Rat B2 Sequences Are Induced in the Hippocampal CA1 Region After Transient Global Cerebral Ischemia*

Global brain ischemia causes cell death in the CA1 region of the hippocampus 3–5 days after reperfusion. The biological pathway leading to such delayed neuronal damage has not been established. By using differential display analysis, we examined expression levels of poly(A) RNAs isolated from hippocampal extracts prepared from rats exposed to global ischemia and found an up-regulated transcript, clone 17a. Northern blot analysis of clone 17a showed an approximately 35-fold increase in the ischemic brain at 24 h after four-vessel occlusion. Rapid amplification of cDNA ends of clone 17a revealed a family of genes (160–540 base pairs) that had the characteristics of rodent B2 sequences. In situ hybridization demonstrated that the elevated expression of this gene was localized predominantly in the CA1 pyramidal neurons. The level of expression in the CA1 region decreased dramatically between 24 and 72 h after ischemia. The elevated expression of clone 17a was not observed in four-vessel occlusion rats treated with the compound LY231617, an antioxidant known to exert neuroprotection in rats subjected to global ischemia. Since delayed neuronal death has the characteristics of apoptosis, we speculate that clone 17a may be involved in apoptosis. We examined the expression level of clone 17a inin vitro models of apoptosis using cerebellar granule neurons that were subjected to potassium removal, glutamate toxicity, or 6-hydroxydopamine treatment and found that clone 17a transcripts were induced in cerebellar granule neurons by glutamate or 6-hydroxydopamine stimulation but not potassium withdrawal.

A short period of global cerebral ischemia in rodents causes neurons in the striatum, hippocampus, and lateral thalamus to die (1). Intriguingly, pyramidal neurons in the CA1 region of the hippocampus undergo delayed neuronal death 3-5 days after the insult (1,2). A similar phenomenon occurs in human cerebral ischemia (3). This time lag provides a window of opportunity for therapeutic interventions after ischemic injury. However, the molecular mechanisms that trigger and lead to delayed neuronal death have not been well established, al-though many hypotheses have been proposed such as excitotoxicity of glutamate, disturbed calcium homeostasis, altered lipid metabolism, free radicals, and mitochondrial involvement (for reviews, see Refs. 4 and 5). There is a growing body of evidence suggesting that apoptotic events occur in both global and focal brain ischemia (6 -8). It is possible that many of these hypotheses may in fact represent different aspects of a common mechanism. One of the approaches that can be used to further our understanding of the molecular mechanism of delayed neuronal death is to establish the gene expression profile of the process. Differential expression of many genes has been observed in ischemic brains, including some immediate early genes, heat shock proteins, and factors controlling apoptosis such as Bcl2, Bcl-x, Bax, and caspases (for reviews, see Refs. 9 and 10).
Differential display (DD) 1 is a technique based on reverse transcription (RT) and PCR (11) and is widely used for identifying genes with altered expression in pathological and special physiological conditions. It has been used successfully in the discovery of several differentially expressed sequences in rodent models of cerebral ischemia. Tsuda et al. (12,13) discovered two genes, serine protease inhibitor, SPI-3, and zinc transporter gene, ZnT-1, up-regulated in gerbil hippocampi 24 h after 5 min of carotid occlusion. Wang and co-workers (14,15) detected adrenomedullin induction using RNAs isolated from rat ipsilateral cortex 2h and 12 h after permanent middle cerebral artery occlusion. In our study, we used a rat model of global cerebral ischemia induced by 30 min of four-vessel occlusion (4VO). Since we are interested in gene induction prior to severe neuronal damage, and it is known that histological evidence of neuronal damage in the CA1 region of this model is not observed at 24 h after reperfusion (16), we used total RNA isolated at 24 h post-ischemia for differential display analysis.

EXPERIMENTAL PROCEDURES
Animals--Transient forebrain ischemia was induced by the 4VO treatment as described by Pulsinelli and Brierley (17). Briefly, Wistar rats (Hilltop Labs, Scottsdale, PA) were prepared for forebrain ischemia by electrocauterizing the bilateral vertebral arteries and placing atraumatic clasps around the common carotid arteries without interrupting the arterial blood flow. Rats were anesthetized for the whole procedure using 2% halothane inhalation. On the following day, forebrain ischemia was induced by tightening the clasps for 30 min. In the case of sham-treated animals, the carotids were exposed but not occluded. Body temperature was maintained at 37°C for 1 h during and after the 4VO treatment by means of heat lamps. Animals were sacrificed by decapitation at 24, 48, and 72 h after ischemia or sham operation. All * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Cultured Rat Cerebellar Granule Neurons (CGN)--Primary cultures of CGNs were prepared from 8-day-old-Harlan Sprague-Dawley rat pups (Harlan Breeders, Indianapolis, IN) as described by Gallo et al. (18). Cells were dissociated from freshly dissected cerebella by mechanical disruption in the presence of trypsin and DNase (Sigma) and were resuspended in basal medium, Eagle's (Life Technologies, Inc.), supplemented with 10% fetal bovine serum, 25 mM KCl, and 0.1 mg/ml gentamicin (Life Technologies, Inc.). Cells were seeded at a density of 1.2-1.5 ϫ 10 6 cells/ml on poly-L-lysine-coated dishes. Cytosine arabinoside (10 M; Sigma) was added to the culture medium 24 h after initial plating to arrest the growth of non-neuronal cells. Cultures were maintained at 37°C in a humidified incubator with 95% air, 5% CO 2 and were fed with glucose (10 mM) after 7 days in vitro and thereafter every 4th day. Cultures generated by this method have been shown to contain 95% granule neurons (19). In all experiments, neurons were used after being cultured 7-8 days in vitro. For low K ϩ -induced neuronal apoptosis, the concentration of KCl in the conditioned medium was switched from 25 to 5 mM for an overnight treatment. For glutamate-induced apoptosis, CGNs were treated with glutamate (Sigma) at a final concentration of 30 M in a 12-well plate (3.8 cm 2 /well) and incubated overnight. In a separate experiment, CGNs were treated with 50 M 6-hydroxydopamine (6-OHDA) (Research Biochemicals, Natick, MA) overnight.
Oligonucleotides and Peptide Synthesis--Oligonucleotides were custom-synthesized by Genosys Corp. (The Woodlands, Texas). Peptide synthesis and rabbit antibody production was provided by Genemed Synthesis, Inc. (South San Francisco, CA).
Total RNA Isolation--Total RNA was isolated from eight 4VO-and eight sham-treated animals. The region of the dorsal hippocampus containing the CA1 layer was carefully dissected from the remainder of the hippocampus, immediately frozen in liquid nitrogen, and stored at Ϫ80°C. Total RNA was isolated using the RNAgents® total RNA isolation system (Promega Corp., Madison, WI) following the manufacturer's instructions.
Differential Display--To remove DNA contamination, 5 g of total RNA from each animal was incubated at 37°C for 30 min with 1 unit of amplification grade RNase-free DNase I (Life Technologies, Inc.) in 10 mM Tris-HCl, pH 7.5, and 10 mM MgCl 2 . The DNA digestion was terminated by addition of 1 l of 0.5 M EDTA and 3 l of 2 M sodium acetate per 50-l reaction. RNA was then phenol-extracted and precipitated (20).
For first strand cDNA synthesis, total RNA (2 g) pooled from four animals was used for each reaction. The reaction contained 20 pmol of T12G, T12C, or T12A as primers. Other reagents were obtained from the SuperScript II RT kit (Life Technologies, Inc.).
The PCR products were incubated with 15 l of loading buffer at 95°C for 2 min. Samples, 3 l each, were electrophoresed at 1700 V on a 6% polyacrylamide gel containing 8.3 M urea until the xylene cyanol marker reached the bottom. After immobilizing the gel on Whatman 3MM paper and drying for 30 min under vacuum at 80°C, the gel was exposed to a Biomax MR film (Eastman Kodak Co.) overnight. The intensity of the bands on the x-ray film is indicative of expression levels of the transcripts. Differentially expressed bands were observed between 4VO-and sham-treated animals. The DD-PCR was repeated with the primer pairs that detected the differential expression. To achieve higher resolution, PCR products containing bands with the confirmed difference were electrophoresed on a 60-cm-long polyacrylamide gel using a genomyxLR high-throughput sequencer (Genomyx Corp., Foster City, CA). The bands of interest were excised and immersed in 100 l of 10 mM Tris buffer containing 1 mM EDTA, pH 8.0. The DNA was eluted by cycles of freezing, thawing, and vortexing. Five l of eluted DNA was amplified with 60 pmol of each primer used for DD-PCR, PCR buffer, 6 nmol of each dNTP, and 2.5 units of AmpliTaq Gold® DNA polymerase (Perkin-Elmer) per 60-l reaction. Thermocycles were per-formed in a Peltier Thermal Cycler, PTC-225 (MJ Research, Inc., as follows: 92°C, 15 min; 40 times (92°C, 15 s; 40°C, 2 min; ramp to 72°C in 1 min; 72°C, 1 min); 72°C, 10 min and kept at 4°C thereafter until further processed. The PCR products were purified with the QIAquick PCR purification kit (Qiagen Inc., Chatsworth, CA) and ligated with pCR 2.1 vector using the original TA cloning kit (Invitrogen Corp., Carlsbad, CA). The plasmids were transformed into MAX Efficiency DH5␣ or DH10B competent cells (Life Technologies, Inc.) according to the manufacturer's instructions. Cells were incubated at 37°C for 18 h in Magnificent Broth (MacConnell Research Corp., San Diego, CA) containing 100 g/ml ampicillin. Plasmid DNAs were purified using the QIAwell ultra plasmid kit (Qiagen). Clones that contained inserts were submitted for automated DNA sequencing (Lilly DNA Technology group, Indianapolis, IN). The nucleic acid and deduced peptide sequences were searched in GenBank TM and GenEMBLTM data bases using the FastA and BLAST algorithms provided in the Genetics Computer Group (GCG) software package (21).
Northern Blot Analysis--Northern blotting was carried out as described originally by Alwine et al. (22) with modifications. Ten g of each RNA sample and 10 g of 0.24 -9.5-kb RNA ladder (Life Technologies, Inc.) was denatured and subjected to electrophoresis on 1.2% agarose gels containing 2.2 M formaldehyde. The gels were routinely stained with ethidium bromide to visualize the RNA ladder and to verify that each lane contained similar amounts of undegraded rRNAs. RNA was transferred onto Zeta-Probe GT nylon membranes (Bio-Rad) at 160 V in 50 mM Tris buffer containing 45 mM boric acid and 0.5 mM EDTA for 1 h at 4°C and cross-linked by baking at 80°C in vacuum for 30 min.
The insert of clone 17a was released from pCR 2.1 with EcoRI, separated on agarose gel, excised, and extracted with the Prep-A-Gene kit (Bio-Rad). Labeling was performed with 50 Ci of [␣-32 P]dCTP (Amersham Pharmacia Biotech) per 20-l reaction using the random primed DNA labeling kit (Roche Molecular Biochemicals). Hybridizations were performed in 50% formamide, 0.12 M Na 2 HPO 4 , 0.25 M NaCl, and 7%(w/v) sodium dodecyl sulfate (SDS) at 43°C overnight. The blots were washed twice, each for 10 min, at room temperature in 30 mM sodium citrate buffer, pH 7.0, containing 0.3 M NaCl and 0.1% SDS. A third wash was conducted at 65°C, for 10 min, in 3 mM sodium citrate buffer, pH 7.0, containing 30 mM NaCl and 0.1% SDS. To check RNA loading in each lane, some membranes were stripped and rehybridized with probes synthesized from a human actin cDNA (CLONTECH, Palo Alto, CA).
Rapid Amplification of cDNA Ends (RACE)--DNA sequences at the 5Ј and 3Ј ends of clone 17a cDNAs were determined by RACE using the rat brain Marathon-Ready cDNA (whole brains pooled from normal Harlan Sprague-Dawley males, 10 -12 weeks of age; CLONTECH). Briefly, a 25-l PCR reaction contained the clone 17a-specific primer (0.2 M), the CLONTECH adaptor primer (0.2 M), 0.25 ng of Marathon-Ready cDNA, KlenTaq PCR reaction buffer, dNTPs (0.2 mM each), and Advantage KlenTaq polymerase mix (CLONTECH). Thermocycles were programmed as follows: 94°C, 1 min; 5 times (98°C, 10 s and 72°C, 1 min); 5 times (98°C, 10 s and 70°C, 1 min); 25 times (98°C, 10 s and 68°C, 1 min); 68°C, 2 min and kept at 4°C thereafter until further processed. The sense primer 5Ј-GAGAGATGGCTCAGCGGT-TAAGAGCAC-3Ј was used as the gene-specific primer 2 (GSP2) for 3Ј-RACE. The antisense primer 5Ј-GAAGAGGGCATCAGATCTC-AT-TACAGATGG-3Ј was used as GSP1 for 5Ј-RACE. The amplification products were electrophoresed on 4% agarose gels and transferred to a Zeta-Probe GT membrane (Bio-Rad) for Southern blot analysis. Biotinylated 5Ј-CCCAGCAACCACATGGTGGCTCACAACC-3Ј, a sequence between GSP1 and GSP2, was used as a probe for Southern blot analysis. The RACE products corresponding to the specific bands on the Southern blot were extracted from the agarose gel matrix with the Prep-A-Gene DNA purification kit (Bio-Rad), ligated into a pCR2.1 vector (Invitrogen), and transformed into Max Efficiency DH5␣ competent cells (Life Technologies, Inc.). Ampicillin-resistant colonies were characterized by PCR amplification with a variety of primer combinations and DNA sequencing.
In Situ Hybridization--For in situ hybridization, five 4VO-and five sham-treated animals were examined. Rat forebrains were rapidly frozen in isopentane chilled with dry ice and serially sectioned at Ϫ20°C in the coronal planes throughout the rostral-caudal axis of the hippocampus. Sixteen-micron-thick sections were thaw-mounted onto gelatin-coated slides and stored at Ϫ80°C until processed for in situ hybridization. Regions evaluated included only those that contained the same coronal sections of the dorsal hippocampus. BamHI-EcoRI fragments of clone 17a were constructed by the expand high fidelity PCR system (Roche Molecular Biochemicals). The 131-bp fragments were inserted into the plasmid vector pT7/T3-18 (Ambion, Austin, TX) with T4 DNA ligase (Life Technologies, Inc.). Vectors containing the insert were linearized with EcoRI or BamHI separately at 37°C, purified with the QIAquick PCR purification kit (Qiagen), and used as the templates to synthesize the probes. Single-stranded RNA probes were synthesized in the presence of 35 S-UTP using the riboprobe in vitro transcription system (T3/T7; Promega). Antisense and sense RNA probes were transcribed with either T3 or T7 RNA polymerase corresponding to the promoter region flanking the insert sequence in the T3/T7-18 vector. After in vitro transcription, RNA probes were treated with RNase-free DNase I to remove the template DNA.
The hybridization was conducted with the SureSite hybridization reagents kit (Novagen, Inc. Madison, WI) following the manufacturer's instructions. Briefly, frozen tissue sections were fixed with 4% paraformaldehyde for 20 min and permeabilized with 1 g/ml proteinase K for 10 min at room temperature. Each slide was covered with 80 l of hybridization buffer containing 10 7 cpm of RNA probes. Hybridization was performed for 18 h at 50°C in a humidified chamber. Background controls were performed on 4VO sections using the sense RNA as a probe. For experimental controls, sections of sham-operated animals were hybridized with the antisense probe. After hybridization, unbound probes were digested with RNase A (20 g/ml) for 30 min at 37°C. The post-hybridization washes were all performed at 50°C. Air-dried slides were exposed to Hyperfilm-␤MAX x-ray film (Amersham Pharmacia Biotech) for 1 h and developed manually in Kodak developer D-19. For higher resolution, sections were covered with 50% Kodak NTB-2 emulsion and exposed for 2.5 h in the dark. The slides were developed, counter-stained with hematoxylin and eosin (H&E), mounted with paramount, and photographed using both light-and dark-field microscopy.

RESULTS
Identification of Clone 17a--To identify differentially expressed genes in response to transient cerebral ischemia, poly(A) RNAs prepared from the dorsal hippocampus of 4VOand sham-operated rats were compared by differential display analysis. We detected both up-and down-regulated transcripts. However, we focused on an ischemia-induced transcript, clone 17a. Fig. 1 displays the result of DD-PCR that contained clone 17a. The products from each DD-PCR reaction were loaded in two lanes. Each lane contained 70 -150 bands. Interestingly, the arbitrary primer sequence was found at both ends of the amplified fragment of clone 17a (Fig. 2). Apparently, the same random primer was used as both 5Ј and 3Ј primers in the DD-PCR, and the anchored oligo(dT)primer was left unused. Conceivably, amplification of clone 17a fragments occurred in all DD-PCR reactions containing the random primer 5Ј-AATCGGGCTG-3Ј, no matter which oligo(dT) primer was used.
Northern Blot Analysis--To confirm the gene expression patterns observed in differential display analysis, the clone 17a fragment was used as the template for probe synthesis in Northern blot analysis. Clone 17a transcripts showed an approximately 35-fold increase in 4VO animals versus sham (Fig.  3). The size range of the transcripts was 160 -350 bases. Data base search of clone 17a in GenBank TM and GenEMBL TM using the BLAST and FastA programs revealed regions of homology to rodent B 2 sequences and a rat somatotropin intron. To check whether somatotropin mRNA is elevated in the 4VO rat, the fifth exon (1909 -2109) of rat presomatotropin gene (Gen-Bank TM accession number J00740 or V01238) was used as a probe for Northern blot analysis. Hippocampal total RNA was loaded, 10 g per lane, and probed with 32 P-labeled fifth exon fragments generated by PCR. A higher level of somatotropin expression was not observed in 4VO rats (data not shown). On the other hand, B 2 sequences are believed to be spread throughout the genome by retrotransposition and are found in both introns and exons of other genes (23).
Clone 17a Matches Characteristics of B 2 Sequences--To get full-length cDNA, the sequence of clone 17a obtained from DD-PCR was used to design primers for 5Ј-and 3Ј-RACE (Fig.  2). The 5Ј-RACE was conducted using GSP1. Heterogeneity at the 5Ј end was observed in both length and base composition (data not shown); however, the consensus sequences started at GAGATG (Fig. 2). By using GSP2, the 3Ј-RACE produced multiple sequences as well (Fig. 4) 1. Identification of clone 17a, whose expression was upregulated in ischemic rats. Total RNA was isolated from CA1 regions in hippocampi of 4VO-and sham-operated rats. RNAs from every four animals were combined for the first strand cDNA synthesis. cDNA was generated with an anchored primer, T12C, and used as the template for PCR amplification. The products from each PCR reaction were resolved in two adjacent lanes on a 6% polyacrylamide gel containing 8.3 M urea. Total eight 4VO and eight sham animals were used in DD-PCR. The arrow indicates the PCR fragments of clone 17a. The numbers on the left margin refer to size markers in bases (b).

FIG. 2.
Nucleotide sequence of clone 17a. The sequence of the arbitrary primer used in differential display analysis is in bold type. GSP1 and GSP2 (underlined) were paired with the adaptor primer used in 5Ј-and 3Ј-RACE, respectively. The direction of primer extension is indicated by the arrows.
(ORF) of either 29-or 40-amino acid peptides. Further analysis of these transcripts revealed promoter regions, box A and B (24), of RNA polymerase III (polIII) within the transcripts. The poly(A) addition signal, transcription stop sequence, and poly(A) tail was found at the ends of most 3Ј sequences. The sequence results indicate that the transcripts are products of polIII and match the characteristics of rodent B 2 sequences.
Hypothetical Peptide--The ORFs seen in clone 17a gene family suggest that peptides could be generated by these sequences. Proteins coded by small open reading frames (Ͻ100 amino acids) belong to a number of important categories, such as ATPase modulators, stress proteins, transcriptional regulators, and antioxidants (for review, see Ref. 25). To see whether the clone 17a gene is translated, a rabbit polyclonal antibody was generated against the synthetic peptide CSS-RGHEFNSQQPHGGSQPSVKRSD deduced from the clone F3 sequence (Fig. 4). The full-length peptide is predicted to have a molecular mass of 4.4 kDa and an isoelectrical point of 8.7. To search for the hypothetical peptide, proteins were extracted from the hippocampus of 4VO and sham rats and analyzed, 80 g per lane, by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting. The peptide was not detected by the antibody raised against the synthetic peptide. Transfection of 293 cells with the clone F3 sequence in a pcDNA3.1 vector yielded no detectable peptide expression. To date, translation of B 2 sequences has not been reported.
Clone 17a Transcripts Are Localized in Pyramidal Neurons--To examine the tissue-specific expression of clone 17a in rat brain, an antisense riboprobe was transcribed from clone 17a and hybridized with rat brain sections. After x-ray film exposure, strong expression was detected mainly in the CA1 region of 4VO hippocampi (Fig. 5A). Some expression was observable in the cortex of 4VO brains as well. Brain sections of sham-treated rats showed a background level expression with the antisense probe (Fig. 5B). The specificity of the hybridization reaction was confirmed by comparison of signals generated from the antisense probe (Fig. 5A) and the sense probe (Fig. 5C) on 4VO brain sections. By using emulsion, higher resolution (1000ϫ) was achieved (Fig. 6). In the CA1 region of a 4VO hippocampus, silver grains were found in the cytoplasm of pyramidal neurons. Since sections were counter-stained with H&E, nuclei are shown as dark gray areas in the bright-field image. In dark-field microscopy, silver grains reflect light and are shown as bright particles in a dark background. In the same brain region of sham animals, silver grains were also observed around nuclei but with a much less density. The same observation was made in brain sections derived from five individual animals. This result confirms the differential expression of clone 17a in 4VO rats discovered by RT-PCR and Northern blot analysis and is consistent with the fact that B 2 poly(A) RNAs are mainly distributed in the cytoplasm (26).
Animals Treated with LY231617--The antioxidant LY231617 has been shown to reduce delayed neuronal death in the CA1 region (27). Although LY231617 possesses antioxidant activity as one of its properties, the precise mechanism through which LY231617 prevents neuronal injury after global ischemia is not clear. To establish that clone 17a gene is closely related to delayed neuronal death, rats were treated with the compound as described above. At 24 h post-ischemia, com- pound-treated animals had a background level expression of clone 17a much less than those of untreated 4VO rats (Fig. 7). This result suggests that clone 17a expression was closely associated with neuronal degeneration. The same blot also demonstrates that a higher level expression of clone 17a occurred around 24 h post-ischemia and was not observed at 72 h post-ischemia (Fig. 7). Decreased expression of clone 17a beyond 24 h post-ischemia was also observed in the CA1 region probed with the antisense RNA (Fig. 8), which suggests that induction of clone 17a transcripts was an early event in delayed neuronal death. Interestingly, an increasing amount of B 2 transcripts in the dentate gyrus was detected over the same period (Fig. 8).
Clone 17a Was Induced by Glutamate and 6-OHDA in CGNs--Delayed neuronal death has the morphological and biochemical characteristics of apoptosis (6,8,28). We postulated that clone 17a may be involved in apoptosis. B 2 induction has been observed in the apoptotic cell death of PC-12 cells upon nerve growth factor deprivation (29). To investigate the possible involvement of clone 17a in apoptosis, in vitro models of neuronal apoptosis in cultured CGNs were examined for the expression of clone 17a. CGNs from early postnatal rats can be maintained in the culture medium containing serum and high potassium (25 mM) (30). The apoptotic cell death of cultured CGNs can be induced by switching potassium to a lower but more physiological concentration (5 mM) (31). Relatively low concentrations of glutamate or 6-OHDA, as applied in this FIG. 5. In situ hybridization of clone 17a in brain sections of 4VO and sham animals. Antisense and sense probes were in vitro transcribed from clone 17a DNA in presence of 35 S-UTP. Expression of transcripts in brain sections of 4VO and sham animals was visualized on x-ray films. A, a 4VO brain section hybridized with the antisense probe. B, a brain section of sham-treated animal probed with the antisense probe. C, a 4VO brain section hybridized with the sense probe. These results were replicated with brain sections from five separate animal experiments.
FIG. 6. Expression of clone 17a in the CA1 regions of hippocampi. Brain sections from 4VO-and sham-treated rats were in situ hybridized with the 35 S-labeled antisense riboprobes derived from the clone 17a DNA fragment. The sections were covered with 50% Kodak NTB-2 emulsion, exposed for 2.5 h in the dark, and counter-stained with H&E. The CA1 regions were photographed with both light-and dark-field microscopy. Clone 17a transcripts were visualized as the light grains in the dark-field images. These results were replicated with brain sections from five separate animal experiments. Magnification, ϫ 1000.  35 S-UTP and was in situ hybridized with brain sections of 4VO rats sacrificed at different time points after ischemia as indicated. Expression of transcripts was visualized on x-ray films. These results were replicated with brain sections from three separate animal experiments. study, are also causative factors of CGN apoptosis (32,33). Glutamate is a known potent excitotoxin critical for the neuronal damage in ischemic stroke (34). 6-OHDA is proposed to be the endogenous toxin involved in the neuronal damage of Parkinson's disease (35). In contrast to low K ϩ model, glutamateinduced apoptosis does not require RNA or protein synthesis but does require post-translational activation (32).
In Northern blot analysis, total RNA isolated from apoptotic and control CGNs was loaded 10 g per lane. Induction of clone 17a transcripts was found most significant in the glutamateand 6-OHDA-induced cell death (Fig. 9, lanes 3 and 4). However, a higher expression level of clone 17a was not detected in the low K ϩ -induced cell death (Fig. 9, lane 2). It is known that apoptosis of CGNs can be induced by different stimuli and environmental conditions and that different intracellular mechanisms are likely to be involved (36). Results in this study indicate that B 2 expression was associated with certain types of apoptosis. Since it is easy to detect, B 2 sequences could be used as a marker to detect or distinguish different types of cell death. Further experiments are needed to establish its specific expression in apoptosis and whether it is apoptotic or protective in delayed neuronal death. DISCUSSION The primary goal of this study is to identify genes involved in ischemic cell death in the rat 4VO model. We are interested, specifically, in early gene products that mediate or control delayed neuronal death. The rats that we used in this study were sacrificed 24 h after the 4VO treatment. At this time point, there is no histological evidence of CA1 neuronal damage (16). Extensive neuronal damage occurs at 72 h after ischemia.
In mammalian genomes, there are families of highly repeated DNA sequences, including long and short interspersed elements (37). B 2 sequences are a family of short interspersed elements consisting of about 10 5 related sequences dispersed throughout the genome (23,38). The consensus region of B 2 sequences is 180 bp long and contains the RNA polymerase III promoter, poly(A) addition signal, and transcription stop signal (24). A member of the B 2 family deviates 3-5% in the consensus region (38). Sequences at the 3Ј end following the consensus sequence are most variable. On Northern blots, B 2 sequences are generally displayed as a smear in the range of 200 -600 bp (39). The function of B 2 sequences has not been established, although a general regulatory role in gene expression and RNA processing has been suggested (38). The RACE products of clone 17a contained multiple sequences with variable 5Ј-and 3Ј-flanking regions and possessed all the structural characteristics of B 2 sequences. Northern blot analysis showed that the size range of clone 17a expressed in 4VO rats was between 160 and 540 bp. Sequence analysis revealed that the RACE products contained the internal promoter regions of polIII and a consensus region of approximately 160 bp with less than 4% deviation. The length of the poly(A) tail obtained by RACE is not the authentic length of poly(A) tail but the length of the oligo(dT) primer used for first strand cDNA synthesis. Although these transcripts contained ORFs, the translated product was not detected. B 2 poly(A) RNAs are believed to be the final functional products involved in regulation of mRNA processing, transport, stability, and translation (26,40,41). In situ hybridization demonstrates that these transcripts were predominantly located in the neurons within the hippocampal CA1 region, indicating that B 2 sequences may be closely related to the distinct effect of 4VO on pyramidal neurons. The specific association of B 2 sequences in delayed neuronal death was further manifested in LY231617-treated 4VO animals, where the compound-protected neurons in the CA1 region showed background levels of clone 17a transcripts.
Although expression of many genes is found to be altered during ischemic injury (10), the mechanism of delayed neuronal death is still vague. Many histological and biochemical observations in delayed neuronal death match features of apoptosis (6,8,28). Induction of several regulatory genes of apoptosis is associated with delayed neuronal death. For example, both mRNA and protein of Bax are expressed prior to delayed neuronal death in the CA1 region (42,43). The antiapoptotic gene bcl-2 is expressed in neurons that have survived delayed neuronal death (44). Increased levels of nuclear factor B in the nuclei of pyramidal neurons are observed after ischemic injury (45). Nuclear factor B has been shown to both promote (46) and suppress (47) apoptosis depending on conditions and cell types. Finally, caspase-3 and other caspase activities are also observed in delayed neuronal death (8). The finding that clone 17a was expressed in the glutamate-and 6-OHDA-induced apoptosis of CGNs is a suggestion of B 2 involvement in apoptosis and further indicates that delayed neuronal death was mediated by apoptosis. However, whether B 2 sequences promote neuronal apoptosis or protect neurons from delayed neuronal death needs further investigation. Since the amount of clone 17a transcripts decreased at 48 and 72 h in situ (Fig. 8) while cell death increases in the same time frame (16), high expression of the transcripts at 24 h post-ischemia is unlikely the consequence of cell death. It is known that DNA fragmentation caused by methyl methanesulfonate does not induce B 2 expression (48). The differences in clone 17a expression among the low K ϩ -, glutamate-, and 6-OHDA-induced cell death in CGNs may be due to the severity of apoptosis or may reflect the differences in their apoptotic pathways. It is known that glutamate-or 6-OHDA-induced apoptosis in CGNs is p53dependent (49,50) and involves loss of mitochondrial function (51,52), whereas low K ϩ -induced apoptosis is p53-independent (53) and does not depend on a loss of mitochondrial function (54). It would be of interest to know whether B 2 sequences play a specific role in the apoptotic cascade and whether altered B 2 expression can affect apoptosis. B 2 expression can be stimulated under various conditions. The level of B 2 transcripts is high in early embryos (40,55) and virally transformed cells (56) and is low or absent in differentiated cell types (57). Transcription of B 2 sequences can be induced in somatic cells by heat shock (58,59) and serum stimulation (60). Interestingly, apoptosis has been observed under similar circumstances (61,62). B 2 sequences are transcription products of polIII. The possible involvement of B 2 sequences in apoptosis can be unveiled further by examining induction of polIII activity in various viral transformations. FIG. 9. Northern blot analysis of clone 17a expression in cultured CGNs. CGNs were cultured as described under "Experimental Procedures." Total RNA was isolated and loaded 10 g per well on a 1.2% agarose gel. The separated RNAs were transferred to a nylon membrane and probed with the 32 P-labeled clone 17a fragments (upper panel). Lane 1, cultured CGNs without a special treatment were used as a control. Lane 2, CGNs were treated with 5 mM KCl in the conditional medium overnight. Lane 3, CGNs were treated overnight in 30 M glutamate. Lane 4, CGNs were treated with 50 M 6-OHDA overnight. The same blot hybridized with the ␤-actin probe is shown as a loading control (lower panel). The size marker on the left margin is presented in bases (b).
Adenovirus E1A (63), simian virus 40 T antigens (64), X-protein of the human hepatitis B virus (65), and Tax protein of the human T-cell leukemia virus type 1 (66) are known to trigger apoptosis in mammalian cells. On the other hand, cells transformed with these agents exhibit an increased level of polIII transcripts (67)(68)(69)(70).
During transcription of B 2 sequences, polIII activity is modulated by transcription factors TFIIIB and TFIIIC (71). The rapid increase in B 2 RNA levels is most likely mediated by these factors. SV40 T antigen is a transactivator of polIII (72), which increases activity of TFIIIC by changing its abundance or phosphorylation state (73). TFIIIC has been shown to be the rate-limiting factor in the formation of polIII initiation complexes (71) and probably the target of trans-activators (74). The ability of X-protein to activate polIII is due to its interaction with TFIIIB rather than TFIIIC (69). During X-protein activation, polIII catalytic activity remains unchanged. TFIIIB is also the target of Tax, which enhances polIII transcription by increasing the effective concentration of TFIIIB (70). However, further studies will be required to determine the upstream activator of clone 17a transcription and its mode of activation in the rat hippocampus after the 4VO treatment.
The action of B 2 sequences can be predicted from their ability to serve as retroposons (75) and to interact with mRNAs (41). B 2 sequences are found to be a cis-acting regulator of gene expression (76). Insertion of Alu sequences, the human counterpart of B 2 sequences, found in the coding region of caspase-8 in tumor cells reduces the effect of the protein on apoptosis (77). B 2 poly(A) itself could play a role in selective inhibition of mRNA translation (78). B 2 sequences may also regulate the stability of certain RNAs, such as mRNAs of c-Fos, c-Myc, and tumor necrosis factor (41). All these could serve as important mechanisms for gene regulation in delayed neuronal death.
Induction of short interspersed elements in mouse, rabbit, and human suggests that these sequences serve a common function in the mammalian system (59). Their expression could be part of a defensive mechanism. This study suggests a new direction for seeking the function of B 2 sequences and may facilitate our effort in elucidating the mechanism of delayed neuronal death in ischemic injury.