species are therefore clearly defined and easily separated electrophoretically. Although one might expect the oligodeoxynucleotide probe described to hybridize to both basic and acidic FGF mRNA sequences present in brain preparations, under the conditions described the probe hybridized to a major band of polyadenylated mRNA of 4.8 kb (Fig. 2). The visualisation of this mRNA species solely in the polyadenylated RNA preparations (Fig. 2, lanes 3 and 4) and not in the total RNA preparations (Fig. 2, lanes 1 and 2) indicates that this hybridization is occurring to a specific polyadenylated RNA species and not to any contaminating rRNA which might be present in both preparations. The hybridizing 4.8 kb band corresponds exactly with the reported size of acidic FGF mRNA and suggests that it is to this mRNA that the probe is hybridizing. Previous attempts to examine FGF gene expression in normal rat brain tissue by mRNA hybridization on GeneSceen (New England Nuclear) to specific cDNA probes for basic and acidic FGF have met with difficulty (A.L., unpublished observations). This may be due to the low level and apparent instability of FGF mRNA (Abraham et al.. 1986a) which is in contrast to the quantity of FGF protein that has been purified from several tissues. The availability of easily prepared ohgodeoxynucleotide probes which distinguish between acidic and basic FGF mRNA and a system of dried gel hybridization which gives a good hybridizing signal corresponding to the FGF mRNAs in normal tissue will prove useful analytical tools. Their application will permit studies of the expression of these genes in experimentally manipulated tissues both in vitro and in vivo. They may also allow the precise cellular localisation of FGF synthesis in normal tissue. Such studies are of great relevance to investigations of the physiological roles of FGF. In this context the observation that the relative amount of the hybridizing mRNA species in rat brain is increased in lesioned cerebral cortex was of interest (Fig. 2). It may reflect increased acidic
Fig. 2. RNA electrophoresis and dried gel hybrldiLation of rat brain mRNA. Lanes 1 and 2 show total mRNA from unlesioned and contralateral lesioned cortex respectively. Lanes 3 and 4 show polyadenylated mRNA isolated from both sources. 10 pg mRNA per lane was fractionated on a 0.5% agarosr gel and hybridized directly with the radiolabelled 40-bass oligodeoxynucleotide probe described. Size marhers were 2% and 18s calf liver ribosomal RNA and IluzdIII-digested XDNA.
27X
FGF gene expression following CNS injury. We have previously implicated a role for FGF in wound healing in the CNS (Berry et al.. 1083) and therefore detection of raised levels of acidic FGF-polyadenylated mRNA in lesioned rat brain may be of significance in this consideration. The application of these techniques to further studies of the wound healing process in the CNS will prove valuable. Acknowledgements This work is supported by the Medical Research Council and the International Spinal Research Trust. Special thanks are due to Dr. Kevin Docherty and the Department of Medicine, University of Birmingham, for technical guidance and provision of facilities. References Abraham. 5.4.. Whang. J.L., Turn&, A., man. J., Gospodarowiu. D. and Fiddes. J. 5. 252?--252X. 4hraham. J.A., Mergia, A.. Whang. J.L.. man, J.. H.jrrrild, K..4.. Gospodarowiu. J.C. (1986b) Science 233. 545-548. Baird. A., Esch. F.. Gospodarowicz. D. (19X5) Biochrmibtr\ 34. 7X55-7859.
Mergin. A.. FriedJ.C. (lOXha) EMBO Tumolo. A.. f’ncdD. and Fiddes. and
(;ulllemln.
R.
Bury. M.. Maxwell. W.L.. Logan. A.. Mathewson. A., McC’onnell. P., Ashhurst. D. and Thomaa, G.H. (19X3) Acta Nrurochirurg. Suppl. 32. 31-53. Esch. F.. Bard. A.. Ling. N.. Ueno. N.. H1l1. F.. Denoroy, L.. Klepper. R.. Gospodarowicr. D.. Bohlen. P. and Guillrmin. R. (19X5) Proc. Natl. Acad. Sci. U.S.A. X2. 6507-651 I. C;o‘;podarowicr. D. (19X7) Nucl. Med. Biol. 14, 421-434. Gospodarowicr. D.. Ncufeld. G and Schweigrrrr. L. (19X6) Mol. Cell. Endocrinol. 46. 1X7-204. Jayc. M., Howk. R.. Rurgcss. W.. Rica, GA. <‘hit], I.-M.. RaLern. M-W., O’Brien. %.I.. Modi. W.S.. Ma&g. T. and Drohan, W.N. (1986) Science 233. 541-545. Maniatis, T.. Fritsh. E.F. and Sambrook. J. (19X2) Molecular Manual. Cold Spring llarhor Cloning: A Laborator> Laboratory 1+x. Cold Spring Harhor. NY. Mergia. A.. Eddy, R.. Abraham. J.A.. Fiddes. J.C. and Shows. T.B. (I 9X6) Biochem. Biophys. Reh. Commun. 13X. 644- 65 I. Ncufcld. (i., Ferrara. N.. Schweigerer. L.. Mitchell. R. and Goapodarowicr. I>. ( 19X7) Endocrinology 121. 597.-603. Pettmann. B.. Labourdette. 6.. Wrihrl, M. and Sensenbrenner. M. (19X6) Neuroaci. Lat. 6X. 175%1X0. Schweigerer. I... Nwfeld, G., Fidda, J.C. and Show. T.B. (19X7il) Nature 325. 257-259. Schueigercr. L., Nrufeld. G. and Gospodarowiu, D. (19X7h) Inwat. Ophthalmol. Visual Sci. 2X. 1X38%1X43. Schwelgrrer. L.. Malcrstein, B., Neufrld, G. and Gospodaro%iu, D. (19x7~) Biochem. Biophys. Rea. (‘ommun. 143, Y.34 940. Senior. R.M.. Hung. S.S.. Griffin. C;.L. and Huang. J.S. ( 1986) Biochem. Blophys. Res. Commun. 141. 67.-72. Winkle. J..4., Fricarl. R., Burgess. W.H., Hawk. R.. Mehlman. T., Wrlnatem. R. and Maciag. T (19X7) Proc. Natl. Acad. SCI. I .s.rz. x4. 7124~7128.
El\evier
and Crllulur
Scientific
Endocrrnolog~,
Publishers
Ireland,
58 (1988)
275-278
215
Ltd.
MC-E 01894
Short Communication
Elevation
of acidic fibroblast
growth factor mRNA in lesioned rat brain Ann L,ogan
Deprrrmwnt
of Ph~xolog~,
The Medrcul
(Received
Kn, words;
Fibroblast
growth
factor;
School.
C~nwrrsi~~~of R~rmr~t~hum. Blrmrtl~hum.
23 December
Oligodeoxynucleotide;
1987: accepted
mRN4;
RI.7 .?TJ. l .. K.
29 April 1988)
Brain lesion: (Rat)
Summary A 40-base oligodeoxynucleotide probe is described which has been prepared corresponding to the amino acid sequence 9-22 of acidic fibroblast growth factor. Following electrophoretic separation of rat brain mRNA under denaturing conditions the probe hybridizes to a major polyadenylated mRNA species of approximately 4.8 kb. This mRNA is the same size as that reported for acidic fibroblast growth factor mRNA. The relative abundance of the hybridizing 4.8 kb mRNA species increases in rat brain 3 days after cortical lesion indicating increased expression of the gene for acidic fibroblast growth factor. _
Introduction The structurally and functionally related acidic and basic fibroblast growth factors (FGFs) are now well characterized. Their complete amino acid and nucleotide sequences are known and recently cDNA and genomic clones for both basic and acidic FGF have been isolated (Abraham et al., 1986b). They are known to stimulate a wide variety of mesoderm- and neuro-ectoderm-derived cells in vitro and to be potent angiogenic agents in vivo (for review see Gospodarowicz et al., 1986; Senior et al., 1986: Gospodarowicz, 1987). Although the precise physiological roles of the FGFs have not been established. their spectrum of activity indicates that they could play a crucial role in the wound healing process in the CNS (Berry et al., 1983). Whilst basic FGF has been detected in a wide range of tissues, acidic FGF has been found
Address for correspondence: Dr. A. Logan, Department of Physiology. The Medical School. University of Birmingham, Birmingham. B15 2TJ. U.K. 0303.7207,‘88,‘$03.50
C 1988 Elsevier Scientific
Publishers
Ireland.
almost exclusively in those of neural origin (Baird et al.. 1985; Gospodarowicz. 1987). To date the precise cellular sites of synthesis of the FGFa remain to be established but there is evidence that the growth factor protein is localised in neurones (Pettmann et al.. 1986). This paper reports the synthesis of a 40-base oligodeoxynucleotide probe corresponding to amino acids 9-22 acidic FGF. Using a dried gel hybridization technique the radiolabelled probe hybridizes with a major band of approximately 4.8 kb of electrophoretically separated. polyadenylated mRNA from adult rat brain which corresponds exactly with the size of acidic FGF mRNA. The intensity of the hybridizing signal is shown to increase when mRNA is prepared from rat brain 3 days following cortical lesion. Materials
and methods
Probe prepurution. A 40-base oligodeoxynucleotide was prepared on a Biotech Instruments BT 8500 automatic synthesiser, using the cyanoethyl phosphoamidite protection strategy.
Ltd
Fig. I. Design of oligodeoxynucleotide probe Amino acids I%31 of basic FGF are aligned with amino acids 9-22 of acidic FGF. The nucleotide sequence known to code for this region of acidic FGF is shown. along with that of the synthesised oligodeoxynucleotide probe that 1s complementary to the coding sequence. Where the two amino acid sequences are Identical, the probe contains the same codon of choice as that found in the acidic FGF genomic fragment: at positions where the sequences differ, the probe incorporates the minimum possible number of nucleotide changes needed to allow the amino acid change.
Deprotection and cleavage from the solid phase was done with aqueous ammonia at 50 o C for 5 h. The oligodeoxynucleotide was designed using the homology between amino acids 18-31 of basic FGF and amino acids 9-22 of acidic FGF (Fig. 1). Where the two amino acid sequences are identical the probe contains the same codon choice as that found in the acidic FGF genomic fraction, at positions where the sequences differ the probe incorporates the minimum possible number of nucleotide changes needed to allow the amino acid change. The oligodeoxynucleotide was 5’ endlabelled with “P using T4 polynucleotide kinase (Maniatis et al.. 1982). Lesions ofudult rut brain. Groups of five adult Wistar rats were anaesthetised with halothane. Using a stereotactic instrument unilateral. cortical, knife cut lesions were made in the right cerebral hemisphere to a depth of approximately 3 mm along a line parallel with the sagittal suture. at least 3 mm lateral to the mid-line. Animals were allowed to recover for 3 days after which they were killed, their brains dissected on ice and the lesioned and contralateral unlesioned hemispheres frozen separately at - 80 o C for later RNA extracti0n.
RNA estruction from rut bruin. Total cellular RNA was extracted from brain tissue by the guanidine thiocyanate method (Maniatis et al., 1982). Polyadenylated RNA was isolated from this preparation using at least two cycles of HybondmAP chromatography (Amersham, U.K.).
RNA electrophoresis und dried gel hyhridirutlon. 10 pg per lane of mRNA was denatured with and dimethyl sulphoxide and size glyoxal fractionated by electrophoresis on a 0.5% agarose gel. 18s and 28s calf liver ribosomal RNA and HbrdIII-digested XDNA were used as size markers (Maniatis et al., 1982). Gels were dried under vacuum onto Whatman 3MM paper. The dried gels were then wetted with distilled water to remove the Whatman paper and sealed in plastic bags for direct hybridization with the radiolabelled oligodeoxynucleotide probe described previously. With no prehybridization, hybridization of 15 pmol of probe took place directly to the gel. in 5 ml SSPE (1 x SSPE = 10 mM sodium phosphate, pH 7.0; 0.X M sodium chloride and 1 mM EDTA) containing 0.3% SDS and 10 pg per ml of sonicated. denatured Escherichia coli carrier DNA, at 5O’C for 16 h. The gels were then washed twice for 30 min at room temperature in 2 x SSPE plus 0.1% SDS. once for 15 min at 40” C in 5 X SSPE plus 0.1% SDS and once for 15 min at 50” C in 5 X SSPE plus 0.1% SDS. After air drying the gels were covered with Saran wrap and exposed to Fuji RX X-ray film with DuPont Cronex II intensifying screens for 4 days at - 70°C. Results and discussion A 40-base oligodeoxynucleotide probe has been constructed from the nucleotide sequence of the amino terminal exon of bovine acidic FGF, taking into account the 55% amino acid sequence homology between basic and acidic FGF (Fig. 1) with the object of examining expression of both acidic and basic FGF genes in a variety of situations. Basic FGF is synthesised initially as a 155 amino acid protein which contains an aminoterminal extension of 9 amino acids not found in the sequenced 146 amino acid form of the protein (Esch et al.. 1985). Acidic FGF is remarkably similar comprising a protein of 140 amino acids with a 15 amino acid amino-terminal extension identical to the basic FGF sequence (Abraham et al., 1986a). In various culture cells and tissues the basic FGF gene gives rise to two polyadenylated mRNAs of 3.7 and 7.0 kb. whereas the acidic
FGF gene appears to encode a single mRNA species of 4.8 kb (Abraham et al., 1986b; Jaye et al.. 1986: Gospodarowicz, 1987; Neufeld et al., 1987; Schweigerer et al., 1987a, b. c; Winkles et al.. 1987). These mRNA species are therefore clearly defined and easily separated electrophoretically. Although one might expect the oligodeoxynucleotide probe described to hybridize to both basic and acidic FGF mRNA sequences present in brain preparations, under the conditions described the probe hybridized to a major band of polyadenylated mRNA of 4.8 kb (Fig. 2). The visualisation of this mRNA species solely in the polyadenylated RNA preparations (Fig. 2, lanes 3 and 4) and not in the total RNA preparations (Fig. 2, lanes 1 and 2) indicates that this hybridization is occurring to a specific polyadenylated RNA species and not to any contaminating rRNA which might be present in both preparations. The hybridizing 4.8 kb band corresponds exactly with the reported size of acidic FGF mRNA and suggests that it is to this mRNA that the probe is hybridizing. Previous attempts to examine FGF gene expression in normal rat brain tissue by mRNA hybridization on GeneSceen (New England Nuclear) to specific cDNA probes for basic and acidic FGF have met with difficulty (A.L., unpublished observations). This may be due to the low level and apparent instability of FGF mRNA (Abraham et al.. 1986a) which is in contrast to the quantity of FGF protein that has been purified from several tissues. The availability of easily prepared ohgodeoxynucleotide probes which distinguish between acidic and basic FGF mRNA and a system of dried gel hybridization which gives a good hybridizing signal corresponding to the FGF mRNAs in normal tissue will prove useful analytical tools. Their application will permit studies of the expression of these genes in experimentally manipulated tissues both in vitro and in vivo. They may also allow the precise cellular localisation of FGF synthesis in normal tissue. Such studies are of great relevance to investigations of the physiological roles of FGF. In this context the observation that the relative amount of the hybridizing mRNA species in rat brain is increased in lesioned cerebral cortex was of interest (Fig. 2). It may reflect increased acidic
Fig. 2. RNA electrophoresis and dried gel hybrldiLation of rat brain mRNA. Lanes 1 and 2 show total mRNA from unlesioned and contralateral lesioned cortex respectively. Lanes 3 and 4 show polyadenylated mRNA isolated from both sources. 10 pg mRNA per lane was fractionated on a 0.5% agarosr gel and hybridized directly with the radiolabelled 40-bass oligodeoxynucleotide probe described. Size marhers were 2% and 18s calf liver ribosomal RNA and IluzdIII-digested XDNA.
27X
FGF gene expression following CNS injury. We have previously implicated a role for FGF in wound healing in the CNS (Berry et al.. 1083) and therefore detection of raised levels of acidic FGF-polyadenylated mRNA in lesioned rat brain may be of significance in this consideration. The application of these techniques to further studies of the wound healing process in the CNS will prove valuable. Acknowledgements This work is supported by the Medical Research Council and the International Spinal Research Trust. Special thanks are due to Dr. Kevin Docherty and the Department of Medicine, University of Birmingham, for technical guidance and provision of facilities. References Abraham. 5.4.. Whang. J.L., Turn&, A., man. J., Gospodarowiu. D. and Fiddes. J. 5. 252?--252X. 4hraham. J.A., Mergia, A.. Whang. J.L.. man, J.. H.jrrrild, K..4.. Gospodarowiu. J.C. (1986b) Science 233. 545-548. Baird. A., Esch. F.. Gospodarowicz. D. (19X5) Biochrmibtr\ 34. 7X55-7859.
Mergin. A.. FriedJ.C. (lOXha) EMBO Tumolo. A.. f’ncdD. and Fiddes. and
(;ulllemln.
R.
Bury. M.. Maxwell. W.L.. Logan. A.. Mathewson. A., McC’onnell. P., Ashhurst. D. and Thomaa, G.H. (19X3) Acta Nrurochirurg. Suppl. 32. 31-53. Esch. F.. Bard. A.. Ling. N.. Ueno. N.. H1l1. F.. Denoroy, L.. Klepper. R.. Gospodarowicr. D.. Bohlen. P. and Guillrmin. R. (19X5) Proc. Natl. Acad. Sci. U.S.A. X2. 6507-651 I. C;o‘;podarowicr. D. (19X7) Nucl. Med. Biol. 14, 421-434. Gospodarowicr. D.. Ncufeld. G and Schweigrrrr. L. (19X6) Mol. Cell. Endocrinol. 46. 1X7-204. Jayc. M., Howk. R.. Rurgcss. W.. Rica, GA. <‘hit], I.-M.. RaLern. M-W., O’Brien. %.I.. Modi. W.S.. Ma&g. T. and Drohan, W.N. (1986) Science 233. 541-545. Maniatis, T.. Fritsh. E.F. and Sambrook. J. (19X2) Molecular Manual. Cold Spring llarhor Cloning: A Laborator> Laboratory 1+x. Cold Spring Harhor. NY. Mergia. A.. Eddy, R.. Abraham. J.A.. Fiddes. J.C. and Shows. T.B. (I 9X6) Biochem. Biophys. Reh. Commun. 13X. 644- 65 I. Ncufcld. (i., Ferrara. N.. Schweigerer. L.. Mitchell. R. and Goapodarowicr. I>. ( 19X7) Endocrinology 121. 597.-603. Pettmann. B.. Labourdette. 6.. Wrihrl, M. and Sensenbrenner. M. (19X6) Neuroaci. Lat. 6X. 175%1X0. Schweigerer. I... Nwfeld, G., Fidda, J.C. and Show. T.B. (19X7il) Nature 325. 257-259. Schueigercr. L., Nrufeld. G. and Gospodarowiu, D. (19X7h) Inwat. Ophthalmol. Visual Sci. 2X. 1X38%1X43. Schwelgrrer. L.. Malcrstein, B., Neufrld, G. and Gospodaro%iu, D. (19x7~) Biochem. Biophys. Rea. (‘ommun. 143, Y.34 940. Senior. R.M.. Hung. S.S.. Griffin. C;.L. and Huang. J.S. ( 1986) Biochem. Blophys. Res. Commun. 141. 67.-72. Winkle. J..4., Fricarl. R., Burgess. W.H., Hawk. R.. Mehlman. T., Wrlnatem. R. and Maciag. T (19X7) Proc. Natl. Acad. SCI. I .s.rz. x4. 7124~7128.