Journal of Neuroscience Research 32: 1-14 (1992)

Isolation of cDNA Clones Encoding Rat Glial Fibrillary Acidic Protein: Expression in Astrocytes and in Schwann Cells D.L. Feinstein, G.A. Weinmaster, and R.J. Milner Department of Neuropharmacology, The Scripps Research Institute, La Jolla, California (D.L.F., R.J.M.) and Molecular Neurobiology Laboratory, The Salk Institute, San Diego, California (G.A.W.)

Glial fibrillary acidic protein (GFAP) expressed by astrocytes in the central nervous system (CNS) has been extensively characterized but the molecular identity of related molecules in the peripheral nervous system (PNS) remains unclear. To examine possible structural differences between CNS and PNS GFAP, we have isolated cDNA clones for rat GFAP from both cultured astrocyte and Schwann cell libraries. Nucleotide sequence analysis indicated that the PNS and CNS GFAP clones contained identical coding regions, with a predicted protein product of 430 amino acids. However, the 5’-untranslated region of clone rGFA15, isolated from the Schwann cell library, was longer than that predicted for brain-derived GFAP mRNA. Primer extension analysis of RNA isolated from the RT4-D6 Schwann cell line indicated that the start site for PNS GFAP mRNA lies 169 bases upstream from that used in the CNS. In addition, tryptic peptide mapping of GFAP prepared from cultured astrocytes and Schwann cells revealed one major peptide fragment present in CNS GFAP but absent from PNS GFAP. These results suggest structural differences between GFAP in these two cell types, at both the nucleic acid and protein level, and are consistent with previous observations of immunochemical differences existing between CNS and PNS GFAP. o 1992 Wiiey-Liss, Inc. Key words: intermediate filament, mRNA, transcription initiation INTRODUCTION Glial fibrillary acidic protein (GFAP) is the major subunit of intermediate filaments in mature astrocytes and has been used extensively as a specific marker for these cells in the mammalian CNS (Eng et al., 1971; Bignami et al., 1972). GFAP is similar in structure to other members of the class I11 family of intermediate filament proteins: desmin, vimentin, and peripherin (Osborn and Weber, 1986; Conway and Parry, 1988; 0 1992 Wiley-Liss, Inc.

Steinert and Roop, 1988; Thompson and Ziff, 1989). These molecules all possess highly conserved a-helical central regions, flanked by divergent amino and carboxyl terminal domains (Geisler and Weber, 1982). The variable termini probably confer cell-specific properties upon these molecules and are likely to be the sites to which isotype-specific antibodies have been generated (Liem et al., 1978; Albrechtsen et al., 1984). Although GFAP is predominantly expressed in astrocytes, GFAP-related molecules have been detected immunocytochemically in peripheral glia, including enteric glia (Jessen and Mirsky, 1980, 1983; Bjorklund et al., 1984b; Jessen et al., 1984) and nonmyelinating Schwann cells (Yen and Fields, 1981; Barber and Lindsay, 1982; Dahl et al., 1982; Bjorklund et al., 1984a; Jessen et al., 1984; Field and McMenamin, 1985; Fields and Yen, 1985). GFAP-related antigens have also been detected in nonneural tissues, including the stellate perisinusoidal cells of rat liver (Gard et al., 1985), the lens epithelium (Hatfield et al., 1985), and other cells of the rodent eye (Bjorklund and Dahl, 1985). (For ease of reference, we refer to non-CNS GFAP collectively as “peripheral-type, or PNS-type GFAP” throughout this paper). The exact relationship between CNS-type and PNS-type GFAP molecules is not known. Although polyclonal antibodies directed against brain-derived GFAP recognize a protein in extracts of PNS tissue which is similar in size (49-50 kDa) to that observed in the CNS (Jessen et al., 1984; Fields and McMenamin, 1985; Fields and Yen, 1985; Noetzel and Agrawal, 1985), some peptide mapping studies have suggested the existence of protein heterogeneity (Davison and Jones, 1981; Yen and Fields, 1985). Received October 11, 1990; revised November 27, 1991; accepted December 2, 1991. Address reprint requests to Robert J. Milner, Ph.D., Department of Neuroscience and Anatomy, The Milton S . Hershey Medical Center, P.O. Box 850, Hershey, PA 17033. D.L. Feinstein is now at the Division of Neurobiology, Cornell University Medical School, 411 East 69th Street, New York, NY 10021.

2

Feinstein et al.

The possibility of structural differences between central and peripheral GFAP has been suggested largely by the demonstration that certain monoclonal antibodies that recognize CNS-derived GFAP fail to react with PNS-type GFAP. This list now includes the mouse antibody GFAP-3, raised against human brain GFAP (Albrechtsen et al., 1984) and originally used by Jessen and co-workers to distinguish PNS from CNS antigens (Jessen et al., 1984); mouse antibody A2D4, raised against chicken brain GFAP (Dahl et al., 1984) and used to distinguish CNS- from PNS-type GFAP in studies of rat olfactory nerve (Barber and Dahl, 1987), mouse lens epithelium (Hatfield et a]., 1985), rat liver cells (Gard et al., 1985), and rat iris cells (Bjorklund and Dahl, 1985); the rat antibody 2.2B10, raised against bovine brain GFAP (Lee et al., 1984), which has been demonstrated to stain CNS but not PNS-derived tumor cells (Trojanowski et al., 1984) and more recently to stain rat brain GFAP but not sciatic nerve GFAP (Mokuno et al., 1989); and mouse antibodies Tp-GFAP3 and 4, raised against a fixed cytoskeleton preparation of human glioma cell line U-251MG (Collins and Moser, 1983), which recognize epitopes present on human, rat, and mouse CNS, but not PNS-derived GFAP (Jie et al., 1986). The maintenance of a PNS-CNS distinction across evolutionarily divergent species suggests that the CNS-specific epitope is a functionally important part of the GFAP molecule. The presence of a GFAP-like mRNA in the PNS was first demonstrated in samples of RNA isolated from RT4-D cells, an ethylnitrosourea-induced PNS tumor line (Freeman and Sueoka. 1987). However, those experiments used a mouse GFAP cDNA as a hybridization probe, which required lower stringency washing conditions, and thus could have detected cross-reacting mRNA species. More recently, we have used a rat brain GFAP cDNA clone (described fully in this paper) as a probe to analysis RNA isolated from rat sciatic nerve and showed that a GFAP mRNA species was present under high stringency conditions (Mokuno et al., 1989). However, those analyses would not have detected small differences in the size or structure of the mRNA, such as those due to alternative splicing events. To define the structure of PNS-type GFAP and to characterize possible molecular differences between CNS and PNS-type GFAPs, we have isolated cDNA clones encoding rat GFAP from both cultured astrocyte and Schwann cell cDNA libraries, and have analyzed GFAP from both sources by tryptic peptide mapping. These studies have provided the complete amino acid sequence of rat GFAP and have revealed differences in the structures of both the GFA protein and mRNA in these two cell types.

MATERIALS AND METHODS Tissue Culture Methods Primary cultures of astrocytes were prepared from the cerebral hemispheres of newborn Sprague-Dawley rats (Charles River Breeding Laboratories, Inc., Wilmington, MA) (Booher and Sensenbrenner, 1972). After removal of the meninges, cells were dissociated by repeated passage through 1.7- and 1 .O-mm needles and plated at los cells per 100-mm dish (Falcon 3003) in basal medium (MEM) containing 10% fetal calf serum. Medium was changed every 3 days. After 14 days in vitro, cultures were confluent and harvested for cytoplasmic RNA purification or were maintained for an additional 1 to 4 days in the presence of 1 mM dibutyryl cyclic AMP with no calf serum (Shapiro, 1973; Sensenbrenner et al., 1980; Wu and DeVellis, 1983). Greater than 95% of the cells exhibited immunoreactivity for GFAP (not shown). In addition, RNA blot analysis of astrocyte RNA with oligodendrocyte or neuron-specific probes demonstrated no detectable presence of these other neural cell types (data not shown). Primary cultures of Schwann cells were grown as described (Brockes et al., 1979). The RT4-D6 Schwann cell line (a gift of Dr. N. Sueoka) was maintained in Dulbeccos's MEM containing 10% fetal calf serum, and passaged at a 1:40 dilution when confluency was reached (approximately 3-4 days).

Immunostaining Cells (cortical astrocytes or RT4-D6 cells) were grown on polylysine-coated glass slides as described above. Cells were washed twice in 0.1 M phosphate buffer, fixed overnight in a solution of 4% paraformaldehyde in 0.1 M phosphate buffer at 4"C, rinsed twice with a Tris-saline solution (TS = 0.9% NaCl in 0.1 M Tris-C1, pH 7.50), and incubated sequentially at 22°C in 0.3% Triton X-100 in TS for 15 min; TS, twice for 5 min; a 1:30 dilution of goat serum (GS) in TS for 30 min; and TS, twice for 5 min. The cells were then incubated with anti-GFAP antisera in a solution of 1% GS in TS for 18 hr at 4°C. Rabbit anti-bovine GFAP sera (Dako) and rat anti-human GFAP monoclonal antibody B2.2,, (a gift of Dr. V. Lee) were both diluted 1:1,000 for use. The cells were then washed with 1% GS in TS, twice for 5 min; incubated with a 1 5 0 dilution of goat anti-rabbit or anti-rat biotinylated IgG for 30 min; washed with 1% GS in TS, twice for 5 min; incubated with peroxidase-avidin complex diluted 1:100 in 1 % GS for 30 min; washed in TS, twice for 5 min; and incubated with a solution of 50 mg of 3,3 '-diaminobenzidine and 10 ~1 of hydrogen peroxide in 100 ml of 100 mM Tris-HC1, pH 7.6, for 6 min. After the DAB reaction, cultures were rinsed in distilled

GFAP Expression in CNS and PNS water, dehydrated in alcohols, and mounted with coverslips with permount.

RNA Isolation and Analysis Total cytoplasmic RNA was isolated from the brains of Sprague-Dawley rats by extraction with phenollchloroformiisoamyl alcohol (Lenoir et al., 1986). Cytoplasmic RNA was isolated from cell cultures by the NP-40 lysis procedure (Sambrook et al., 1989) as previously described (Feinstein et al., 1991) or by the guanidine isothiocyanate procedure (Chomczymski and Sacchi, 1987). Enrichment for poly(A)+ mRNA was achieved by chromatography on oligo(dT) cellulose (Aviv and Leder, 1972). RNA blot analysis was carried out as previously described (Lenoir et al., 1986); final washing of the blots was performed in 0.2 X SSC/1% SDS at 65°C for 1 hr. The efficiency of RNA transfer and location of the 28 S and 18 S ribosomal mRNAs was determined by staining the filters with methylene blue (Sambrook et al., 1989). Radioactive probes were prepared by random hexamer priming of isolated DNA fragments (Feinberg and Vogelstein, 1983). cDNA Library Construction A cDNA library was generated from mRNA derived from astrocytes grown for 14 days in vitro and then for an additional 4 days in the presence of 1 mM dibutyryl CAMP. Double-stranded cDNA was synthesized from 2 pg of mRNA using a NotI-oligo(dT) primer and commercial reagents (Boehringer Mannheim) according to established procedures (Gubler and Hoffman, 1983). The blunt-ended double-stranded cDNA was ligated to phosphorylated Not1 linkers (Pharmacia), passed over an Ultrogel U2A column to remove excess linkers, and digested with NotI restriction endonuclease (37°C for 12 hr). The material was passed over a second Ultrogel column. precipitated with isopropanol in 2.5 M ammonium acetate, and ligated to Lambda-Zap phage arms (Stratagene) which had been digested with NotI restriction endonuclease and treated with calf intestinal phosphatase. Recombinant phage molecules were packaged in vitro (Stratagene Gigapack Gold) and titered in LE392 cells. The library had an initial complexity of approximately 500,000 recombinants, with greater than 90% containing inserts as determined by blue/white screening on IPTG and XGAL spread plates. The library was amplified once and stored at -80°C in 7% DMSO. Selected recombinants from this library were recovered in the plasmid bluescript SK minus with use of helper phage R408 following recommended procedures (Stratagene). A cDNA library containing inserts of known orientation was constructed from Schwann cell poly(A) RNA in the bacteriophage vector Lambda-Zap I1 using the UniZap cDNA cloning kit (Stratagene). Recombinant +

3

phage molecules were packaged in vitro using Gigapack Gold I1 (Stratagene), titered in XL1 cells, and amplified once before storage. The initial complexity of the library before amplification was over 6 million. Selected clones were recovered as described above.

cDNA Library Screening Approximately 300,000 recombinants of the amplified astrocyte cDNA library were hybridized with the '2P-labeled 1,260-bp SafI-Hind111 fragment of mouse GAFP cDNA G, (Lewis et al., 1984). The cDNA clone encoding mouse GFAP was a gift of Dr. Sally Lewis. The final washing of the filters was carried out under low stringency conditions (2 X SSC, 0.1% SDS at 42°C). Antibody screening of the same library was carried out as described (Young and Davis, 1983) using a rabbit antibovine GFAP antibody (DAKO) at 1:1,000 dilution. Phage were grown in BB4 cells (Stratagene) and fusion proteins induced by incubation with IPTG. Visualization of positive plaques was accomplished by incubation with '251-labeled protein-A (New England Nuclear) and autoradiography on XAR film. Approximately 1 million recombinants of the Schwann cell cDNA library were hybridized with the 347-bp PvuII fragment of cDNA clone rGFA5, labeled with 32P by random priming (Feinberg and Vogelstein, 1983). The final washing was carried out at high stringency conditions (0.1 X SSC, 1% SDS at 65°C). Determination of the nucleotide sequences of isolated cDNA clones was carried out by the dideoxy chain termination procedure (Sanger et al., 1977). Primer Extension Primer extension analysis of the CNS-type GFAP mRNA was performed by modification of the method of Giorgi et al. (1983). Aliquots (10 pg) of total cytoplasmic RNA were combined with the oligonucleotide, 5'-GGCAGGGAGTGGAGGCGTCATTCGAGAC-3' (complementary to nucleotides 136-163 of the sequence shown in Fig. l ) , labeled with 32P on the 5' terminus. The mixture was heated for 2 min at 90"C, annealed for 60 min at 65°C in 200 mM KCl, and extended with reverse transcriptase for 60 min at 42°C. The reverse transcription reaction was carried out in a final volume of 40 ml in 50 mM Tris-HC1 (pH 8.5), 8 mM MgCI,, 30 mM KC1, 0.5 mM dithiothreitol, 500 p M each dNTP, 50 pgiml actinomycin D, and 600 U/ml AMV reverse transcriptase. The products were analyzed by electrophoresis on 6% polyacrylamide-7 M urea sequencing gels. Primer extension analysis of the PNS-type GFAP mRNA was accomplished by mixing approximately 10 mg of total cytoplasmic RT4-D6 RNA with 400 ng of oligonucleotide 5 '-CCCTGCTTCTGCTGGCTCCT-3 ' (complementary to nucleotides 2-21 of the sequence shown in Fig. 1). The mixture was allowed to anneal for

4

Feinstein et al.

2 min at 65°C in 10 mM Tris-C1, pH 7.5 and 1 mM EDTA, and then placed on ice. The reverse transcription reaction was carried out in a final volume of 20 p1 in 50 mM Tris-HC1 (pH 8.3), 7 mM MgCl,, 40 mM KC1, 1.O mM dithiothreitol, 1 mM each dNTP, 50 Kg/ml actinomycin D, 600 U/ml AMV reverse transcriptase, and 10 mCi [32P]dATP (3,000 Ci/mmol) for 60 min at 42°C. The extended products were purified and then analyzed as described above.

clones rGFA5 and &FA7 were determined using both synthetic oligonucleotide primers and specific deletion constructs. The nucleotide sequences of the cDNAs were identical, except that the rGFA7 sequence extended an additional 260 bp upstream from the 5' end of clone rGFA5 (Fig. 1). Clone rGFA7 contained an open reading frame of 1,285 bp that was unbounded on the 5' side and could encode a protein of at least 429 residues (Fig. 2A). The strong similarity of this sequence to the corresponding region of the mouse GFAP gene (Balcarek and Isolation and Analysis of GFAP Cowan, 1985) at both nucleotide and amino acid levels Cultures of primary rat brain astrocytes and (Fig. 3 ) indicated that cDNA clones rGFA5 and rGFA7 Schwann cells were radiolabeled for 24 hr in methionine- encode rat GFAP. In addition, this sequence matched free Dulbecco-Vogt modified Eagle's medium, supple- (with one nucleotide difference at position 427) the parmented with 10% fetal calf serum, 20 pg/ml CM-GGF tial nucleotide sequence of a cDNA clone of rat GFAP (Lemke and Brockes, 1984), 2 mM forskolin, and 100 reported by Nichols and colleagues (Nichols et al., pCi/ml [35S]methionine (1,000 Ci/mmol, Amersham 1990). Corp.). Radiolabeled cells were lysed in RIPA buffer (Weinmaster et al., 1984) and [35S]methionine-con- Expression of GFAP mRNA The GFAP cDNA hybridized strongly to a single taining GFAP was isolated by immunoprecipitation with rabbit antibodies against bovine GFAP (DAKO). Radi- band of approximately 2.8 kb in RNA samples from both olabeled proteins were resolved on one-dimensional whole brain and cultured astrocytes (Fig. 4), as has been gels, extracted from the gel, and digested with trypsin. reported previously (Rataboul et al., 1988). A very weak The digests were separated in two dimensions on cellu- signal was also present at approximately 2.0 kb but this lose thin-layer plates by electrophoresis in the first di- band is probably the result of nonspecific hybridization mension and chromatography in the second dimension, to ribosomal RNA (Tardy et al., 1989). Following treatas described previously (Weinmaster et al., 1984). Ra- ment of cultured astrocytes with dibutyryl cAMP for 4 diolabeled peptides were detected by autoradiography on days, there was a large increase in the relative abundance X-ray film at -70"C, following treatment with EnHance of the GFAP mRNA, consistent with the increase reported for the protein itself (Shapiro, 1973; Sensenbren(Amersham). ner et al., 1980; Wu and De Vellis, 1983). The increase in mRNA suggests that changes in gene expression acRESULTS company the morphological effects of elevated cyclic Isolation of cDNA Clones Encoding AMP levels. Astrocyte GFAP An RNA of similar size was also detected by GFAP Total cytoplasmic RNA was prepared from primary cDNA probes in samples from cultured Schwann cells cultures of rat brain astrocytes that had been treated with (Fig. 5 ) . Probes corresponding to the 5' and 3' portions I mM dibutyryl cAMP for 4 days prior to RNA isolation, of the coding region each hybridized to a 2.8-kb mRNA a treatment known to increase significantly the level of in both astrocytes and Schwann cells. The 3'-probe also expression of GFAP (Shapiro, 1973; Sensenbrenner et detected a predominant mRNA species of 2.3 kb in al., 1980; Wu and DeVellis, 1983). A cDNA library Schwann cells as well as in astrocyte samples after proconstructed from this RNA was hybridized with a ra- longed exposure times (not shown). The close sequence diolabeled fragment of the mouse GFAP cDNA clone similarity of GFAP to vimentin, whose mRNA is (Lewis et al., 1984) under reduced stringency. Six inde- roughly 2.3 kb in size (Quax et al., 1983), suggests that pendent clones were obtained and characterized by re- this smaller band corresponds to the rat vimentin mRNA. striction endonuclease digestion: the largest clone (rGFA5) contained a cDNA insert of approximately Isolation of cDNA Clones Encoding GFAP From 2,500 bp (Fig. I). The same library was screened for Schwann Cells A cDNA library derived from cultured rat Schwann clones that expressed GFAP antigenic determinants, using a rabbit antibody against bovine GFAP. This resulted cells was screened for clones encoding GFAP. Approxin the isolation of clone rGFA7, which contained an imately 1 million recombinant phage were hybridized insert of over 2,700 bp and a similar pattern of restriction with the 347-bp PvuII fragment of rGFA5 to yield nine positive clones. Digestion of these clones .with the enenzyme sites as rGFA5 (Fig. 1). The nucleotide sequences of the cDNA inserts in zyme S a d , which cleaves once within the GFAP se-

GFAP Expression in CNS and PNS bp

200

400

600

800

5’ probe

sc P

rGFA5 rGFA7 rGFA 75

1000

1200

1400

1600

5

2700

3. probe P

s

Ps

P

P Ps

------

Fig. 1 . Structure of rat GFAP cDNA clones. The cDNA inserts from clones rGFA5, 7, and 15 are shown: the hatched areas indicate the coding region and the unshaded areas indicate 5’ and 3’-untranslatedregions. Common restriction endonuclease sites are also shown (Ps, Pstl; P, PvuII; Sc, S a d ; S, StuI). The bars indicate the positions and sizes of cDNA fragments used as hybridization probes for RNA blots. quence and once in the vector polylinker, indicated that the mouse (23/25 bases identical) (Balcarek and Cowan, the 5’ end of the insert in clone rGFAl5 was 20-30 bp 1985) and human (20/25 bases identical) (Brenner et al., longer than that of clone rGFA7 (Fig. 1). Apart from this 1990) GFAP genes, confirming the fact that this region difference, the cDNA inserts of clones rGFA7 and was derived from a GFAP mRNA, and was not due to rGFA15 had identical digestion patterns with over 20 cloning or sequencing artifacts (Fig. 2B). different restriction endonucleases, suggesting that their The deduced amino acid sequence of rat GFAP corresponding mRNA species are products of the same (Fig 2A) is very similar to the predicted sequences of gene. To rule out the possibility that alternative splicing mouse (Lewis et al., 1984; Balcarek and Cowan, 1985) could produce highly similar but distinct mRNAs, the and human GFAP (Reeves et al., 1989; Brenner et al., complete coding sequence as well as the terminal 250 bp 1990) and the partial amino acid sequence of pig GFAP [extending into and including the poly(A)tail] of clone (Geisler and Weber, 1983) (Fig. 3). This analysis asrGFA15 were determined. These regions (a total of ap- sumes the single base error described (Brenner et al., proximately 1,585 bp of rGFA15 were sequenced) were 1990) in the published nucleotide sequence of the mouse identical to the corresponding DNA sequences deter- GFAP gene (Balcarek and Cowan, 1985). Rat and mouse mined for clone rGFA7. We conclude that clone rGFA 15 GFAP differ at only 19 of 430 residues (96% identity, differs from clone rGFA7 solely by the addition of 30 bp including a one residue gap inserted into each sequence for best alignment) and the rodent sequences are each at its 5’ end. slightly more divergent from the human (91%) identity) Sequence of Rat GFAP mRNA and Protein and pig (88% identity) GFAPs. With a few exceptions, The nucleotide sequences obtained from clones the amino acid differences among these sequences are rGFA5, rGFA7, and rGFA 15 were combined to provide conservative changes and none of the changes alters the a composite sequence of 2,692 bp for rat GFAP mRNA heptad repeat motif of the central rod domain that is (Fig. 2A). This sequence contains an open reading frame characteristic of intermediate filaments (Geisler and Weof 1,290 bp, encoding a protein of 430 amino acid res- ber, 1983; Conway and Parry, 1988). A substantial proidues, preceded by a 5’-untranslated region of 25 bp. The portion of the differences are clustered within the amino3’-untranslated region is 1,374 bp in length, terminating terminal 40 residues that form the head region of the in a stretch of at least 70 adenosine residues. The se- molecule. However, the initial amino-terminal segment quence of the 3’-untranslated region for rat extends 30 (residues 1-16) is more highly conserved among the four bases further than the sequence published for mouse species, suggesting a functional role for this part of the GFAP (Lewis et al., 1984), and includes within this protein. segment a polyadenylation signal 22 bases upstream from the poly(A)+ tail. The 5’-untranslated region of rat Determination of Transcriptional Start Sites GFAP mRNA was defined solely by the nucleotide sePrimer extension analysis (Fig. 6A) of RNA dequence of clone GFA15, derived from Schwann cells. rived from either cultured astrocytes or whole rat brain This sequence is similar to the corresponding regions of indicated that the transcription of the GFAP gene in the

A

MetGl uArgArgArgI1eThrSerAlaArgArgSerTyrAlaSerSerG1uThrMetValArg

caggagccagcaGAAGCAGGGCAAGATGGA~GGAGACGTATCACCTCTGCACGCCGCTCCTAT~CTCCTCCGAGACGAT~TCA~ 88

GlyHisGlyProThrArgHisLeuGlyThrIleProArgLeuSerLeuSerArgMetThrProProLeuProAlaArgValAspPheSer G G C C A T G G T C C T A C C A G A C A C C T G G G T A C C G C G C C

,78

LeuAlaGlyAlaLeuAsnA1aGlyPheLysG 1 uThrArgA1aSerGluArgAlaG1uMetMetG1uLeuAsnAspArgPheA1aserwr CTGGCCGGG(;CGCTCAATGCCGGCTTCAAAGt~GACTCGGGCCAGCGAGCGCGCGGAGATGATGGAGCTCAATGACCC~~T~GCTAGCTAC 268

I1eG1uLysValArgPheLeuG1uGlnGlnAsnLysAlaLefrA1 aAlaGluLeuAsnGlnLeuArgAlaLysG1uProThrLysLeuA1a ATCGAGAAGGTCCGCTTCCTGGMCAGC~~CAAGGCGCTGGCA~TGAGCTGMCCAGCTTCGAGCCAAGGAGCCCACC~CTGGCT

358

AspVal WrGlnA2 aG1 uLeuArgG1 uLeuArgLeWgLeuAspG1nLeuThrThrAsnSerAlaArgLeuG1uVa1G1uArgAspAsn GACGTTTACCAGGCAGAACTTCGGA~T~(~~TGCGTCTGCTACCAACAGTGCCCGGCTGGAGGTGGAGAGGGACAA 448 T

LeuThrGlnAspLeuGlyThrLeWgGlnL!/sLeuGlnAspGl uThrAsnLeu4rgLeuGluAlaG1uAsnAsnLeuA1aVa1TyrArg C T C A C A C A G G A C C T C ~ A C C C T G A G G C A G A n G C T C C A A G A 538

GlnGluAlciAspG1 uA1aThrLeuAlaArgValAspLeuG1 uArgLysValG1uSerLeuG2 uG1uG2uIleGlnPheLeuArgLysI1e C A G G A G G C G G A T G A A G C C A C C T T G G C T C G T G ' r G G A T C T G T C

628

HisG1uG1 uG1uVa1ArgG 1 uLeuGl nG1 uG 1 nLeuA1aG1nG1nG1nVa1Hi sVa1G1UMetAsp Va1A1aLysProAspLeuThrA1a CATGAGGAGGAAGTTCGAGAACTCCAGGAGCRGCTGGCTG~CCAGCAGCAGGTCCACGTGGAGATGGATGTGGCCMGCCAGACCTCACA~G718

A l a LeuArgGluIleArgThrG1nTyrG1uA 1aValA1aThrSerAsnMetGlnGluThrGluG1uTrpTyrArgSerLysPheAlaAsp ~TCTGAGAGAGATTCGCACTCAGTACGAGG(IAGTGGCCACCAGTAACATGCAAGAAACAGAAGAGTGGTGGTATCCAAC

808

LeuThrAspVa1A1aSerArgAsnAla G1 uLeuVa1ArgGlnA1aLysHisG1 uA 1aAsnAspTyrArgArgG1nLeuGlnAlaLeuThr CTCACAGACGTTGCTTCCCGCMCGCAGAAGCTCGTCCGCCAGGCCMGCACGAGGCT~~GACTATC~C~~CGCAACT~A~CTTGACC 898 WsAspLeuGlUSerLeuArgGlyThrAsnG 1 uSerLeuGluArgGlnMetArgG1uGlnG1uG2uArgHisAlaArgG1uSerAlaSer TGCGACCTTGAGTCCTTGCGCGGCACGAACGAGTCCTTGGAGAGGCAAATGCGCGMCAGGAGGAGCGCCACGCTC~AGTCGGCCAGT 988

TyrGlnG1uAl aLeuAlaArgLeuG1uG1 uG-1uGlyGinSerLeuLysG1uG1 uMetAlaArgHisLeuGlnG1uQ7rGlnAspLeuLeu TACCAGGAGGCACTCGCTCGGCTGGAGGAGGAGG~CAAAGCCTCAAGGAGGAGATGGCCCGCCACCTGCAGGAGT~~CCAGGATCTACTC 1078

AsnValLysLeuAlaLeuAspIl eG1uIleA.1aThrTyrArgLysLeuLeuG1uGlyGluG1uAsnArgI1eThrIleProVa1GlnThr MCGTTAAGCTAGCCCTGGACATCGAGATCG(ICACCTACAGGAAATTGCTGCTGGAGGGCG~G~CCGCATCACCATTCCTGTACAGACT

1168

PheSerAsnLeuG1n I1eArgG1 uThrSerLeuAspThrLysSerVa1SerGluGlyHisLeuLysArgAsnI1eVa1Va1LysThrVa1 TTCTCCAACCTCCAGATCCGAG~CCA~~~GGACACCAAATCTGTGT~AG~G~CACCTC~GAGG~CATCGT~T~GACGG~ 1258

G1 uMetArgAspGlyGluVal I1eLysGluSerLysGlnG1uHisLysAspVa1Met* ** GAGATGCGGGATGGCGAGGTCATTAAGGAGT(IGAAGCAGGAGCAGGATGTG

1348

C~;GTGTGAGGGCCTAAAGCTCCCTCCTCAGA'CAGTCTTGTTTGCTAGGCCC~TTCCCATC~ACACCAGT~TCCCCTT~TTCTGTTTT 1438 AGTGCCC A C G G C T C G G T C A G T G C G G A G T C T C R T G G A C G G C C 1528 AGGAGGGGAGG~CCCCCCTCCCCCAGCTGGT~~AGAACTGG~GAGG~GACAGGGG~A~GAGACTTAACAAATCCCTTCCTTCA 1618 CTTCTTGTTCTATGGAAACCGTTGCCAGAGCTGGAGGTCTCTGGG~CTGG~CTTTTGMG~TTTCATAG~ATGCTGGA~GCAAGAC~ 1708

CATTCAGACAGAAAGGAAAAGATCCCGAGGCI;AAGAATCTCTAGCCAGAGGCCTAGGCATCTGGAAGAACC~TGACGATGTAGGAGTGGG

1798

TAGGGCAGACTTGCTACCTGG~T~CACT~~AGGCAGTCCTG~G~CC~CCTCTTGAG~ATGACCCTC~TGTATCGGCCCACTGA~ 1888 A

G

C

C

C

T

G

C

A

G

G

T

T

G

A

T

G

C

C

A

G

A

~

G

T

G

T

G

~

~

C

T

T

G

1978

G

GAGGGTATAGAGGAGGCTCTCTGGCCCCTAA~TATGGATGAGTCGAGAGGGGGGAGCCCAG~AAGGCTACCCCGCTCAGGCTGCAGGGGT 2068 GCCATGGCGGA~GAACCGGTGGAGAT~CTT~GACAATGG~.~TTGGAAGCGGTGTGTAGG~CTAGTTACTCTTGGCTCTG~TCCTTGGA 21 58 A T C M G G A A A T G A C C T G T C T C T C A A A G A C C T ~ A A A C A G G G2248 GTCTATCCTGGTTGCTCAGTCECAACTGCGCATCACCCTGGGCTTCTCAACCTGGAGTC.ACAACCATCCTTCT(;AGGC~CCATCCCA 2338 CAACCACTAGCTGTTGTTCTCTAGCCAAGGC(~CCATTCCCTTTCTTATGCATGTACGGAGTI;TCGCCTAGA~TTAA~~GTC~ATCCTGTT 2428 TGAAGTTGGGCAA~'TC,ACACGTTGTGTTCAAGCAGCCTGGTGTGGAGT~CTT~GTATTAGTG~ACCCTCTCGGAAGCTGGTTGGTGGGCAG2518 G'~'GAGG~G~T,TGGAGCTGAAAGTGTCCCC~~CAGTTGTCCTTTCCTCCCCC.TCTAAGGTCCCTCCTTTTCCCCAGGACATCGTACACTC 2608 C C C C C C T T G T C A C C T C T G C T A A C C T T C A G A G ( ~ A G T A C T G T C ~ C C T T T A C T C ~ C T ~ G C A G G ~ T ~ G A C ~ G T G T C A 2698 GA~T~ Fig. 2A.

GFAP Expression in CNS and PNS

7

sion experiments using RNA from adult rat sciatic nerve and with the same oligonucleotide used to determine the CNS start site were unsuccessful, yielding multiple extended products. This may have been due to the limited Fig. 2. A: The nucleotide and derived amino acid sequence of amount of RNA available and to a high degree of seccDNAs for rat GFAP mRNA. The sequence shown is compiled ondary structure present in the 5’-untranslated region refrom the sequences of cDNA clones rGFA5 (positions 291gion of the mRNA (see Discussion). To overcome these 2,698), rGFA7 (positions 31-2,698), and rGFAl5 (positions I-1,332). Note that the first 12 nucleotides (shown in lower difficulties, we isolated larger quantities of RNA from case) are present only in the cDNA isolated from Schwann the RT4-D6 Schwann cell line. The RT4-D6 cell line, an cells, while the transcriptional start site in astrocytes (Fig. 6) ethyl nitrosourea induced PNS tumor line (Freeman and and mouse brain (Balcarek and Cowan, 1985) is located at Sueoka, 1987) expresses high levels of GFAP in culture position 13. The 3’-untranslated region was sequenced in only as demonstrated by immunostaining with a rabbit polyone direction. B: Alignment of the 5’-untranslated regions and clonal antibody, as do cultured cortical astrocytes (Fig. corresponding genomic sequences of human, mouse, and rat 7A,C). However, when stained with monoclonal antiGFAP. body 2.2B,,, which distinguishes CNS from PNS type GFAP (Mokuno et al., 1989), positive staining was seen CNS starts at nucleotide 13 of the sequence shown in only with the astrocytes (Fig. 7B,D). Hence, the RT4Figure 1 , resulting in a 5’-untranslated region that is 12 D6 cell line maintains expression of the PNS type of bases in length. This is identical to the initiation site GFAP in vitro, and was therefore selected for further determined for transcription of the GFAP gene in mouse mRNA studies. The results of primer extension analysis of RT4-D6 brain (Balcarek and Cowan, 1985) and one nucleotide different from the start site of the human GFAP gene derived RNA are shown in Figure 6B. The primer used (Brenner et al., 1990). in this experiment was complementary to bases 2-21 of The extended 5’-untranslated region region of clone rGFA15, hence it will not hybridize to the CNS cDNA clone &FA1 5 suggested that the transcriptional mRNA species which lacks bases 1-12. A major exstart site for PNS-type GFAP mRNA was located up- tended product of approximately 178 bases was synthestream from that used in the CNS. Initial primer exten- sized by the reverse transcriptase with this primer when

B

e ,

............

Rat CAGGAGCCAGCA GAAGCAGGGC AAG Mouse ..G.e...oe .Go Human G... .GC*..A.o .Go

RAT MOUSE HUMAN PIG

73

RAT MOUSE HUMAN PIG

148

RAT MOUSE HUMAN

223

RAT MOUSE HUMAN PIG

298

RAT MOUSE HUMAN PIG

373

RAT

I T I P V 0 T F S N L 0 I K E’T S L D T K S V S E :1 H L K K N I V V Ic T V EM R D G E V I K E S K C L H F L)V * M

429

.~

Fig. 3. Amino acid similarities between rat, mouse, porcine, and human GFAP molecules. The predicted amino acid sequences derived from rat (this paper), mouse (Lewis et al., 1984), and human (Reeves et al., 1989)GFAP cDNAs, as well as the partial sequence determined by protein chemical methods for pig GFAP (Geisler and Weber, 1983). Only residues

that differ from rat are shown; absent residues are indicated by 0 . The sequence of mouse GFAP between positions I and 22 was derived from the published genomic sequence for mouse GFAP (Balcarek and Cowan, 1985), assuming a single base deletion at position 326 of the mouse sequence (Brenner et al., 1990).

8

Feinstein et al.

1

2

3

28s +

18S+

Fig. 4. Expression of GFAP mRNA in cultured astrocytes. RNA samples (10 pg total cytoplasmic RNA) were separated by electrophoresis, transferred to nitrocellulose, and hybridized with a "P-labeled probe corresponding to the 5'-region of clone rGFA7 (positions 31-91, Fig. 2). RNA samples: postnatal day 25 whole brain (lane 1); astrocytes cultured for 14 days in vitro (lane 2); astrocytes cultured for 14 days in vitro and treated with dibutyryl CAMPfor 4 days (lane 3). the RT4-D6 RNA was used. Low amounts of a smaller, approximately 100 base, extended product were sometimes observed, as well as much higher molecular weight (