Brain‐specific expression of human microtubule‐associated protein 1A (MAP1A) gene and its assignment to human chromosome 15

We isolated several cDNA fragments by immunoscreening a human cDNA library with our monoclonal antibody, BG5, that showed neuronal staining on human and rat brain sections. A 1,570 bp sequence of one cDNA fragment showed 75% homology to the rat microtubule‐associated protein 1A (MAP1A) cDNA sequence. This rat MAP1A‐like human cDNA was highly specific to the adult brain among human tissues tested, and was expressed in various brain regions including white matter. The size of the mRNA detected with Northern blot analysis in adult human brain equaled 10 kb. The gene of this cDNA was assigned to human chromosome 15 that has a syntenic region of mouse chromosome 2, where the mouse MAP1A gene has been assigned. These results indicate that this rat MAP1A‐like cDNA is a portion of human MAP1A and is a conserved molecular species among humans and rodents. © 1995 Wiley‐Liss, Inc. This article is a US Government work and, as such, is in the public domain in the United States of America.


INTRODUCTION
Studies on the molecular mechanisms of Alzheimer disease (AD) have shown that AD occurs by multifactorial causes that may include a mutation of amyloid precursor protein (APP) (Goate et al., 1991), allele e4 of ApoE protein (Corder et al., 1993), abnormally phosphorylated tau protein (Goedert et al., 1991), and decrease of the levels of functional molecules such as cytochrome oxidase (Chandrasekaran et al., 1994). Yet it is still necessary to isolate several other molecules and antigens to fully understand the molecular mechanism underlying the pathogenesis of AD because at least three responsible genes for familial AD cases have not been isolated (e.g., gene on chromosome 19 for the late onset familial AD; Farrer and Stice, 1993; gene on chromosome 14 for the early onset familial AD; Schellenberg et al., 1992).
In an attempt to identify unique epitopes for AD (Fukuyama et al., 1994a;Fukuyama and Rapoport, 1993), we have isolated various monoclonal antibodies raised against selected brain regions that are susceptible to AD (Rapoport, 1990). One antibody, which showed immunoreactivity to neurons of rat and human brain, was used to isolate corresponding cDNA from a human hgt 1 1 cDNA library. Here we report the partial sequence of clones thus isolated, the brain-specific expression of the mRNA, and the chromosomal assignment of the gene.

MATERIALS AND METHODS Isolation of Immunoreactive Clones
A hgtll cDNA library (human amygdala, Clontech, Palo Alto, CA) was screened by a monoclonal antibody BGS that was obtained in a previous experiment (Fukuyama et a]., 1994a) as described by Young and Davis (1983) with minor modification. After tertiary screening of immunoreactive plaques, we obtained two hgtll clones, BGSi-2 and BGSi-4. Inserted cDNA sizes were determined by polymerase chain reaction (PCR) using hgtl 1-forward and reverse primers (Clontech). Clones BGSi-2 and BGSi-4 showed a similar sequence in the initial sequencing experiment, and thus were assumed to encode the same protein. The shorter BGSi-2 cDNA was subsequently used for Northern blot analysis and gene assignment, whereas the larger BG5i-4 cDNA was sequenced.

Sequencing of cDNA Fragments
The PCR-amplified BGSi-4 cDNA fragment was sequenced using an automated sequencer (DNA Technology, Bethesda, MD). Sequence primers were designed to cover entire portions of the BG5i-4 clone. The BGSi-2 cDNA fragment also was partially sequenced by a cycle sequencing procedure combined with chemiluminescent detection (Circum Vent Phototope Kit, New England Biolabs, Beverly, MA). The sequence data were aligned by an assembly program on a Macintosh computer (AssemblyLIGN, IBI, New Haven, CT). The homologous sequence to that of our cDNA was searched in GenBank (release 83.0) and EMBL (release 38.0) data bases, using appropriate software (Convex, Genetic Computer Group, Madison, WI) on a Macintosh computer.

Northern Blot Analysis
Northern blot filters charged by mRNAs isolated from several different tissues and several brain parts of human were purchased from Clontech. After hybridizing BG5i-2 cDNA under highly stringent conditions using premade Hybrisol I (Oncor, Gaithersburg, MD), filters were washed at 52°C and exposed to X-ray films (Eastman Kodak Company, Rochester, NY) as described (Fukuyama et al., 1994b). Filters were subsequently hybridized with human p-actin cDNA probe (Clontech).

Gene Assignment
A genomic Southern blot filter of the human-rodent monochromosomal hybrid panel was purchased from Oncor. The membrane was charged by 24 kinds of DNAs as well as by host DNAs of human, mouse, and hamster after they were digested with PstI restriction enzyme (Oncor). This panel was hybridized and washed under the same stringent conditions as used with the Northern blot analysis (see above).

RESULTS
Inserted cDNA sizes determined by the PCR technique were 1.5 kb for BGSi-4 clone and 0.8 kb for the BG5i-2 clone, respectively (data not shown). Sequence data of the 1.5 kb BG5i-4 clone and its sequence-data base search showed marked similarity only to the sequence of rat microtubule-associated protein 1A (MAPlA) (GenBank-accession number M83196) ( was very similar (82% homology) (Fig. IB). The sequence of the BGSi-2 clone corresponded to a 295-1,017 portion of BGSi-4 cDNA (data not shown). Northern blot analysis of mRNAs from various human tissues, using the BG5i-2 clone, showed that this clone was expressed almost exclusively in brain (Fig. 2,  BG5i-2, left column). The mRNA size of the hybridized signal was about 10 kb. A minor level of expression of this 10 kb-message was detected in heart and skeletal muscle. The 10 kb-message was expressed in each brain region that was examined (Fig. 2, BG5i-2, right column). Actin cDNA was used to show the level of electrophoresed RNA (Fig. 2, p-actin).
Gene assignment utilizing a human-rodent monochromosomal hybrid-panel whose DNAs were digested first with PstI showed that human-specific 3.6 kb hybridization signal only on hybrid DNA containing human chromosome 15 (15*, Fig. 3).

DISCUSSION
Sequence comparison to the data base showed strong homology of our cDNAs only to rat microtubuleassociated protein 1A (MAPIA) cDNA. The mRNA size detected by our cDNA was about 10 kb and was the same size as rat MAPlA mRNA (Langkopf et al., 1992). MAPlA has been known to be the major MAP in brains of adult rats and hamsters (Garner et al., 1990;Langkopf et al., 1992), which is also specific to the brain compared with other tissues of these species (Langkopf et al., 1992;Lewis et al., 1986). Northern blot analysis using BG5 clones agreed with this evidence regarding brain specificity and mRNA size. Because of these results, and of our chromosomal assignment data (see below), we concluded that our cDNA is most likely a portion of human MAP 1 A.
The sequences of our two cDNA fragments corresponded to the sequence of the central portion of rat MAPlA cDNA. It is known that an amino terminal portion of MAPlA and MAPlB is the binding domain to microtubules, and that a carboxyl terminal portion in-cludes the light chains, LCl and LC2 (Langkopf et al.. 1992). Sequences of MAPlA and MAPlB genes in these portions are well conserved, whereas the central portion between them is not (Garner et a]., 1990;Langkopf el. al., 1992). This evidence suggests that the central portion of MAPlA also is antigenically diverged and therefore could encode a species-specific epitope that was recognized with our monoclonal antibody in the screening process of cDNAs. Indeed, our cDNAs include the 5'-portion which is divergent from the corresponding sequence of rat MAPlA (56% homology for 400 bp at 5'-portion and 82% for the last, 1,170-bp segment). Thus, it is of interest to map BGS-epitope within this central portion of MAPl A. However, entire sequence comparison of human MAPlA gene to that of rodent MAPl A is necessary to confirm this hypothesis.
In the brain samples tested, levels of mRNA of human MAPlA were about the same. A substantial hybridization signal of human MAPl A also was detected in white matter sample (corpus callosum). Thus, it is likely that MAPlA is expressed in glial cells as well as in ASSIGNMENT OF HUMAN MAPlA GENE TO CHROMOSOME 15 * M F m h 1 2 3 4 5 6 7 8 9 10 11 1213 1 4 1 5 1 6 1 7 1 8 1 9 2 0 2 1 2 2 X Y 3.6 kb ___ -* Fig. 3. Assignment of the gene of BG5i-2 cDNA clone to human chromosome 15. Control DNAs of human (M, male; F, female), mouse (m), and hamster (h), as well as DNAs from hybrid cells which have single human chromosomes (1 through Y), are electrophoresed after they are digested with PstI. Human chromosomes 1, 2, 16, 17, 20 and 21 were maintained in human-mouse hybrid cells, whereas the other human chromo-somes were held in human-hamster (Oncor). The filter of the hybrid panel was hybridized with a shorter sized clone, BG5i-2. A 3.6 kb human specific PsrI band was found only in the hybrid line containing human chromosome 15 (*). Arrows indicate weak rodent signals that cross-hybridized with the cDNA clone.
neurons. Indeed, its mRNA has been detected in oligodendroglia in the rat brain (Tucker et al., 1989).
MAPlB is the major isoform of MAP1 in the embryonic brain. Its expression decreases dramatically during development, while expression of MAP1 A increases (Garner et al., , 1990Oblinger et al., 1994;Safaei and Fischer, 1989;Schoenfeld et al., 1989). Biochemical properties of MAPlA are homologous to those of MAPIB. However, genes for MAPlA and MAPlB are distinct in rodents (Garner et al., 1990;Langkopf et al., 1992). Recently, human MAPlB gene was assigned to human chromosome 5 (Lien et al., 1991). Therefore, assignment of our gene to human chromosome was of interest. We could have assigned the gene of our BG5i-2 cDNA clone to human chromosome 15 by our hybridization experiment using a human-rodent hybrid panel. Therefore, we suggest that MAPlA and MAPlB genes likely are distinct in humans as well as in rodents.
Interestingly, human chromosome 15 is syntenic to a region of mouse chromosome 2 (O'Brien and Marshall-Graves, 1991), where mouse MAPlA gene is located. Human MAPlB gene was assigned to human chromosome 5q13 (Lien et al., 1991), which again includes the mouse syntenic region of chromosome 13 (O'Brien and Marshall-Graves, 1991), on which mouse MAP1 B gene was assigned (Garner et al., 1990). These chromosomal assignments, together with sequence conservation of MAPlA and MAPlB among humans and rodents, suggest that MAPlA and MAPlB are well conserved molecular species that evolved from a common ancestral gene.
The complete human MAPlB cDNA and its gene have been isolated by Lien et al. (1994). These authors also reported a human MAPlB-related gene by searching sequence homology in the expressed sequence tagged library (EST database) with their MAPIB sequence, and found that EST 01683, 01974, and 00368 genes are homologous to some extent to MAPlB gene (Lien et d . , 1994). A further homology search showed that these clones have strong similarity to the rat MAPIA sequence (Lien et al., 1994). The human MAPlB-related EST clone was assigned to human chromosome 1.5~13-qter. Although the authors mentioned its specific expression in the adult brain, they did not show any Northern blot data. Nevertheless, their data obtained without cloning experiments suggested that MAP1 B-related gene is a human counterpart of rat MAPl A, namely human MAP1 A. Because chromosomal assignment, brain specific expression, and the size of mRNA detected by our clone are very similar to their information, we believe that we have confirmed their MAPlB-related EST clones as portions of MAP 1 A gene.
MAPS have been implicated in the pathogenesis of AD (Goedert et al., 1991). MAPlB (alternately called MAPS) was shown to be involved in AD pathology, and its expression is altered in AD brain (e.g., Takahashi et al., 1991). However, up to the present, MAPlA expression in AD brain has not been reported. We currently are analyzing the levels of protein and mRNA of MAPlA in control and AD brains, using Western blot and Northern blot techniques.