LAR tyrosine phosphatase receptor: proximal membrane alternative splicing is coordinated with regional expression and intraneuronal localization

Molecular Brain Research 60 Ž1998. 1–12

Research report

LAR tyrosine phosphatase receptor: proximal membrane alternative splicing is coordinated with regional expression and intraneuronal localization Jari Honkaniemi a

a,b

, Julie S. Zhang a , Tao Yang a , Cheng Zhang a , Michelle A. Tisi a , Frank M. Longo a,)

Neurology SerÕice, VA Medical Center and UCSF V-127, 4150 Clement St., San Francisco, CA 94121, USA b Department of Neurology, UniÕersity of Tampere, 33521 Tampere, Finland Accepted 23 June 1998

Abstract Examination of null-mutant Drosophila and Leukocyte Common Antigen-Related ŽLAR.-deficient transgenic mice has demonstrated that the LAR protein tyrosine phosphatase ŽPTP. receptor promotes neurite outgrowth. In the absence of known ligands, the mechanisms by which LAR-type PTP receptors are regulated are unknown. We hypothesized that an alternatively spliced eleven amino acid proximal membrane segment of LAR ŽLAR alternatively spliced element-a; LASE-a. contributes to regulation of LAR function. Human, rat and mouse LAR cDNA sequences demonstrated that the predicted eleven amino acid inserts in rat and mouse are identical and share nine of eleven residues with the human insert. LASE-a splicing led to the introduction of a Ser residue into LAR at a position analogous to Ser residues undergoing regulated phosphorylation in other PTPs. In-situ studies revealed increasingly region-specific expression of LASE-a containing LAR transcripts during postnatal development. RT-PCR analysis of cortical and hippocampal tissue confirmed that the proportion of LAR transcripts containing LASE-a decreases during development. Immunostaining of cultured PC12 cells, cerebellar granule neurons, dorsal root ganglia and sciatic nerve sections with antibody directed against the LASE-a insert demonstrated signal in cell bodies but little if any along neurites. In contrast, staining with antibody directed to a separate domain of LAR showed accumulation of LAR along neurites. The findings that LASE-a splicing is conserved across human, rat and mouse, that the LASE-a insert introduces a Ser at a site likely to be targeted for regulated phosphorylation and that developmentally regulated splicing is coordinated with specific regional and intraneuronal localization point to important novel potential mechanisms regulating LAR-type tyrosine phosphatase receptor function in the nervous system. q 1998 Elsevier Science B.V. All rights reserved. Keywords: Protein tyrosine phosphatase; Leukocyte common antigen related gene; LAR; Proximal membrane; PC12 cell; Cell adhesion molecule

1. Introduction Tyrosine phosphorylation by receptor and intracellular protein tyrosine kinases ŽPTKs. constitutes a fundamental process regulating neuronal survival and development. PTK signaling is counteracted or in some cases augmented by the action of protein tyrosine phosphatases ŽPTPs.. Elucidation of mechanisms regulating PTP function will provide novel fundamental insights into neuronal survival and development. In the case of most PTK cell surface receptors, activity is induced by growth factor ligand activation or the formation of homodimeric or heterodimeric complexes ) Corresponding [email protected]

author.

Fax:

q 1-415-750-2273;

E-mail:

w26x. In contrast, the mechanisms regulating PTP receptor activity remain to be identified w2,7,25x. The observation that recombinantly expressed PTP intracytoplasmic catalytic domains demonstrate high activity in the absence of extracellular domains indicates that ligand–extracellular domain interactions or other protein–protein interactions mediated by heretofore unidentified domains might downregulate constitutive activity. The proximal membrane domain of PTP and other cell surface receptors constitutes an important candidate site regulating receptor activity. Crystal studies of the PTPa receptor identified a segment in the proximal membrane domain adjacent to the N-terminus of the first PTP catalytic domain, designated as the ‘N-terminal wedge’, that upon receptor dimerization is predicted to insert into the

0169-328Xr98r$ - see front matter q 1998 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 9 - 3 2 8 X Ž 9 8 . 0 0 1 5 1 - X

2

J. Honkaniemi et al.r Molecular Brain Research 60 (1998) 1–12

active site of the dyad-related monomer w3x. The observation that this inhibitory wedge is conserved in many receptor PTPs suggested that dimerization and associated active-site blockage might constitute a general mechanism for down regulating PTP activity. Mutagenesis studies of this wedge region in the CD45 PTP receptor have confirmed that this domain is required for ligand-induced inhibition of enzyme activity w17x. A second possible mechanism regulating PTP receptor activity via the proximal membrane domain involves phosphorylation of specific Ser residues in this region leading to decreased catalytic activity w9,24x. It has also been proposed that certain PTP functions are dependent upon interaction with other membrane-associated proteins. In this regard it is of interest to note that the proximal membrane domain of PTPl is required for association with b-catenin, a membrane associated protein undergoing tyrosine dephosphorylation and involved in cadherin mediated cell adhesion w5x. An important opportunity for assessing the possibility that the PTP receptor proximal membrane domain plays a critical role in PTP function consists of the discovery that the proximal membrane region of the Leukocyte Common Antigen-Related ŽLAR. PTP receptor undergoes alternative splicing w28x. LAR is the prototype member of the subgroup of PTP receptors that contain Ig-like and fibronectin

type III-like ŽFNIII. cell adhesion domains w4,23x. LAR modulates signaling by the insulin, epidermal growth factor and hepatocyte growth factor tyrosine kinase receptors w14x. Studies in our laboratory have shown that LAR modulates NGF and BDNF Trk-mediated neurotrophin signaling ŽTisi and Longo; Yang and Longo, unpublished data.. The discoveries that LAR is expressed by mammalian neurons and that its expression is regulated during neural development and by NGF w16,19,28,29x suggested that LAR might modulate mammalian neuronal survival andror neurite outgrowth. This hypothesis was supported by studies showing reduced size of basal forebrain cholinergic neurons and loss of cholinergic innervation of the dentate gyrus in LAR-deficient transgenic mice w27x and aberrant motor neuron pathfinding in Drosophila LAR loss-of-function mutants w13x. LAR-deficient transgenic mice have also been found to have impaired outgrowth of sensory fibers during sciatic nerve regeneration ŽXie and Longo, unpublished data.. As is the case with most PTP receptors, the extracellular ligands interacting with LAR have not been identified and the mode by which LAR is regulated is unknown. The proximal membrane region of LAR contains an alternatively spliced 33 bp exon encoding an eleven amino acid ŽSSAPSCPNISS. segment designated as LAR Alternatively Spliced Element-a ŽLASE-a; w28x; Fig. 1A.. Mea-

Fig. 1. LAR alternatively spliced element-a ŽLASE-a.. ŽA. Full-length ; 8 kb LAR transcripts consist of three Ig-like domains Žloops., eight FNIII domains Žhatched boxes., a transmembrane segment ŽTM. and two intracellular protein tyrosine phosphatase domains ŽPTP 1 and 2.. LASE-a is a 33 bp alternatively spliced exon encoding an 11 amino acid segment insert located in the proximal membrane region 41 residues downstream from the transmembrane domain and 44 residues upstream from PTP-1. LASE-c is a 27 bp alternatively spliced exon encoding a 9 amino acid segment located in the fifth FNIII domain. ŽB. LASE-a cDNA sequence in rat and mouse is identical. Two base changes in the human cDNA sequence predict Val for Ala and Thr for Ile substitutions as shown. ŽC. The proximal membrane region of mouse PTPa contains a 33 amino acid segment known as the N-terminal wedge that is conserved in LAR Žconserved residues underlined; rat amino acid sequence shown. and a large number of other PTPs. The N-terminal wedge constitutes part of the PTPa dimer interface involved in PTP dimerization. LASE-a is located adjacent to the N-terminal wedge and introduces Ser residues at the first and tenth positions of the insert Žarrows.. The Ser in the tenth position is homologous to PTPa Ser-204, which undergoes PKC-mediated phosphorylation.

J. Honkaniemi et al.r Molecular Brain Research 60 (1998) 1–12

surement of the ratio of LAR RT-PCR products with and without the LASE-a insert in cortical, cerebellar and nonneuronal tissue showed that LASE-a splicing occurs preferentially in the nervous system and that the proportion of LAR transcripts containing LASE-a decreases during development. Given the potential importance of the proximal membrane region in regulating LAR-type PTP receptor function and protein–protein interaction, the present study established whether LASE-a splicing occurs in a context in which it might function during neural development. Four fundamental questions essential for the subsequent design of direct functional studies of LASE-a were addressed: Ži. Is the LASE-a insert conserved across rat, human and mouse LAR? Žii. How is the expression of LASE-a containing isoforms spatially and developmentally regulated? Žiii. Given that LAR is expressed by multiple cell types, are LASE-a containing isoforms expressed selectively by neurons? and Živ. As shown with other LAR protein isoforms, are LASE-a-containing protein isoforms present in neurites?

2. Materials and methods 2.1. Isolation and sequencing of mouse and human LASE-a containing cDNA clones A 33-mer oligonucleotide probe corresponding to the sequence of the rat LASE-a insert was used to screen a human hippocampus cDNA library using standard protocols Ž2-year-old female, Lambda ZAPII; Stratagene.. cDNA was sequenced by the chain termination method using double stranded DNA template with successive oligonucleotide primers ŽSequenase 2.0 kit protocol and reagents from United States Biochemical.. Mouse LAR cDNA sequence flanking and including the LASE-a insert was obtained by performing RT-PCR using mouse brain poly ŽA. RNA as a template and previously described primers corresponding to the sequence of rat LAR flanking LASE-a w28x. PCR product of the expected size was subcloned into the pCR vector ŽInvitrogen; manufacturer’s reagents and protocol. and sequenced as described above. 2.2. In situ hybridization After decapitation, brain, spinal cord and dorsal root ganglion ŽDRG. tissue was harvested and frozen on dry ice. Fourteen micrometer thick coronal sections were cut with a cryostat, mounted onto Fisherbrand SuperfrostrPlus slides ŽFisher Scientific, Pittsburgh, USA. and air-dried at room temperature. Probes consisting of antisense and sense 33-mer oligonucleotides corresponding to the rat 33 bp LASE-a insert Žantisense: 5X -TTGAGATATTCGGGCAACTGGGGGCACTGGAAC-3X . were labeled at the 3X-end with 35 S-dATP ŽDuPont-NEN Research Products,

3

Boston, USA. using terminal deoxynucleotidyltransferase ŽDuPont-NEN.. The sections were hybridized at 428C for 18 h in a solution of 4 = SSC, 50% formamide, 1 = Denhardt’s solution, 1% sarcosyl, 0.02 M phosphate buffer ŽpH 7.0., 10% dextran sulphate, 500 mgrml heat denatured salmon sperm DNA, 200 mM dithiothreitol and 1 = 10 7 c.p.m.rml of the labeled probe. After hybridization, sections were washed 4 times for 15 min each in 1 = SSC at 558C and thereafter left to cool for 1–3 h at room temperature. Sections were then dipped in distilled water and subsequently in 50, 60 and 90% ethanol, air dried at room temperature and exposed to Kodak SB5 autoradiographic film ŽKodak, Rochester, CT. for 3 weeks. To visualize transcripts at the cellular level, sections were coated with autoradiographic emulsion ŽKodak., stored at 48C for 8–9 weeks and then developed with Kodak D-19 developer and Kodak GBX fixative. Sections were stained with hematoxylin and eosin before being embedded with Aquamount mountant ŽAtomergic Chemicals, Farmingdale, NY.. 2.3. RT-PCR Cortex and hippocampus tissue was dissected from Sprague–Dawley rats at the ages indicated in figure legends, frozen on dry ice powder and stored at y708C. PolyŽA. RNA was directly isolated from tissue by incubation of samples in tissue lysis buffer with oligo dT cellulose Žtype 3, Collaborative Research. using a standard protocol and reagents ŽInvitrogen FastTrack polyŽA. RNA isolation kit.. To remove any contaminating DNA, polyŽA. RNA was incubated with DNase Ž2U in a total volume of 400 ml; Promega. at 378C for 30 min. The proportion of LAR mRNA with and without LASE-a was determined by semi-quantitative RT-PCR performed as described in w28x. PCR primers flanking LASE-a where used to generate LAR product with and without the LASE-a insert. Upper Žwith LASE-a. and lower Žwithout LASE-a. bands visualized on ethidium bromide-stained gels were photographed and negatives scanned by quantitative densitometry to measure the ratio of PCR product with and without LASEa. Since amplification efficiencies of PCR reactions using the same primer pair amplifying products differing in size by less than several hundred bp remain similar in both the exponential and plateau phases of the reaction, the ratio of the products can be used to measure changes in the relative proportions of the starting transcripts w8x. 2.4. PC12 cell culture PC12 cells ŽSPC12; w20x. were grown in Dulbecco’s modified Eagle’s medium Žwith 3.7 grl NaHCO 3 , 4.5 grl glucose, 0.584 grl L-glutamine. containing 10% fetal bovine serum, 5% horse serum and penicillin–strepto-

4

J. Honkaniemi et al.r Molecular Brain Research 60 (1998) 1–12

mycin in untreated tissue culture flasks. When cells were approximately 40% confluent, differentiation was induced by changing to medium containing 2% fetal calf serum and

1% horse serum and adding b-NGF to a final concentration of 50 ngrml. NGF was kindly supplied by W. Mobley at UCSF.

Fig. 2. Regional expression of LAR transcripts containing LASE-a. Coronal brain sections from new born ŽP0, A–E., 21 day old ŽP21, F–I. and adult ŽJ–M. rats were assessed with in-situ hybridization using LASE-a antisense ŽA–C, F–G, J–K. and sense ŽD–E, H–I, L–M. oligonucleotide probes. At P0, LASE-a was evident in a subset of gray matter regions including the parietal cortex ŽPC., lateral septal nucleus ŽLS., occipital cortex ŽOC., thalamus ŽTH. and the pyramidal ŽPY. and dentate gyri ŽDG. of the hippocampus. Faint signal was associated with the inferior colliculus ŽIC. and caudate putamen ŽCP.. Signal was also present along the lateral ventricle ŽLV. and aqueduct ŽAQ.. No signal was detected in regions destined to become white matter tracts including the corpus callosum Žcc.. At P21 and in the adult, LASE-a expression was largely confined to the pyramidal ŽPY. and dentate gyri ŽDG. of the hippocampus. Signal barely above background was present in the cortex and cerebellar granule cell layers. P0 coronal ŽD, E. sections, P21 coronal ŽH. and cerebellar ŽI. sections and adult coronal ŽL. and cerebellar ŽM. sections hybridized with sense probe showed only background signal.

J. Honkaniemi et al.r Molecular Brain Research 60 (1998) 1–12

2.5. Cerebellar neuronal cell culture Seven-day-old ŽPD 7. female Sprague–Dawley rats were anesthetized by isoflurane and decapitated. The cerebellum was removed and dissociated with 20 Urml of papain and 0.5 mgrml of DNase in modified Eagle’s medium ŽMEM. without serum at 378C for 20 min. Cells were washed and further dissociated by trituration to single-cell suspension in MEM containing 10% fetal calf serum. Approximately 200 000 cells in MEM containing 10% fetal calf serum and 100 U penicillin and streptomycin were seeded in poly-Llysine coated 16 mm wells. 1-b-D-Arabinofuranosylcytosine Ž10 mM. was added to the culture after 18–22 h to prevent replication of non-neuronal cells. Cells were fixed after 7 days and immunostained with LASE-a antibody as described above. 2.6. LASE-a and LASE-c antibodies The generation and characterization of LASE-a and LASE-c antibodies was previously described w28,29x. Rabbits were immunized with synthetic peptides corresponding to the LASE-a or LASE-c inserts. Antibodies were isolated by immunoaffinity purification using peptide coupled to a Sulfolink Gel ŽPierce.. 2.7. Western blot analysis Protein extract was prepared from PC12 and cerebellar granule cells and DRG tissue. Cells and minced tissue fragments were homogenized in RIPA buffer with proteinase inhibitors, incubated on ice for 30 min, sonicated for 10 s and microcentrifuged at 14 000 rpm for 30 min. Approximately 100 mg of protein supernatant extract from each sample was loaded onto a 6% SDS polyacrylamide gel and fractionated by electrophoresis. Proteins were electrotransferred to a PVDF membrane ŽHoffman gel transfer apparatus and protocol.. Primary antibody was detected using the Western Exposure Chemiluminescent Detection System ŽClontech, Palo Alto, CA..

5

Cells were rinsed in the same buffer without antibody and incubated in biotinylated goat anti-rabbit antibody Ž1:200 dilution. for 30 min followed by ABC complex ŽVector Laboratories, Burlingame, CA. for 30 min. Diaminobenzidine containing 1% nickel sulfate was used as a chromogen to visualize immunostaining. After staining, cells were dehydrated by 70% and 95% ethanol and preserved in 50% glycerol in PBS. For immunofluorescent staining, cells were fixed in methanol at y208C for 10 min, rinsed in PBS and incubated for 2 h in the following blocking buffer: 1% BSA, 1% goat serum and 5% non-fat dried milk in PBS with 0.2% tween ŽPBST.. Cells were incubated with LASE-a Ž29 mgrml. or LASE-c Ž1.7 mgrml. affinity-purified antibody in blocking buffer overnight at 48C. Cells were washed three times in PBST, incubated with secondary fluorescein goat anti-rabbit IgG Ž15 mgrml; Vector Laboratories. for 1 h at room temp, washed three times with PBST and photographed under light and fluorescent microscopy. 2.9. Immunostaining of dorsal root ganglia and sciatic nerÕe Adult female Sprague–Dawley rats were anesthetized with 80 mgrkg of ketamine and 8 mgrkg of xylazine, and perfused transcardially first with 100 ml of saline then 3 min with ice-cold 4% paraformaldehyde in 0.1 M phosphate buffered saline ŽPBS.. After perfusion, dorsal root ganglia and sciatic nerves were excised and further incubated in the same fixative at 48C for 3 h. Tissue was cryoprotected with 20% sucrose in 0.1 M PBS at 48C

2.8. Immunostaining of cultured cells PC12 cells and cultured rat cerebellar granule neurons grown on 16 mm diameter 24-well plates ŽCorning, Corning, NY.. Cells were fixed using 2% paraformaldehyde in culture media for 30 min, followed by 4% paraformaldehyde in PBS for 30 min and permeabilized in methanol at y208C for 20 min. The wells were then washed twice in PBS for 10 min and incubated with 0.5% gelatin in PBS at room temperature for 1 h to block nonspecific binding. Fixed cells were incubated overnight with LASE-a antibody diluted in 0.1 M PBS containing 1% normal goat serum and 0.3% Triton-X 100. Affinity-purified polyclonal antibody raised against c-Fos protein Žkindly provided by Dr. Steven Sagar, UCSF. was used as a negative control.

Fig. 3. Developmental regulation of LAR transcripts containing LASE-a. ŽA. Poly ŽA. RNA was extracted from cortex and hippocampus of P0, P21 and adult rats. RT-PCR was performed using primers flanking the LASE-a insert yielding products of 208 bp and 175 bp with and without LASE-a, respectively. Five to ten microliters of each PCR product was fractionated in 6% acrylamide gels, visualized by staining with ethidium bromide and photographed. Photographic negatives were analyzed by scanning densitometry and the ratio of the signal from the upper band Žwith LASE-a. over that from the lower band Žwithout LASE-a. was calculated. The mean ratio "S.E.M. from four separate RT-PCR reactions is shown.

6

J. Honkaniemi et al.r Molecular Brain Research 60 (1998) 1–12

overnight and frozen in methylbutane at y308C. Fourteen mm thick sections were cut with a cryostat and mounted onto Superfrost slides. Sections were incubated overnight with LASE-a antibody in a buffer containing 0.1 M PB, 0.3% Triton-100 and 1% normal goat serum. LASE-a antibody incubated with LASE-a peptide was added to adjacent sections as a negative control. After primary antibody incubation, incubation with biotinylated goat anti-rabbit antibody and subsequent washing steps were performed as described in cultured cell staining. Stained sections were dehydrated in 70% and 95% ethanol and embedded in Aquamount mountant.

3. Results 3.1. The LASE-a insert is present in human and mouse LAR The originally determined human LAR sequence was derived from cDNA clones isolated from a lymphocyte library and did not contain the LASE-a insert w23x. While

this finding was consistent with subsequent studies of rat LAR transcripts in multiple organ tissues showing that LASE-a is preferentially spliced in nervous system tissue w28x, we first determined whether LASE-a is present in LAR transcripts in the human nervous system. A human brain cDNA library was screened using a probe corresponding to the rat LASE-a sequence. The first positive clone isolated contained an ; 5 kb insert. Sequencing of ; 2 kb of this insert demonstrated sequence that overlapped and was identical to that of the intracytoplasmic region of human LAR w23x. In addition, a 33 bp insert was present at the site analogous to the LASE-a site in rat LAR cDNA ŽFig. 1B.. Present in the insert were two base changes predicting an Ala to Val substitution and an Ile to Thr substitution in the human LASE-a insert. Available partial mouse LAR cDNA sequence predicts amino acid sequence identical to that of rat LAR in the N-terminal wedge region; however, mouse LAR clones containing the LASE-a insert have not been described w22x. Sequencing of the RT-PCR mouse LAR cDNA product derived in the present study demonstrated that the sequences of the 33 bp encoding the rat and mouse LASE-a inserts are identical ŽFig. 1B..

Fig. 4. Cellular expression of LASE-a containing LAR transcripts in dorsal root ganglion ŽDRG. and spinal cord. Transverse sections from P21 DRG and P21 spinal cord were hybridized with LASE-a probe, coated with autoradiographic emulsion and stained with hematoxylin and eosin. ŽA. LASE-a signal was detected in all DRG neurons Žlarge arrows. but not in the smaller, dark-staining satellite cells Žsmall arrow. Žbar s 25 mm.. ŽB. Lumbar spinal cord section. LASE-a signal was detected in a small number of neurons in the anterior horn Žbar s 500 mm..

J. Honkaniemi et al.r Molecular Brain Research 60 (1998) 1–12

In all three species, LASE-a contained a Ser at the first and tenth positions of the eleven amino acid insert Ždesignated as Ser-1 and Ser-10 of LASE-a.. It is of particular interest to note that of the eleven residues present in LASE-a, only the Ser residues are conserved in PTPa. In the PTParLAR alignment, these Ser residues align with Ser-195 and Ser-204 of PTPa ŽFig. 1C.. In PTPa , Ser-204 is phosphorylated by protein kinase C and thus may provide a target for modulation of PTPa function w24x. Regulated phosphorylation of PTPa Ser-195 has not been described. It is also of interest to note that introduction of the LASE-a segment into the LAR amino acid sequence is compatible with concomitant alignment of LAR residues with PTPa residues on both sides of the LASE-a insert. It will be of interest to determine whether the corresponding eleven residues in PTPa are also encoded by an alternatively spliced exon. 3.2. The spatial distribution of LASE-a transcript expression is deÕelopmentally regulated Previous in-situ studies of LAR distribution were conducted using LAR probes corresponding to either the 3X

7

untranslated region w16x or the catalytic domain w19x, both of which likely identified all, rather than particular, LAR isoforms. These studies revealed that in early postnatal development Žpostnatal days 0 and 4. LAR was diffusely expressed in gray matter including the cortex and midbrain, subcortical and brainstem nuclei. In the adult, LAR expression was confined to specific regions including: the basal forebrain diagonal band, reticular nucleus of the thalamus, oculomotor nucleus, hippocampus, cerebellar Purkinje neurons, deep cerebellar nuclei and superior, medial and inferior vestibular nuclei w16,19,29x. To determine whether the regional expression of LAR isoforms containing LASE-a is similar to that of LAR, LASE-a expression was assessed by in-situ hybridization using a 33 bp probe corresponding to LASE-a ŽFig. 2.. Previous Northern analysis using the same probe demonstrated specific hybridization to the expected ; 8 kb LAR full-length transcript in cortex and cerebellum tissue harvested at P0, P21 and adult stages w28x. At P0, LASE-a signal was detected in parietal and occipital cortex, ventricular lining, lateral septum, thalamus, and pyramidal and dentate cells of the hippocampus. Faint signal was detected in the caudate-putamen and inferior colliculus. Regions in which white matter tracts develop such as the corpus

Fig. 5. Western analysis of LASE-a containing LAR protein. Approximately 100 mg of protein extract derived from PC12 cells, postnatal day 7 cerebellar granule cells and DRG tissue were applied to each lane. PC12 cells were treated with NGF at 50 ngrml for 3 days. Blots were incubated with affinity-purified LASE-a antibody. ŽA. The predominant band at ; 85 kDa Žarrow. is consistent with the 85 kDa intracellular portion of LAR. ŽB. Similar levels of LASE-a containing LAR isoforms were expressed in control and NGF-exposed PC12 cells Župper and lower panels show the 85 kDa band from separate Western analyses. ŽC. LASE-a antibody recognizes the 85 kDa intracellular portion of LAR in granule cells and DRG tissue.

8

J. Honkaniemi et al.r Molecular Brain Research 60 (1998) 1–12

callosum demonstrated no signal. By P21 and adult stages, detection of LASE-a expression was primarily limited to the hippocampus with faint signal barely above background associated with the cortex and cerebellar granule cell layers. The limited distribution of LASE-a signal was in marked contrast to the more widespread distribution of LAR signal in adult brain demonstrated in previous studies. This finding suggested that expression of LASE-a occurred selectively or preferentially in a relatively narrow subset rather than in all LAR expressing neurons. The demonstration by RT-PCR studies Ždescribed below. that LASE-a transcripts are present in adult cortex and the possibility that the 33 bp LASE-a oligonucleotide probe was less sensitive than the riboprobe used to detect LAR in previous studies indicated that while LASE-a in-situ signal was more readily detected in hippocampus, expression is also likely to occur in the cortex. 3.3. RT-PCR demonstrates LASE-a containing transcripts in cortex and hippocampus Since low levels of in-situ signal associated with the hippocampus, such as that found for LASE-a in the adult,

can be non-specific, LASE-a mRNA expression in the hippocampus was further assessed using RT-PCR ŽFig. 3.. The proportion of LAR transcripts containing LASE-a in the cortex had been previously characterized; therefore, LASE-a expression in the hippocampus and cortex was compared. The proportion of LAR transcripts in the hippocampus containing LASE-a was highest at P0, ; 4.5 to 1. During postnatal development this proportion declined to a ratio of ; 2.5 to 1 in the adult, a ratio similar to that found in the cortex. Thus, RT-PCR studies confirmed that LASE-a containing LAR transcripts are present in the adult hippocampus and that LASE-a splicing in the hippocampus is developmentally regulated. 3.4. LASE-a mRNA expression in DRG and spinal cord Previous studies demonstrated that LAR mRNA expression in the nervous system is limited to neuronal cells w16,19x. As a prelude to characterization of LASE-a protein expression in the sciatic nerve, LASE-a mRNA expression in P21 and adult DRG and spinal cord was assessed ŽFig. 4.. In-situ analysis demonstrated abundant LASE-a signal

Fig. 6. Immunostaining of PC12 and cerebellar granule neurons with anti-LASE-a. ŽA. PC12 cells were cultured with NGF for 5 days and immunostained as described in Section 2. Phase microscopy was adjusted to optimize visualization of cell processes. Cell bodies were associated with abundant LASE-a staining; however, little if any staining was detected along cell processes Žarrows.. ŽB. Negative control PC12 cultures were processed identically except that the primary antibody consisted of affinity-purified polyclonal antibody raised against c-Fos. ŽC. Cerebellar granule cells Ž; 90% neurons. harvested from postnatal day-7 cerebellum and cultured for 7 days demonstrated abundant LASE-a staining in cell bodies and little or no staining along neurites. ŽD. Negative control cerebellar granule cell cultures were incubated with anti-c-Fos as the primary antibody.

J. Honkaniemi et al.r Molecular Brain Research 60 (1998) 1–12

associated with neurons but not satellite cells or other non-neurons in DRG. In the spinal cord, signal was associated with only a small number of anterior horn neurons in the lumbar segment. In contrast to LAR expression, which was readily apparent in the intermediolateral neurons of the spinal cord, no LASE-a signal was found in these neurons. Thus as found in the brain, the distribution of LASE-a signal in the spinal cord was more limited than that found for LAR expression.

3.5. LAR protein containing the LASE-a domain is present in PC12 cells, cerebellar granule cells and DRG tissue Western blot analysis of protein extracts prepared from PC12 and cerebellar cells and DRG tissue using the LASE-a antibody detected the expected 85 kDa protein constituting the LAR cytoplasmic subunit ŽFig. 5.. Similar levels of

9

LASE-a protein were detected in control and NGF treated PC12 cells. 3.6. Cellular pattern of LASE-a protein expression in cultured PC12 and cerebellar granule cells Immunostaining of cultured PC12 and cerebellar granule cells with LASE-a antibody revealed LASE-a containing protein in cell bodies with signal only slightly above background along processes ŽFig. 6.. Previous immunostaining studies of LAR in PC12 and cerebellar granule cells were conducted with antibody directed against a nine residue segment located in the extracellular region of LAR known as LAR alternatively spliced element-c ŽLASE-c; w29x.. The primarily cell body distribution of LASE-a containing isoforms described here was in marked contrast to abundant LASE-c signal detected along processes and in growth cones of these cells cultured under the same conditions.

Fig. 7. Comparison of cerebellar granule neuron immunostaining patterns using anti-LASE-a and anti-LASE-c antibodies. Cultures of cerebellar granule cells Ž; 90% neurons. were harvested from postnatal day 7 cerebellum, cultured for 7 days and incubated with LASE-a or LASE-c antibodies. Immunostaining with anti-LASE-a ŽA and B. demonstrated a relatively diffuse pattern of staining in cell bodies with a relative absence of staining along cell processes Žarrows.. In contrast, incubation with anti-LASE-c ŽC and D. revealed a punctate pattern of staining in cell bodies and abundant staining along neurites Žarrows..

10

J. Honkaniemi et al.r Molecular Brain Research 60 (1998) 1–12

In order to directly compare the intracellular distributions of LASE-a and LASE-c containing isoforms, cerebellar granule cell cultures from a common cell preparation were stained with either LASE-a or LASE-c antibodies

ŽFig. 7.. LASE-a signal was observed in a diffuse distribution in cell bodies and barely detected in cell processes. In contrast, LASE-c signal was found in a punctate distribution in cell bodies and readily detected along neurites.

Fig. 8. Immunostaining of DRG and sciatic nerve. Transverse sections from dorsal root ganglia and sciatic nerve were immunostained with affinity-purified LASE-a antibody ŽA, C, E.. For negative controls adjacent sections were stained with LASE-a antibody incubated with LASE-a peptide ŽB, D.. ŽA. Staining was associated with neurons while within the region of the nerve trunk ŽNT. no staining of processes was detected Žbar s 150 mm.. ŽC. At a higher magnification of the field shown in A, staining is detected in intracytoplasmic and membrane regions of neurons but not in surrounding satellite cells and other non-neurons Žbar s 150 mm.. ŽE. At a more distal site within the sciatic nerve trunk ŽNT. no staining was detected. ŽF. In contrast, immunostaining using antibody directed against the LASE-c insert located in the LAR extracellular domain demonstrated abundant signal within axon profiles.

J. Honkaniemi et al.r Molecular Brain Research 60 (1998) 1–12

3.7. LAR protein containing the LASE-a domain is present in cell bodies but not processes of DRG neurons in ÕiÕo To examine the intra-neuronal distribution of LASE-a protein in vivo, sections of DRG and sciatic nerve were immunostained by LASE-a antibody ŽFig. 8A–D.. Staining was evident in all DRG neurons. The wide range of the degree of staining between different neurons raised the possibility that a subpopulationŽs. of neurons expresses relatively more LASE-a; however, across size distributions no correlation between neurons size and degree of staining was evident. Consistent with in-situ studies, no protein signal was detected in satellite cells or other non-neurons. In parallel to the distribution of LASE-a staining found in cultured neurons, LASE-a protein, while readily detected in DRG cell bodies, was not detected in sciatic nerve neurites ŽFig. 8E.. This pattern was in marked contrast to the sciatic nerve pattern observed using LASE-c w29x. Previous DRG and sciatic nerve immunostaining using LASE-c antibody demonstrated LAR protein in neuronal cell bodies and along processes Žshown in Fig. 8F.. Thus, the in vitro and in vivo differential staining patterns shown here indicate that splicing of the LASE-a insert is coordinated with differential subcellular localization of LASE-a containing LAR protein isoforms compared to other LAR protein isoforms.

4. Discussion This study demonstrates that splicing of the LASE-a insert is conserved across human, rat and mouse LAR transcripts. The predicted amino acid sequence and its conservation across species raise important novel possibilities regarding potential mechanisms for the regulation of LAR function. The introduction of Ser residues at the first and tenth position of the LASE-a insert result in the presence of Ser moieties in one-to-one alignment with Ser-195 and Ser-204 of PTPa. Ser-204 of PTPa is directly phosphorylated by protein kinase C in vivo w24x. The observations that at least one other PTP receptor, PTP´, as well as PTPa and LASE-a containing LAR have a Ser at the position equivalent to Ser-204 suggest the possibility that Ser phosphorylation of PTP proximal membrane domains might constitute a common mechanism regulating PTP function. This possibility is supported by the demonstration that phosphorylation of Thr-654 in the proximal membrane domain of the EGF receptor by protein kinase C decreases EGF-dependent activation of the EGF receptor w6x. Phosphorylation of the cytosolic PTP, PTP-PEST at Ser-39, similarly located near the catalytic domain, has also been shown to decrease its enzyme activity w9x. Similar enzymatic activity studies will be required to determine whether Ser phosphorylation of the receptor PTPs modulates their activity.

11

The location of the LASE-a insert within the proximal membrane region of LAR points to additional mechanisms by which LASE-a splicing might affect LAR function. Structural modifications introduced by the inclusion of the LASE-a insert in the region adjacent to the catalytic domain could affect enzyme activity directly via cis mechanisms or indirectly via trans-acting processes such as modification of LAR proteinrprotein interactions. Enzyme kinetic studies of recombinant isoforms of PTPm demonstrated that isoforms in which the proximal membrane region was deleted had 2-fold less activity suggesting a cis-acting modulatory function for this region w10x. As shown in Fig. 1 the LASE-a insert is adjacent to the region corresponding to the N-terminal wedge defined in PTPa. The suggestion that the N-terminal wedge inhibits catalytic activity of the adjacent PTPa in PTPa dimer complexes raises the possibility that the presence of LASE-a may modify the position and function of the adjacent LAR N-terminal wedge. A third possible process through which LASE-a could influence LAR activity is by mediating heterophilic LAR proteinrprotein interactions or modulating its access to specific substrates. Co-immunoprecipitation studies indicate that LAR and other LAR-type PTPs interact with cadherin–catenin complexes w1,15x. The observation that the proximal membrane region of PTPk is required for its association with b-catenin w5x raises the important possibility that LASE-a splicing in the proximal membrane region of LAR might in turn influence association between LAR and b-catenin. The association of alternative splicing and differential subcellular localization of a receptor PTP has not been previously reported. Thus the novel observation that LASE-a containing isoforms appeared largely confined to neuron cell bodies in vitro and in vivo is of particular interest. Studies of cultured cells and neural tissue have demonstrated that other LAR isoforms ŽFig. 8, present study and Ref. w29x. and a chick analog for LAR w21x are present along neurites and in growth cones. Several cytoplasmic PTPs have been found to contain protein sequence motifs located at either the carboxy- or amino-termini that are associated with protein localization in defined subcellular compartments w18x. In the case of one cytoplasmic PTP, alternative splicing of PTP-S4 resulting in the deletion of the carboxy-terminal 34 amino acids generates an isoform that localizes to the nucleus w12x. It is also of interest to note that basolateral localization of the EGF receptor in a model of polarized epithelial cells is regulated by a topogenic sequence present in the proximal membrane region w11x. Sequence analysis of the LASE-a insert did not demonstrated significant homology with other PTPs or other proteins and thus raised the important possibility that LASE-a type alternative splicing of the proximal membrane segment in receptor PTPs might constitute an additional mechanism influencing their subcellular localization. The present findings will encourage additional studies involving LASE-a protein measurements in well-defined

12

J. Honkaniemi et al.r Molecular Brain Research 60 (1998) 1–12

subcellular fractions and LAR isoform localization using high-resolution imaging techniques. In-situ and RT-PCR analysis demonstrated that both the proportion of LAR transcripts and the absolute level of LAR transcripts containing LASE-a decrease during postnatal development. Moreover, the distribution of neurons expressing LAR transcripts containing LASE-a, particularly at the P21 and adult stages, appears to comprise a subset of the larger population of neurons expressing LAR in g en eral. T h e co n co m itan t p ro cesses o f developmentally-regulated and region-specific expression of LASE-a containing LAR transcripts revealed in this study supports the hypothesis that LASE-a splicing contributes to the role played by LAR during neuronal development. This spatial–temporal coordination along with the introduction of Ser residues at positions found to be strategic in the regulation of other PTPs and the specific location of the LASE-a insert in the proximal membrane domain suggest novel mechanisms by which LASE-a splicing might regulate LAR localization and activity.

Acknowledgements Supported by NIA R01 AG09873 ŽF.L.., the Veterans Administration ŽF.L.., and the Finnish Neurology Foundation and the Finnish Academy of Sciences ŽJ.H... We thank Dr. Frank Sharp in the UCSFrVAMC Department of Neurology for reviewing histological findings.

References w1x B. Aicher, M.M. Lerch, T. Muller, J. Schilling, A. Ullrich, Cellular redistribution of protein tyrosine phosphatases LAR and PTPs by inducible proteolytic processing, J. Cell Biol. 138 Ž1997. 681–696. w2x D. Barford, Z. Jia, N.K. Tonks, Protein tyrosine phosphatases take off, Nature Structure Biology 2 Ž1995. 1043–1053. w3x A.M. Bilwes, J. den Hertog, T. Hunter, J.P. Noel, Structural basis for inhibition of receptor protein-tyrosine phosphatase-a by dimerization, Nature 382 Ž1996. 555–559. w4x S.M. Brady-Kalnay, N.K. Tonks, Protein tyrosine phosphatases as adhesion receptors, Current Biology 7 Ž1995. 650–657. w5x J. Cheng, K. Wu, M. Armanini, N. O’Rourke, D. Dowbenko, L.A. Lasky, A novel protein-tyrosine phosphatase related to the homotypically adhering k and m receptors, J. Biol. Chem. 272 Ž1997. 7264–7277. w6x C. Cochet, G.N. Gill, J. Meisenhelder, J.A. Cooper, T. Hunter, C-kinase phosphorylates the epidermal growth factor receptor and reduces its epidermal growth factor-stimulated tyrosine protein kinase activity, J. Biol. Chem. 259 Ž1984. 2553–2558. w7x S.J. Fashena, K. Zinn, The ins and outs of receptor tyrosine phosphatases, Current Biology 5 Ž1995. 1367–1369. w8x K.P. Foley, M.W. Leonard, J.D. Engel, Quantitation of RNA using the polymerase chain reaction, Trends in Genetics 9 Ž1993. 380–385. w9x A.J. Garton, N.K. Tonks, PTP-PEST: a protein tyrosine phosphatase regulated by serine phosphorylation, EMBO J. 13 Ž1994. 3763–3771. w10x M.F.B.G. Gebbink, M.H.G. Verheijen, G.C.M. Zondag, I. van Etten, W.H. Moolenaar, Purification and characterization of the cytoplasmic domain of human receptor-like protein tyrosine phosphatase RPTPm, Biochemistry 32 Ž1993. 13516–13522.

w11x M. Hobert, C. Carlin, Cytoplasmic juxtamembrane domain of the human EGF receptor is required for basolateral localization in MDCK cells, J. Cellular Physiology 162 Ž1995. 434–446. w12x S. Kamatkar, V. Radha, S. Nambirajan, R.S. Reddy, G. Swarup, Two splice variants of a tyrosine phosphatase differ in substrate specificity, DNA binding and subcellular localization, J. Biol. Chem. 271 Ž1996. 26755–26761. w13x N.X. Krueger, D. Van Vactor, H.I. Wan, W.M. Gelbart, C.S. Goodman, H. Saito, The transmembrane tyrosine phosphatase DLAR controls motor axon guidance in Drosophila, Cell 84 Ž1996. 611– 622. w14x D.T. Kulas, B.J. Goldstein, R.A. Mooney, The transmembrane protein-tyrosine phosphatase LAR modulates signaling by multiple receptor tyrosine kinases, J. Biol. Chem. 271 Ž1996. 748–754. w15x R.M. Kypta, H. Su, L.F. Reichardt, Association between a transmembrane protein tyrosine phosphatase and the cadherin–catenin complex, J. Cell Biol. 134 Ž1996. 1519–1529. w16x F.M. Longo, J.A. Martignetti, J.M. Le Beau, J.S. Zhang, J.P. Barnes, J. Brosius, Leukocyte common antigen-related receptor-linked tyrosine phosphatase. Regulation of mRNA expression, J. Biol. Chem. 268 Ž1993. 26503–26511. w17x R. Majeti, A.M. Bilwes, J.P. Noel, T. Hunter, A. Weiss, Dimerization-induced inhibition of receptor protein tyrosine phosphatase function through an inhibitiry wedge, Science 279 Ž1998. 88–91. w18x L.J. Mauro, J.E. Dixon, ‘Zip codes’ direct intracellular protein tyrosine phosphatases to the correct cellular ‘address’, Trends in Biochemical Science 19 Ž1994. 151–155. w19x M. Sahin, J.J. Dowling, S. Hockfield, Seven protein tyrosine phosphatases are differentially expressed in the developing rat brain, J. Comp. Neurol. 351 Ž1995. 617–631. w20x D. Schubert, S. Heinemann, Y. Kidikoro, Cholinergic metabolism and synapse formation by a rat nerve cell line, Proc. Natl. Acad. Sci. U.S.A. 74 Ž1977. 2579–2583. w21x A.W. Stoker, B. Gehrig, F. Haj, B.-H. Bay, Axonal localization of the CAM-like tyrosine phosphatase CRYPa: a signaling molecule of embryonic growth cones, Development 121 Ž1995. 1833–1844. w22x R.Q.J. Schaapveld, A.M.J.M. van den Maagdenberg, J.T.G. Schepens, D.O. Weghuis, A.D. van Kessel, B. Wieringa, W.J.A.J. Hendriks, The mouse gene ptprf encoding the leukocyte common antigen-related molecule LAR: cloning, characterization, and chromosomal localization, Genomics 27 Ž1995. 124–130. w23x M. Streuli, N.X. Krueger, L.R. Hall, S.F. Schlossman, H. Saito, A new member of the immunoglobulin superfamily that has a cytoplasmic region homologous to the leukocyte common antigen, J. Exp. Med. 168 Ž1988. 1523–1530. w24x S. Tracy, P. van der Geer, T. Hunter, The receptor-like protein-tyrosine phosphatase, RPTPa , is phosphorylated by protein kinase C on two serines close to the inner face of the plasma membrane, J. Biol. Chem. 270 Ž1995. 10587–10594. w25x D. Van Vactor, Protein tyrosine phosphatases in the developing nervous system, Current Opinion in Cell Biology 10 Ž1998. 174–181. w26x F.U. Weiss, H. Daub, A. Ullrich, Novel mechanisms of RTK signal generation, Current Opinion in Genetics and Development 7 Ž1997. 80–86. w27x T.T. Yeo, T. Yang, S.M. Massa, J.S. Zhang, J. Honkaniemi, L.L. Butcher, F.M. Longo, Deficient LAR expression decreases basal forebrain cholinergic neuronal size and hippocampal cholinergic innervation, J. Neuroscience Res. 47 Ž1997. 348–360. w28x J.S. Zhang, F.M. Longo, LAR tyrosine phosphatase receptor: alternative splicing is preferential to the nervous system, coordinated with cell growth and generates novel isoforms containing extensive CAG repeats, J. Cell Biol. 128 Ž1995. 415–431. w29x J.S. Zhang, J. Honkaniemi, T. Yang, T.T. Yeo, F.M. Longo, LAR tyrosine phosphatase receptor: a developmental isoform is present in neurites and growth cones and its expression is regional- and cell-specific, Molecular and Cellular Neuroscience 10 Ž1998. 271– 286.