- Email: [email protected]
Oligodendrocyte Transmembrane Protein: A Novel Member of the Glutamate-Binding Protein Subfamily
Biochemical and Biophysical Research Communications 283, 900 –907 (2001) doi:10.1006/bbrc.2001.4859, available online at http://www.idealibrary.com on
Oligodendrocyte Transmembrane Protein: A Novel Member of the Glutamate-Binding Protein Subfamily Sara Szuchet, 1 David C. Plachetzki, and Kristen S. Eaton Department of Neurology and Brain Research Institute, University of Chicago, Chicago, Illinois 60637
Received April 9, 2001
Oligodendrocytes (OLGs) are cells from the central nervous system that synthesize, assemble, and maintain myelin, the multilamellar membrane that surrounds axons and facilitates the fast conduction of nerve impulses. We have shown that OLGs initiate their myelinogenic phenotype upon adhesion to GRASP, a heparin-binding glycoprotein that we purified from horse serum. In an attempt to identify the genes implicated in establishing this phenotype, we isolated a novel 3500 bp cDNA related to, but distinct from, a subfamily of glutamate-binding proteins (GBP). The cDNA encodes a protein of 511 amino acids, whose predicted sequence can be modeled as a tetrahelical integral protein with a large external loop and with the N- and C-termini located inside the cell. We have named this protein oligodendrocyte transmembrane protein (OTMP). Transcription of the message is induced upon OLG acquiring a myelinogenic phenotype (i.e., upon adhesion). The temporal expression in conjunction with the structural and biochemical features of OTMP is suggestive of a signaling receptor with a role in myelinogenesis. © 2001 Academic Press Key Words: central nervous system; glial cells; myelination; regeneration; signaling receptors; heparan sulfate proteoglycans; adhesion-induced gene transcription; Src homology-binding proteins.
Oligodendrocytes (OLGs) are cells from the central nervous system that synthesize, assemble and maintain myelin, the multilamellar membrane that surrounds axons and facilitates the fast conduction of nerve impulses. Little is known about the molecular and biochemical pathways that direct OLGs to enwrap axons with myelin. Similarly, little is known about the potential of OLGs to regenerate following the destruction of myelin, such as occurs in, e.g., multiple sclerosis. In an attempt to shed some light on these pro1 To whom correspondence should be addressed at Department of Neurology, MC 2030, 5841 South Maryland Avenue, Chicago, IL 60637. Fax: (773) 702-9076. E-mail: [email protected].
0006-291X/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
cesses, we have developed a model, albeit idealized, consisting of pure cultures of ovine OLGs isolated from young, myelinated brains (1). Using this model, we have shown that OLGs recover from the trauma of separation from myelin and reassemble multilamellar membranes—a process we have named myelin palingenesis, i.e., remyelination (regeneration) (2). Moreover, as remyelination by OLGs recapitulates the developmental steps of myelination, this in vitro model can also be utilized to study the mechanism of myelination (3, 4). In fact, we have discovered that the state of OLG differentiation (or regeneration) can be modulated in our model by facilitating or interfering with their attachment to a substratum consisting of a heparin-binding glycoprotein, named GRASP, that we purified from horse serum (5). Not only can this model be used to study the morphological and biochemical changes that occur in OLGs upon adherence to a substratum, but also to identify alterations in the expression of genes that are important for (re)myelination. Indeed, taking advantage of such a strategy and using techniques such as differential display and cDNA library screening, we have identified several genes which are upregulated upon OLG adhesion to GRASP—an event that marks the commencement of their myelinogenic phenotype (6, 7). We have characterized several of the genes implicated in establishing this phenotype (6). We have also demonstrated that adhesion induces OLGs to assemble a cell-associated matrix and have characterized the major matrix constituents as heparan sulfate proteoglycans (HSPGs) (7). Because these HSPGs are very minor components, in an attempt to establish their function, we investigated whether they belong to families of well-characterized HSPGs such as syndecans, perlecan or glypicans. However, when we screened an OLG cDNA library with an oligonucleotide probe designed to detect syndecans, we isolated a novel, 3500 bp cDNA, unrelated to syndecans but related, in lieu, to a glutamate-binding protein (GBP). GBP was first isolated bochemically as a complex of four or more proteins (8). Although the original
900
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
interpretation, and hence the name, that this was the glutamate-binding subunit of the N-methyl-Daspartate (NMDA) receptor did not hold, the finding that antisense oligonucleotides directed against the GBP cDNA block L-glutamate and NMDA-induced Ca 2⫹ influx and neurotoxicity in hippocampal neurons and in cerebellar granule cells supports a receptor role for GBP (9). A number of other cDNAs that share conserved domains with GBP have now been identified and sequenced (10 –12). Herein we demonstrate that the isolated cDNA encodes a novel protein of 511 amino acids, whose predicted sequence can be modeled as a tetrahelical integral protein with a large external loop and with the Nand C-termini located inside the cell. We have named this protein oligodendrocyte transmembrane protein (OTMP). Transcription of the message is induced upon OLG acquiring a myelinogenic phenotype. The temporal expression in conjunction with the structural and biochemical features of OTMP is suggestive of a signaling receptor with a role in myelinogenesis. METHODS Oligodendrocyte cultures. OLGs were isolated from 3- to 6-month-old lamb brains according to our standard procedure (1) and maintained in culture as described by Szuchet and Yim (13). Freshly isolated OLGs were plated on tissue culture petri dishes, where they do not adhere but form floating clusters. After 3–5 days, nonadherent OLGs were harvested, centrifuged, resuspended in medium (DMEM plus 20% horse serum, 2 mM L-glutamine and antibiotics) and seeded on polylysine-coated petri dishes; cultures were kept at 37°C. Alternatively, OLGs were adhered on plates coated with GRASP after precoating with polylysine (5) and maintained in a synthetic medium without serum. Construction and screening of an OLG cDNA library. We followed the protocols provided by the vendors (Stratagene, La Jolla, CA) to generate a unidirectional cDNA expression library from adhered OLG in the Uni-ZAP XR vector. This vector can accommodate DNA inserts from 0 to 10 kb. The titer of the packaged and unamplified library is 2.1 ⫻ 10 6 pfu/ml. The titer of the amplified library is 1.15 ⫻ 10 10 pfu/ml (6). We synthesized a set of degenerate oligonucleotide primers based on the syndecan cytoplasmic sequence YRMKKKDEGSY that is 100% conserved across species (14). Primers were end-labeled with [␥- 32P]dATP and used for library screening. We screened a total of 1.4 ⫻ 10 6 plaques. Screening of the library was performed following the procedure given by Sambrook et al. (15), modified to incorporate recommendations by the vendor NEN Research Products (Boston, MA). To obtain single clones, each positive plaque was subjected to three rounds of screening. Clones were identified by Southern blot analysis with individual probes. Southern blot analysis. Plasmids were purified; the insert cDNA was excised with restriction enzymes EcoRI and XhoI, resolved on a 1% agarose gel, transferred onto a membrane and hybridized with the same [␥- 32P]dATP end-labeled primers utilized for library screening. Northern blot analysis. Total RNA from both floating and attached OLGs were isolated following the single step procedure described by Chomczynsky and Sacchi (16). Poly(A) ⫹ RNA was obtained with the aid of an mRNA isolation kit (Qiagen, Chatswath, CA). mRNA originating from adhered and non-adhered OLGs were resolved on an agarose-formaldehyde gel, transferred to a nylon
FIG. 1. Agarose gel electrophoresis of clone 75-5-1. The insert cDNA was excised with EcoRI and XhoI, resolved on a 1% agarose gel. Lanes: 1 and 2, mol. wt. std. (digests of 174 and DNA, respectively); 3, digest of clone 75-5-1 arrowheads point to the two fragments of insert cDNA; arrow indicates pBluescript; 4, pBluescript without insert (control); 5, excised cDNA hybridized with the same probe employed for screening the library.
membrane, and hybridized with a 600 bp cDNA fragment, obtained by cutting the isolated cDNA with EcoR1 and BsrG1, according to a protocol that was based on steps taken from Sambrook et al. (15) and from the manufacturer’s recommendations for GeneScreen. Sequencing. The BigDye Terminator kit was used with an ABI 377 DNA Sequencer.
RESULTS AND DISCUSSI0N Isolation and Characterization of Clone 75-5-1 In an attempt to molecularly characterize surfaceassociated HSPGs that are induced upon OLGs adhering to GRASP (5, 7), we investigated whether HSPGs implicated in the differentiation of other cell types play any part here. Hence, we screened the cDNA library from adhered OLG with oligonucleotide probes complementary to conserved sequences of syndecan, perlecan and glypican—the most widely distributed HSPGs (17). The cDNA insert (⬃3500 nts) of one of the isolated clones (clone 75-5-1) hybridized specifically in a Southern blot (Fig. 1) with a probe (33 nts) designed to detect syndecans (14). However, subsequent sequencing revealed that this cDNA codes for a novel protein, unrelated to the syndecans, sharing similarity only in the region of the probe. We sequenced both strands of clone 75-5-1, and found a unique open reading frame of 1533 nts, a stop codon at 1534 and a polyadenylation signal at position 1675 followed by a poly-A tail. The remainder of the sequence was no longer in frame. Nevertheless, a second poly-A signal and tail were localized toward the 3⬘-end of the clone. Conceivably, this cDNA might have two out of frame initiation sites that encode two different proteins. Herein, we focus on the first 1675 bp. Comparison of this sequence (Fig. 2) with entries in GenBank (18) revealed selective areas of high similarity/identity with an emerging subfamily of
901
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Nucleotide and predicted amino acid sequence of OTMP. Predicted transmembrane helices are underlined. GenBank Accession No. AF292563. 902
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Northern blots. OLG mRNAs were processed as described under Methods. Lanes: 1, nonadhered OLGs; 2, adhered OLGs. (A) Hybridized with a cDNA fragment from clone 75-5-1 (60-h exposure). (B) After reprobing with a -actin cDNA (24-h exposure).
genes coding for proteins related to the “glutamatebinding protein” (GBP). These include rat GBP (19); a mouse protein (LAG) (10); the Drosophila NMDA receptor-associated protein (NMDARA1) (11) and the neural membrane protein NMP35 (12). The OTMP Expressed mRNA and Protein Northern blots confirmed OTMP mRNA expression by both non-adhered and adhered OLGs (Fig. 3A). Furthermore, an adhesion-dependent second transcript was detected in the adhered OLGs (arrow in Fig. 3A) that corresponds closely in size (⬃3300 nt) to the cDNA of clone 75-5-1 (isolated from a cDNA library generated from adhered OLGs). The appearance of the larger of the two mRNAs at a time when OLGs are poised to initiate their myelinogenic program is suggestive of a role in this event. At this time, we have no clue as to the significance of the difference in the number of mRNAs between nonadhered and adhered OLGs. Reprobing the membrane with -actin attests to the quality of the mRNAs (Fig. 3B). Initially, we thought that OTMP might be the ovine ortholog of GBP. We therefore used an anti-rat GBP polyclonal Ab (generous gift from Dr. E. K. Michaelis) to probe for the synthesis of OTMP by OLGs. A single band of ⬃65K was detected by Western blots in lysates from adhered OLGs (not illustrated). This could be taken to signify that only one mRNA was translated. Future work will address this issue. Cultured OLGs, somata and processes were immunostained by this Ab (not illustrated). Although, as argued below, OTMP is distinct from GBP, the extent of sequence similarity between the two proteins would be expected to yield the observed cross-reactivity. Nonetheless, it is pertinent to ask: how do we know that the 65K band corresponds to OTMP and not GBP? Supporting an argument that the anti-GBP Abs may be detecting only OTMP is the observation that GBP mRNA was not detected in glial cells (20, 21). Generation of OTMP-specific Abs should provide a direct answer to this quandary. The amino acid sequence of OTMP predicts a polypeptide chain with a molecular mass of 56 kD; a pI of ⬃10
and a plasma membrane localization. There is no consensus motif for N-glycosylation but there are a number of sites that could be O-glycosylated. Based on the molecular mass of the processed protein (65 kD), ⬃12% of it is contributed by carbohydrates. Alternatively, since OTMP lacks a canonical signal peptide, it is possible that there is another initiator ATG upstream to the assumed one (Fig. 2), i.e., the polypeptide chain may be larger than 56 kD. OTMP Is a Novel Transmembrane Protein That Shares Conserved Domains with an Extensive Family The availability of a large number of predicted protein sequences has permitted scientists to group them into families based on conserved motifs. It is an established fact that conservation of sequence underlies conservation of structure and function. Therefore, knowledge that an unknown protein belongs to a given family can afford significant insight into its potential function. When the predicted OTMP amino acid sequence (Fig. 2) was analyzed with the Pfam software (22) that detects conserved domains, it was found that residues 141–361 are part of a conserved motif referred to as UPF0005. This motif is present in a significant number of integral proteins— mostly uncharacterized— containing (generally) seven transmembrane segments; members of the GBP subfamily belong to this group. The relationship between OTMP and the GBP sub-family can be appreciated when their predicted amino acid sequences are compared (Fig. 4). This comparison brings to light a number of interesting points (Fig. 4). First, the five proteins share sequence identity in selected areas (highlighted in Fig. 4). These areas represent largely, but not totally (see below), transmembrane segments. Second, there is a stretch of ⬃85 amino acids, where the sequences are essentially identical; this segment corresponds to the C-terminus of the three small proteins (LAG, NMDAR1, and NMP35). Although GBP and OTMP extend for another ⬃150 amino acids, they no longer exhibit much identity or similarity. This limited and striking identity makes a compelling argument that, in each of these proteins, these ⬃85 amino acids are of structural and/or functional significance. In sharp contrast, a large fragment at the N-terminus lacks any significant similarity, thereby defining each of these proteins as a distinct entity with the exception of LAG, which, most probably, is the mouse ortholog of GBP. The data illustrated in Fig. 4 affirm OTMP as a bona fide member of the GBP subfamily. The Topology of OTMP Is Distinct from That of Known Members of the GBP Subfamily OTMP was analyzed with software that predicts transmembrane helices (TMHMM) (23) using the ExPASy Molecular Biology Server. The resulting topology (Fig. 5A)
903
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
904
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 5. Comparison of the structures of OTMP (A) with GBP (B). Gray columns indicate transmembrane helices. Helix positions for OTMP are: 157–179, 272–294, 301–323, and 335–357. GBP helices are at amino acids: 139 –161, 171–189, 202–224, 228 –246, 253–275, 284 –306, and 319 –341. Dotted lines mark residues located inside the cell, solid lines mark those facing the outside.
gives some insight into the localization, orientation and possible function of OTMP. OTMP can be modeled as a tetrahelical integral protein; there is an external loop of 91 amino acids between the first and second helix and another one of 10 amino acids between the third and fourth helix. The 5 amino acids that separate helix 2 from 3 face the inside. Both the N- and C-termini are oriented intracellularly. This topology (strongly favored by a number of other software programs) sets OTMP apart from other members of this family. This is clearly in evidence in Fig. 5B, where the topology of GBP, obtained with the same software, is shown. First, GBP is modeled as a heptahelical protein as are the other proteins presented in Fig. 4; and second, the C-terminus is predicted to be on the outside (cf. respective probabilities in Fig. 5). Aligning
Figs. 5A and 5B shows that helices 1, 2, 3, and 4 of OTMP approximately overlap (respectively) with helices 1, 5, 6 and 7 of GBP with 91–94% sequence identity (see highlighted nucleotides in Fig. 4). It might be functionally significant that three GBP helices have been replaced by an extracellular loop in OTMP. It is noteworthy that the amino acid sequence of this loop is dissimilar from its counterparts in the other proteins (positions 180 –272 in Fig. 4). It might also be functionally significant that part of this loop (positions 233–262) is H-rich (10H/30 residues). No DNA or RNA binding motif was detected. It should be pointed out that Michaelis and co-workers (9) represent GBP as a tetrahelical protein. In our hands, using different software, the heptahelical model always had the highest probability.
FIG. 4. Comparison of members of the GBP subfamily. The Pileup mode of the Wisconsin Package (10.1-UNIX) from the Genetic Computer Group was used to align these proteins. Only identical residues are highlighted. GenBank accession numbers for the listed proteins are: LAG (mouse Lag mRNA) AF182040; GBP (rat NMDA receptor glutamate-binding subunit mRNA) S61973; OTMP (ovine oligodendrocyte transmembrane protein) AF292563; NMDAR1 (Drosophila N-methyl-D-aspartate receptor-associated protein) L37377; and NMP35 (rat neural membrane protein 35) AF044201. 905
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Are These Proteins Functionally Related? Among the proteins presented here, GBP is the best characterized. Although the initial assignment of GBP as the glutamate-binding subunit of the NMDA receptor (8, 24) proved to be incorrect (25), several independent lines support the notion that GBP may be part of a heteromeric complex that might be involved in glutamate-induced neuronal excitation and toxicity (20, 21, 9). NMP35 has properties in common with GBP: they display overlapping patterns of expression and are similarly regulated during development (12). Little to nothing is known about the function of the other proteins. The fact that more members of this family are coming to light in species from Drosophila to mammals may be taken as an indication of functional relevance. As aforementioned, OTMP has a topology that distinguishes it from the other members of the GBPsubfamily and, yet, it shares with them extensive sequence identity. We can only speculate on the implication of this finding. Taken as an independent entity, OTMP exhibits structural features in common with signal transduction receptors. Its large external— H-rich—loop may constitute a ligand-binding domain. The N- and C-termini— oriented internally— have sites for S/T and Y phosphorylation; they are also P-rich. Interestingly, the N-terminus carries the amino acids that Ren et al. (26), defined as essential and conserved in all Src (SH3) homology-binding proteins. Collectively, these characteristics are suggestive of a role in signal transduction. Work in progress is designed to define the function of OTMP. ACKNOWLEDGMENTS
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
We thank Dr. Bharaty Sanyal for making the oligodendrocyte cDNA library that was used in these studies and Mr. Paul Polak for the initial screening experiments. We are grateful to Dr. E. K. Michaelis for a generous gift of anti-GBP antibodies and to Dr. Miriam Domowicz for reading the manuscript and making insightful suggestions. This research was supported by Grant RG-2677A-7/1 from the National Science Foundation.
16.
17.
18.
REFERENCES 1. Szuchet, S., Stefansson, K., Wollmann, R. L., Dawson, G., and Arnason, B. G. W. (1980) Cultured oligodendrocytes: A role for cell-substratum interaction in phenotypic expression. Brain Res. 200, 151–164. 2. Arvanitis, D., Dumas, M., and Szuchet, S. (1992) Myelin palingenesis. 1. Electron microscopical localization of myelin/ oligodendrocyte proteins in multilamellar structures by the immunogold method. Dev. Neurosci. 14, 328 –335. 3. Szuchet, S., Yim, S. H., and Monsma, S. (1983) Lipid metabolism of isolated oligodendrocytes maintained in long-term culture mimics events associated with myelinogenesis. Proc. Natl. Acad. Sci. USA 80, 7019 –7023. 4. Yim, S. H., Szuchet, S., and Polak, P. E. (1986) Cultured oligo-
19.
20.
21.
906
dendrocytes: A role for cell-substratum interaction in phenotypic expression. J. Biol. Chem. 261, 11808 –11815. Schirmer, E. C., Farooqui, J., Polak, P. E., and Szuchet, S. (1994) GRASP: A novel heparin-binding serum glycoprotein that mediates oligodendrocyte-substratum adhesion. J. Neurosci. Res. 39, 457– 473. Sanyal, B., Polak, P., and Szuchet, S. (1996) Differential expression of the heavy-chain ferritin gene in non-adhered and adhered oligodendrocytes. J. Neurosci. Res. 46, 187–197. Szuchet, S., Watanabe, K., and Yamaguchi, Y. (2000) Differentiation/regeneration of oligodendrocytes entails the assembly of a cell-associated matrix. Int. J. Dev. Neurosci. 18, 705–720. Ly, A., and Michaelis, E. (1991) Solubilization, partial purification, and reconstitution of glutamate- and N-methyl-Daspartate-activated cation channels from brain synaptic membranes. Biochemistry 30, 4307– 4316. Michaelis, E. K. (1998) Molecular biology of glutamate receptors in the central nervous system and their role in excitotoxicity, oxidative stress and aging. Prog. Neurobiol. 54, 369 – 415. Goldberger, O., Smith, T., Minta, J., and Goldberger, G. (2001) Identification and phylogeny of a novel family of eukaryotic membrane proteins related to a glutamate binding protein (GBP). GenBank AF182040. Pellicena-Palle, A., and Salz, H. (1995) The putative Drosophila NMDARA1 gene is located on the second chromosome and is ubiquitously expressed in embryogenesis. Biochim. Biophys. Acta 1261, 301–303. Schweitzer, B., Taylor, V., Welcher, A. A., McClelland, M., and Suter, U. (1998) Neural membrane protein 35 (NMP35): A novel member of a gene family which is highly expressed in the adult nervous system. Mol. Cell. Neurosci. 11, 260 –273. Szuchet, S., and Yim, S. H. (1984) Cultured oligodendrocytes: A role for cell-substratum interaction in phenotypic expression. J. Neurosci. Res. 11, 131–144. Bernfield, M., Kokenyesi, R., Kato, M., Hinkes, M. T., Spring, J., Gallo, R. L., and Lose, E. J. (1992) Biology of the syndecans: A family of transmembrane heparan sulfate proteoglycans. Annu. Rev. Cell Biol. 8, 365–393. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Chomczynsky, P., and Sacchi, N. (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol– chloroform extraction. Anal. Biochem. 162, 156 –159. Tumova, S., Woods, A., and Couchman, J. R. (2000) Heparan sulfate proteoglycans on the cell surface: Versatile coordinators of cellular functions. Int. J. Biochem. Cell Biol. 32, 269 –288. Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D. J. (1997) Gapped BLAST and PSI-BLAST: A new generation of protein database search programs. Nucleic Acids Res. 25, 3389 –3402. Kumar, K., Tilakaratne, N., Johnson, P., Allen, A., and Michaelis, E. (1991) Cloning of cDNA for the glutamate-binding subunit of an NMDA receptor complex. Nature 354, 70 –73. Sato, C., Kitajima, K., Inoue, S., Seki, T., Troy, F. A. II, and Inoue, Y. (1995) Characterization of the antigenic specificity of four different anti(␣2-8-linked polysialic acid) antibodies using lipid-conjugated oligo/polysialic acids. J. Biol. Chem. 270, 18923–18928. Pal, R., Eaton, M. J., Islam, S., Hake-Frendscho, M., Kumar, N., and Michaelis, E. K. (1999) Immunocytochemical and in situ hybridization studies of the expression and distribution of three subunits of a complex with N-methyl-D-aspartate receptor-like properties. Neuroscience 94, 1291–1311.
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
22. Bateman, A., Birney, E., Durbin, R., Eddy, S. R., Howe, K. L., and Sonnhammer, E. L. L. (2000) The Pfam protein families database. Nucleic Acids Res. 28, 263–266. 23. Sonnhammer, E. L. L., von Heijne, G., and Krogh, A. (1998) A hidden Markov model for predicting transmembrane helices in protein sequences. In Proceedings of the Sixth International Conference on Intelligent Systems for Molecular Biology (Glasgow, J., Littlejohn, T., Major, F., Lathrop, R., Sankoff, D., and Sensen, C., Eds.), pp. 175–182, AAAI Press, Menlo Park, CA.
24. Wang, H., Kumar, K., and Michaelis, E. (1992) Isolation of glutamate-binding proteins from rat and bovine brain synaptic membranes and immunochemical and immunocytochemical characterization. Neuroscience 46, 793– 806. 25. Hollmann, M., and Heinemann, S. (1994) Cloned glutamate receptors. In Annual Review of Neuroscience (Cowan, W. M., Ed.), pp. 31–108, Annual Reviews, Inc., Palo Alto, CA. 26. Ren, R., Mayer, B. J., Cicchetti, P., and Baltimore, D. (1993) Identification of a ten-amino acid proline-rich SH3 binding site. Science 259, 1157–1161.
907
Oligodendrocyte Transmembrane Protein: A Novel Member of the Glutamate-Binding Protein Subfamily Sara Szuchet, 1 David C. Plachetzki, and Kristen S. Eaton Department of Neurology and Brain Research Institute, University of Chicago, Chicago, Illinois 60637
Received April 9, 2001
Oligodendrocytes (OLGs) are cells from the central nervous system that synthesize, assemble, and maintain myelin, the multilamellar membrane that surrounds axons and facilitates the fast conduction of nerve impulses. We have shown that OLGs initiate their myelinogenic phenotype upon adhesion to GRASP, a heparin-binding glycoprotein that we purified from horse serum. In an attempt to identify the genes implicated in establishing this phenotype, we isolated a novel 3500 bp cDNA related to, but distinct from, a subfamily of glutamate-binding proteins (GBP). The cDNA encodes a protein of 511 amino acids, whose predicted sequence can be modeled as a tetrahelical integral protein with a large external loop and with the N- and C-termini located inside the cell. We have named this protein oligodendrocyte transmembrane protein (OTMP). Transcription of the message is induced upon OLG acquiring a myelinogenic phenotype (i.e., upon adhesion). The temporal expression in conjunction with the structural and biochemical features of OTMP is suggestive of a signaling receptor with a role in myelinogenesis. © 2001 Academic Press Key Words: central nervous system; glial cells; myelination; regeneration; signaling receptors; heparan sulfate proteoglycans; adhesion-induced gene transcription; Src homology-binding proteins.
Oligodendrocytes (OLGs) are cells from the central nervous system that synthesize, assemble and maintain myelin, the multilamellar membrane that surrounds axons and facilitates the fast conduction of nerve impulses. Little is known about the molecular and biochemical pathways that direct OLGs to enwrap axons with myelin. Similarly, little is known about the potential of OLGs to regenerate following the destruction of myelin, such as occurs in, e.g., multiple sclerosis. In an attempt to shed some light on these pro1 To whom correspondence should be addressed at Department of Neurology, MC 2030, 5841 South Maryland Avenue, Chicago, IL 60637. Fax: (773) 702-9076. E-mail: [email protected].
0006-291X/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
cesses, we have developed a model, albeit idealized, consisting of pure cultures of ovine OLGs isolated from young, myelinated brains (1). Using this model, we have shown that OLGs recover from the trauma of separation from myelin and reassemble multilamellar membranes—a process we have named myelin palingenesis, i.e., remyelination (regeneration) (2). Moreover, as remyelination by OLGs recapitulates the developmental steps of myelination, this in vitro model can also be utilized to study the mechanism of myelination (3, 4). In fact, we have discovered that the state of OLG differentiation (or regeneration) can be modulated in our model by facilitating or interfering with their attachment to a substratum consisting of a heparin-binding glycoprotein, named GRASP, that we purified from horse serum (5). Not only can this model be used to study the morphological and biochemical changes that occur in OLGs upon adherence to a substratum, but also to identify alterations in the expression of genes that are important for (re)myelination. Indeed, taking advantage of such a strategy and using techniques such as differential display and cDNA library screening, we have identified several genes which are upregulated upon OLG adhesion to GRASP—an event that marks the commencement of their myelinogenic phenotype (6, 7). We have characterized several of the genes implicated in establishing this phenotype (6). We have also demonstrated that adhesion induces OLGs to assemble a cell-associated matrix and have characterized the major matrix constituents as heparan sulfate proteoglycans (HSPGs) (7). Because these HSPGs are very minor components, in an attempt to establish their function, we investigated whether they belong to families of well-characterized HSPGs such as syndecans, perlecan or glypicans. However, when we screened an OLG cDNA library with an oligonucleotide probe designed to detect syndecans, we isolated a novel, 3500 bp cDNA, unrelated to syndecans but related, in lieu, to a glutamate-binding protein (GBP). GBP was first isolated bochemically as a complex of four or more proteins (8). Although the original
900
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
interpretation, and hence the name, that this was the glutamate-binding subunit of the N-methyl-Daspartate (NMDA) receptor did not hold, the finding that antisense oligonucleotides directed against the GBP cDNA block L-glutamate and NMDA-induced Ca 2⫹ influx and neurotoxicity in hippocampal neurons and in cerebellar granule cells supports a receptor role for GBP (9). A number of other cDNAs that share conserved domains with GBP have now been identified and sequenced (10 –12). Herein we demonstrate that the isolated cDNA encodes a novel protein of 511 amino acids, whose predicted sequence can be modeled as a tetrahelical integral protein with a large external loop and with the Nand C-termini located inside the cell. We have named this protein oligodendrocyte transmembrane protein (OTMP). Transcription of the message is induced upon OLG acquiring a myelinogenic phenotype. The temporal expression in conjunction with the structural and biochemical features of OTMP is suggestive of a signaling receptor with a role in myelinogenesis. METHODS Oligodendrocyte cultures. OLGs were isolated from 3- to 6-month-old lamb brains according to our standard procedure (1) and maintained in culture as described by Szuchet and Yim (13). Freshly isolated OLGs were plated on tissue culture petri dishes, where they do not adhere but form floating clusters. After 3–5 days, nonadherent OLGs were harvested, centrifuged, resuspended in medium (DMEM plus 20% horse serum, 2 mM L-glutamine and antibiotics) and seeded on polylysine-coated petri dishes; cultures were kept at 37°C. Alternatively, OLGs were adhered on plates coated with GRASP after precoating with polylysine (5) and maintained in a synthetic medium without serum. Construction and screening of an OLG cDNA library. We followed the protocols provided by the vendors (Stratagene, La Jolla, CA) to generate a unidirectional cDNA expression library from adhered OLG in the Uni-ZAP XR vector. This vector can accommodate DNA inserts from 0 to 10 kb. The titer of the packaged and unamplified library is 2.1 ⫻ 10 6 pfu/ml. The titer of the amplified library is 1.15 ⫻ 10 10 pfu/ml (6). We synthesized a set of degenerate oligonucleotide primers based on the syndecan cytoplasmic sequence YRMKKKDEGSY that is 100% conserved across species (14). Primers were end-labeled with [␥- 32P]dATP and used for library screening. We screened a total of 1.4 ⫻ 10 6 plaques. Screening of the library was performed following the procedure given by Sambrook et al. (15), modified to incorporate recommendations by the vendor NEN Research Products (Boston, MA). To obtain single clones, each positive plaque was subjected to three rounds of screening. Clones were identified by Southern blot analysis with individual probes. Southern blot analysis. Plasmids were purified; the insert cDNA was excised with restriction enzymes EcoRI and XhoI, resolved on a 1% agarose gel, transferred onto a membrane and hybridized with the same [␥- 32P]dATP end-labeled primers utilized for library screening. Northern blot analysis. Total RNA from both floating and attached OLGs were isolated following the single step procedure described by Chomczynsky and Sacchi (16). Poly(A) ⫹ RNA was obtained with the aid of an mRNA isolation kit (Qiagen, Chatswath, CA). mRNA originating from adhered and non-adhered OLGs were resolved on an agarose-formaldehyde gel, transferred to a nylon
FIG. 1. Agarose gel electrophoresis of clone 75-5-1. The insert cDNA was excised with EcoRI and XhoI, resolved on a 1% agarose gel. Lanes: 1 and 2, mol. wt. std. (digests of 174 and DNA, respectively); 3, digest of clone 75-5-1 arrowheads point to the two fragments of insert cDNA; arrow indicates pBluescript; 4, pBluescript without insert (control); 5, excised cDNA hybridized with the same probe employed for screening the library.
membrane, and hybridized with a 600 bp cDNA fragment, obtained by cutting the isolated cDNA with EcoR1 and BsrG1, according to a protocol that was based on steps taken from Sambrook et al. (15) and from the manufacturer’s recommendations for GeneScreen. Sequencing. The BigDye Terminator kit was used with an ABI 377 DNA Sequencer.
RESULTS AND DISCUSSI0N Isolation and Characterization of Clone 75-5-1 In an attempt to molecularly characterize surfaceassociated HSPGs that are induced upon OLGs adhering to GRASP (5, 7), we investigated whether HSPGs implicated in the differentiation of other cell types play any part here. Hence, we screened the cDNA library from adhered OLG with oligonucleotide probes complementary to conserved sequences of syndecan, perlecan and glypican—the most widely distributed HSPGs (17). The cDNA insert (⬃3500 nts) of one of the isolated clones (clone 75-5-1) hybridized specifically in a Southern blot (Fig. 1) with a probe (33 nts) designed to detect syndecans (14). However, subsequent sequencing revealed that this cDNA codes for a novel protein, unrelated to the syndecans, sharing similarity only in the region of the probe. We sequenced both strands of clone 75-5-1, and found a unique open reading frame of 1533 nts, a stop codon at 1534 and a polyadenylation signal at position 1675 followed by a poly-A tail. The remainder of the sequence was no longer in frame. Nevertheless, a second poly-A signal and tail were localized toward the 3⬘-end of the clone. Conceivably, this cDNA might have two out of frame initiation sites that encode two different proteins. Herein, we focus on the first 1675 bp. Comparison of this sequence (Fig. 2) with entries in GenBank (18) revealed selective areas of high similarity/identity with an emerging subfamily of
901
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Nucleotide and predicted amino acid sequence of OTMP. Predicted transmembrane helices are underlined. GenBank Accession No. AF292563. 902
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Northern blots. OLG mRNAs were processed as described under Methods. Lanes: 1, nonadhered OLGs; 2, adhered OLGs. (A) Hybridized with a cDNA fragment from clone 75-5-1 (60-h exposure). (B) After reprobing with a -actin cDNA (24-h exposure).
genes coding for proteins related to the “glutamatebinding protein” (GBP). These include rat GBP (19); a mouse protein (LAG) (10); the Drosophila NMDA receptor-associated protein (NMDARA1) (11) and the neural membrane protein NMP35 (12). The OTMP Expressed mRNA and Protein Northern blots confirmed OTMP mRNA expression by both non-adhered and adhered OLGs (Fig. 3A). Furthermore, an adhesion-dependent second transcript was detected in the adhered OLGs (arrow in Fig. 3A) that corresponds closely in size (⬃3300 nt) to the cDNA of clone 75-5-1 (isolated from a cDNA library generated from adhered OLGs). The appearance of the larger of the two mRNAs at a time when OLGs are poised to initiate their myelinogenic program is suggestive of a role in this event. At this time, we have no clue as to the significance of the difference in the number of mRNAs between nonadhered and adhered OLGs. Reprobing the membrane with -actin attests to the quality of the mRNAs (Fig. 3B). Initially, we thought that OTMP might be the ovine ortholog of GBP. We therefore used an anti-rat GBP polyclonal Ab (generous gift from Dr. E. K. Michaelis) to probe for the synthesis of OTMP by OLGs. A single band of ⬃65K was detected by Western blots in lysates from adhered OLGs (not illustrated). This could be taken to signify that only one mRNA was translated. Future work will address this issue. Cultured OLGs, somata and processes were immunostained by this Ab (not illustrated). Although, as argued below, OTMP is distinct from GBP, the extent of sequence similarity between the two proteins would be expected to yield the observed cross-reactivity. Nonetheless, it is pertinent to ask: how do we know that the 65K band corresponds to OTMP and not GBP? Supporting an argument that the anti-GBP Abs may be detecting only OTMP is the observation that GBP mRNA was not detected in glial cells (20, 21). Generation of OTMP-specific Abs should provide a direct answer to this quandary. The amino acid sequence of OTMP predicts a polypeptide chain with a molecular mass of 56 kD; a pI of ⬃10
and a plasma membrane localization. There is no consensus motif for N-glycosylation but there are a number of sites that could be O-glycosylated. Based on the molecular mass of the processed protein (65 kD), ⬃12% of it is contributed by carbohydrates. Alternatively, since OTMP lacks a canonical signal peptide, it is possible that there is another initiator ATG upstream to the assumed one (Fig. 2), i.e., the polypeptide chain may be larger than 56 kD. OTMP Is a Novel Transmembrane Protein That Shares Conserved Domains with an Extensive Family The availability of a large number of predicted protein sequences has permitted scientists to group them into families based on conserved motifs. It is an established fact that conservation of sequence underlies conservation of structure and function. Therefore, knowledge that an unknown protein belongs to a given family can afford significant insight into its potential function. When the predicted OTMP amino acid sequence (Fig. 2) was analyzed with the Pfam software (22) that detects conserved domains, it was found that residues 141–361 are part of a conserved motif referred to as UPF0005. This motif is present in a significant number of integral proteins— mostly uncharacterized— containing (generally) seven transmembrane segments; members of the GBP subfamily belong to this group. The relationship between OTMP and the GBP sub-family can be appreciated when their predicted amino acid sequences are compared (Fig. 4). This comparison brings to light a number of interesting points (Fig. 4). First, the five proteins share sequence identity in selected areas (highlighted in Fig. 4). These areas represent largely, but not totally (see below), transmembrane segments. Second, there is a stretch of ⬃85 amino acids, where the sequences are essentially identical; this segment corresponds to the C-terminus of the three small proteins (LAG, NMDAR1, and NMP35). Although GBP and OTMP extend for another ⬃150 amino acids, they no longer exhibit much identity or similarity. This limited and striking identity makes a compelling argument that, in each of these proteins, these ⬃85 amino acids are of structural and/or functional significance. In sharp contrast, a large fragment at the N-terminus lacks any significant similarity, thereby defining each of these proteins as a distinct entity with the exception of LAG, which, most probably, is the mouse ortholog of GBP. The data illustrated in Fig. 4 affirm OTMP as a bona fide member of the GBP subfamily. The Topology of OTMP Is Distinct from That of Known Members of the GBP Subfamily OTMP was analyzed with software that predicts transmembrane helices (TMHMM) (23) using the ExPASy Molecular Biology Server. The resulting topology (Fig. 5A)
903
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
904
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 5. Comparison of the structures of OTMP (A) with GBP (B). Gray columns indicate transmembrane helices. Helix positions for OTMP are: 157–179, 272–294, 301–323, and 335–357. GBP helices are at amino acids: 139 –161, 171–189, 202–224, 228 –246, 253–275, 284 –306, and 319 –341. Dotted lines mark residues located inside the cell, solid lines mark those facing the outside.
gives some insight into the localization, orientation and possible function of OTMP. OTMP can be modeled as a tetrahelical integral protein; there is an external loop of 91 amino acids between the first and second helix and another one of 10 amino acids between the third and fourth helix. The 5 amino acids that separate helix 2 from 3 face the inside. Both the N- and C-termini are oriented intracellularly. This topology (strongly favored by a number of other software programs) sets OTMP apart from other members of this family. This is clearly in evidence in Fig. 5B, where the topology of GBP, obtained with the same software, is shown. First, GBP is modeled as a heptahelical protein as are the other proteins presented in Fig. 4; and second, the C-terminus is predicted to be on the outside (cf. respective probabilities in Fig. 5). Aligning
Figs. 5A and 5B shows that helices 1, 2, 3, and 4 of OTMP approximately overlap (respectively) with helices 1, 5, 6 and 7 of GBP with 91–94% sequence identity (see highlighted nucleotides in Fig. 4). It might be functionally significant that three GBP helices have been replaced by an extracellular loop in OTMP. It is noteworthy that the amino acid sequence of this loop is dissimilar from its counterparts in the other proteins (positions 180 –272 in Fig. 4). It might also be functionally significant that part of this loop (positions 233–262) is H-rich (10H/30 residues). No DNA or RNA binding motif was detected. It should be pointed out that Michaelis and co-workers (9) represent GBP as a tetrahelical protein. In our hands, using different software, the heptahelical model always had the highest probability.
FIG. 4. Comparison of members of the GBP subfamily. The Pileup mode of the Wisconsin Package (10.1-UNIX) from the Genetic Computer Group was used to align these proteins. Only identical residues are highlighted. GenBank accession numbers for the listed proteins are: LAG (mouse Lag mRNA) AF182040; GBP (rat NMDA receptor glutamate-binding subunit mRNA) S61973; OTMP (ovine oligodendrocyte transmembrane protein) AF292563; NMDAR1 (Drosophila N-methyl-D-aspartate receptor-associated protein) L37377; and NMP35 (rat neural membrane protein 35) AF044201. 905
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Are These Proteins Functionally Related? Among the proteins presented here, GBP is the best characterized. Although the initial assignment of GBP as the glutamate-binding subunit of the NMDA receptor (8, 24) proved to be incorrect (25), several independent lines support the notion that GBP may be part of a heteromeric complex that might be involved in glutamate-induced neuronal excitation and toxicity (20, 21, 9). NMP35 has properties in common with GBP: they display overlapping patterns of expression and are similarly regulated during development (12). Little to nothing is known about the function of the other proteins. The fact that more members of this family are coming to light in species from Drosophila to mammals may be taken as an indication of functional relevance. As aforementioned, OTMP has a topology that distinguishes it from the other members of the GBPsubfamily and, yet, it shares with them extensive sequence identity. We can only speculate on the implication of this finding. Taken as an independent entity, OTMP exhibits structural features in common with signal transduction receptors. Its large external— H-rich—loop may constitute a ligand-binding domain. The N- and C-termini— oriented internally— have sites for S/T and Y phosphorylation; they are also P-rich. Interestingly, the N-terminus carries the amino acids that Ren et al. (26), defined as essential and conserved in all Src (SH3) homology-binding proteins. Collectively, these characteristics are suggestive of a role in signal transduction. Work in progress is designed to define the function of OTMP. ACKNOWLEDGMENTS
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
We thank Dr. Bharaty Sanyal for making the oligodendrocyte cDNA library that was used in these studies and Mr. Paul Polak for the initial screening experiments. We are grateful to Dr. E. K. Michaelis for a generous gift of anti-GBP antibodies and to Dr. Miriam Domowicz for reading the manuscript and making insightful suggestions. This research was supported by Grant RG-2677A-7/1 from the National Science Foundation.
16.
17.
18.
REFERENCES 1. Szuchet, S., Stefansson, K., Wollmann, R. L., Dawson, G., and Arnason, B. G. W. (1980) Cultured oligodendrocytes: A role for cell-substratum interaction in phenotypic expression. Brain Res. 200, 151–164. 2. Arvanitis, D., Dumas, M., and Szuchet, S. (1992) Myelin palingenesis. 1. Electron microscopical localization of myelin/ oligodendrocyte proteins in multilamellar structures by the immunogold method. Dev. Neurosci. 14, 328 –335. 3. Szuchet, S., Yim, S. H., and Monsma, S. (1983) Lipid metabolism of isolated oligodendrocytes maintained in long-term culture mimics events associated with myelinogenesis. Proc. Natl. Acad. Sci. USA 80, 7019 –7023. 4. Yim, S. H., Szuchet, S., and Polak, P. E. (1986) Cultured oligo-
19.
20.
21.
906
dendrocytes: A role for cell-substratum interaction in phenotypic expression. J. Biol. Chem. 261, 11808 –11815. Schirmer, E. C., Farooqui, J., Polak, P. E., and Szuchet, S. (1994) GRASP: A novel heparin-binding serum glycoprotein that mediates oligodendrocyte-substratum adhesion. J. Neurosci. Res. 39, 457– 473. Sanyal, B., Polak, P., and Szuchet, S. (1996) Differential expression of the heavy-chain ferritin gene in non-adhered and adhered oligodendrocytes. J. Neurosci. Res. 46, 187–197. Szuchet, S., Watanabe, K., and Yamaguchi, Y. (2000) Differentiation/regeneration of oligodendrocytes entails the assembly of a cell-associated matrix. Int. J. Dev. Neurosci. 18, 705–720. Ly, A., and Michaelis, E. (1991) Solubilization, partial purification, and reconstitution of glutamate- and N-methyl-Daspartate-activated cation channels from brain synaptic membranes. Biochemistry 30, 4307– 4316. Michaelis, E. K. (1998) Molecular biology of glutamate receptors in the central nervous system and their role in excitotoxicity, oxidative stress and aging. Prog. Neurobiol. 54, 369 – 415. Goldberger, O., Smith, T., Minta, J., and Goldberger, G. (2001) Identification and phylogeny of a novel family of eukaryotic membrane proteins related to a glutamate binding protein (GBP). GenBank AF182040. Pellicena-Palle, A., and Salz, H. (1995) The putative Drosophila NMDARA1 gene is located on the second chromosome and is ubiquitously expressed in embryogenesis. Biochim. Biophys. Acta 1261, 301–303. Schweitzer, B., Taylor, V., Welcher, A. A., McClelland, M., and Suter, U. (1998) Neural membrane protein 35 (NMP35): A novel member of a gene family which is highly expressed in the adult nervous system. Mol. Cell. Neurosci. 11, 260 –273. Szuchet, S., and Yim, S. H. (1984) Cultured oligodendrocytes: A role for cell-substratum interaction in phenotypic expression. J. Neurosci. Res. 11, 131–144. Bernfield, M., Kokenyesi, R., Kato, M., Hinkes, M. T., Spring, J., Gallo, R. L., and Lose, E. J. (1992) Biology of the syndecans: A family of transmembrane heparan sulfate proteoglycans. Annu. Rev. Cell Biol. 8, 365–393. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Chomczynsky, P., and Sacchi, N. (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol– chloroform extraction. Anal. Biochem. 162, 156 –159. Tumova, S., Woods, A., and Couchman, J. R. (2000) Heparan sulfate proteoglycans on the cell surface: Versatile coordinators of cellular functions. Int. J. Biochem. Cell Biol. 32, 269 –288. Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D. J. (1997) Gapped BLAST and PSI-BLAST: A new generation of protein database search programs. Nucleic Acids Res. 25, 3389 –3402. Kumar, K., Tilakaratne, N., Johnson, P., Allen, A., and Michaelis, E. (1991) Cloning of cDNA for the glutamate-binding subunit of an NMDA receptor complex. Nature 354, 70 –73. Sato, C., Kitajima, K., Inoue, S., Seki, T., Troy, F. A. II, and Inoue, Y. (1995) Characterization of the antigenic specificity of four different anti(␣2-8-linked polysialic acid) antibodies using lipid-conjugated oligo/polysialic acids. J. Biol. Chem. 270, 18923–18928. Pal, R., Eaton, M. J., Islam, S., Hake-Frendscho, M., Kumar, N., and Michaelis, E. K. (1999) Immunocytochemical and in situ hybridization studies of the expression and distribution of three subunits of a complex with N-methyl-D-aspartate receptor-like properties. Neuroscience 94, 1291–1311.
Vol. 283, No. 4, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
22. Bateman, A., Birney, E., Durbin, R., Eddy, S. R., Howe, K. L., and Sonnhammer, E. L. L. (2000) The Pfam protein families database. Nucleic Acids Res. 28, 263–266. 23. Sonnhammer, E. L. L., von Heijne, G., and Krogh, A. (1998) A hidden Markov model for predicting transmembrane helices in protein sequences. In Proceedings of the Sixth International Conference on Intelligent Systems for Molecular Biology (Glasgow, J., Littlejohn, T., Major, F., Lathrop, R., Sankoff, D., and Sensen, C., Eds.), pp. 175–182, AAAI Press, Menlo Park, CA.
24. Wang, H., Kumar, K., and Michaelis, E. (1992) Isolation of glutamate-binding proteins from rat and bovine brain synaptic membranes and immunochemical and immunocytochemical characterization. Neuroscience 46, 793– 806. 25. Hollmann, M., and Heinemann, S. (1994) Cloned glutamate receptors. In Annual Review of Neuroscience (Cowan, W. M., Ed.), pp. 31–108, Annual Reviews, Inc., Palo Alto, CA. 26. Ren, R., Mayer, B. J., Cicchetti, P., and Baltimore, D. (1993) Identification of a ten-amino acid proline-rich SH3 binding site. Science 259, 1157–1161.
907