Expression of GTP-dependent and GTP-independent tissue-type transglutaminase in cytokine-treated rat brain astrocytes.

Tissue-type transglutaminases (TGases) were recently shown to exert dual enzymatic activities; they catalyze the posttranslational modification of proteins by transamidation, and they also act as guanosine triphosphatase (GTPase). Here we show that a tissue-type TGase is expressed in rat brain astrocytes in vitro, and is induced by the inflammation-associated cytokines interleukin-1β and to a lesser extent by tumor necrosis factor-α. Induction is accompanied by overexpression and appearance of an additional shorter clone, which does not contain the long 3′-untranslated region and encodes for a novel TGase enzyme whose C terminus lacks a site that affects the enzyme's interaction with guanosine triphosphate (GTP). Expression of two clones revealed that the long form is inhibited noncompetitively by GTP, but the short form significantly less so. The different affinities for GTP may account for the difference in physiological function between these two enzymes.

Tissue-type transglutaminases (TGases) 1 were recently shown to have dual activities. They act as G-proteins (1-3) and catalyze the posttranslational modification of proteins by transamidation of available glutamine residues (4). They are also involved in repair processes, where their primary role is to stabilize, by cross-linking, proteins in the vicinity of the wound, thereby increasing tissue resistance to proteolytic activity and to changes in environmental conditions (5)(6)(7)(8).
Recent studies of developing and regenerating nerves have led to an increasing appreciation of the role of Ca 2ϩ -associated changes mediated by TGases (9 -11), and it was suggested that these are associated with glial cell formation (12,13). In developing rat cerebellar tissue-type TGase, immunoreactivity is detectable in axons of the external granular layer during the early postnatal period (14,15). In PC12 cells, induction of neurite formation is accompanied by an elevation in TGase activity (16). Peripheral nerve transection results in an increase of TGase activity close to the site of injury, probably associated with nonneuronal elements (13).
TGase activity has been detected in the brain in cytosolic fractions, in synaptic vesicles, and as a membrane-associated enzyme (9, 16 -21). It was proposed that either different types of TGases are expressed in the nervous system or posttranslational modifications alter the localization of the enzyme (22). It was further suggested that in the central nervous system the regulation of TGase function depends on substrate selection and availability (10). These and other studies illustrate the importance of regulation of TGases in general and in the nervous system in particular (23).
We have shown that rat astrocytes, transplanted into transected optic nerves of syngeneic adult rats at the time of transection, facilitate recovery of the response to light, provided that the transplanted astrocytes have been preactivated with a combination of two cytokines, tumor necrosis factor-␣ (TNF-␣) and interleukin-1␤ (IL-1␤). 2 We suggested that such preactivation might induce even mature astrocytes to be growth-permissive, for example, by up-regulating intracellular and extracellular substances needed for growth and downregulating growth-inhibitory substances. Among such substances are growth factors, enzymes, extracellular matrix molecules, and cell adhesion molecules. TGase may well be included among these substances, since it was shown to be facilitatory for regeneration (24,25), and might be a product of astrocytes induced by the two cytokines.
The aim of this study was to find out whether rat brain astrocytes express TGase and, if so, what type and whether or not the enzyme is up-regulated in response to cytokines that induce acquisition of growth-supportive characteristics. We show that in mature astrocytes low levels of TGase transcript (3.7 kb) are expressed. The cytokines IL-1␤ and, to a lesser extent, TNF-␣ up-regulate TGase transcript levels and induce expression of an additional 2.4-kb transcript. Cloning of rat astrocyte TGase cDNAs revealed two almost identical fulllength TGase-encoding cDNAs, exhibiting marked homology with tissue-type TGases. One form is shorter and encodes for an enzyme with a different C terminus than the longer form and consequently with a lower guanosine triphosphate (GTP) dependence and an increased Ca 2ϩ dependence. This shorter form could not be detected either in untreated astrocytes or in other tissues expressing tissue-type TGase. Its detection in cytokine-treated astrocytes suggests that either it is unique to these cells or the conditions of its induction in other tissues are not yet known.

EXPERIMENTAL PROCEDURES
Preparation of Rat Brain Astrocyte Cultures-Astrocyte cultures were prepared by a modification of the procedure of McCarthy and deVellis (26). Cells dissociated from the cerebral cortex of 2-day-old rats were cultured in poly-D-lysine-coated tissue culture flasks (two brains/85 cm 2 flask) containing Dulbecco's modified Eagle's medium, 2 * This study was supported by the Daniel Heumann Fund for Spinal Cord Research (USA) and the Alan T. Brown Foundation of Nerve Paralysis (USA). 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.
Polymerase Chain Reaction (PCR) Analysis of TGase in Rat Brain, Astrocytes, and Oligodendrocytes-Total RNA was prepared from neonatal rat brain and from mature astrocytes and oligodendrocytes by the RNAzol B method with the aid of a Biotec Lab Kit. Primers were synthesized on the basis of conserved amino acid sequences in the active site region of the TGase family; the downstream primer was 5Ј-CGG-GTGGG(C)A(G)AG(T)GCCCAGGCAGCGCAGCACTGTGA(GCT)A(T)-GGG(T)CACAG(C)CAGCAAAGACCCAGCACTGGCCATA-3Ј, and the upstream primer was 5Ј-GGG(T)CAGTTTGAAGATGGG(C)ATCCT-GGA-3Ј. The reaction was carried out as follows; 1 g of total RNA was incubated at 65°C for 10 min and then on ice for 5 min. The RNA was added to a reverse transcriptase reaction mixture and incubated for 20 min at 42°C, followed by 5 min at 95°C for enzyme and RNA denaturation. The reaction was then transferred to ice. PCR mixture was added to the cDNA and covered with mineral oil. The reaction was transferred to a Cetus PCR apparatus for 30 cycles of 3 min each (1 min of annealing at 72°C, 1 min of elongation at 72°C, and 1 min of denaturation at 95°C). The products were precipitated and purified by electrophoresis on a 1% agarose gel.
In Situ Hybridization in Rat Astrocytes-Cells grown on coverslips were fixed for 20 min by 4% paraformaldehyde in 4% sucrose in phosphate-buffered saline (PBS) (warmed to 37°C). The coverslips were then rinsed in PBS, transferred to 70% ethanol and stored at 4°C until use. Before hybridization, cells were rehydrated by sequential 10-min immersion in PBS containing 5 mM MgCl 2 , in 200 mM Tris containing 100 mM glycine, pH 7.5, and in 50% deionized formamide in 2 ϫ saline sodium citrate (SSC) at 60°C. Prehybridization was carried out for 1-3 h at 50°C in hybridization buffer (100 l/coverslip) consisting of 50% formamide, 300 mM NaCl, 20 mM Tris, pH 8.0, 5 mM EDTA, 1 ϫ Denhardt's solution, 10% dextran sulfate, 10 mM dithiothreitol, and 10 mM vanadyl sulfate ribonucleoside complex (BioLabs). The RNA probe was mixed with 1 mg/ml salmon sperm DNA, and the solution was incubated at 95°C for 3 min and added to the hybridization buffer (20 ml) at a final probe concentration of 50 -100 ng/coverslip. Hybridization was allowed to proceed for 16 h at 50°C in a humid chamber. The coverslips were then floated off in 1 ml of 300 mM NaCl, 20 mM Tris, pH 8.0, 5 mM EDTA, gathered on a rack, and washed twice with the same buffer for 15 min each at 50°C. To allow hydrolysis of the nonspecifically bound probe (background), coverslips were treated with 20 mg/ml RNase in 500 mM NaCl and 10 mM Tris, pH 8.0, for 30 min at 37°C, followed by three 15-min rinses at 50°C in 150 mM NaCl, 20 mM Tris, pH 7.5, and 5 mM EDTA, and two 15-min rinses at 50°C in 15 mM NaCl and 20 mM Tris, pH 7.5. Cells were incubated for 30 min at room temperature with antidigoxygenin/alkaline phosphatase-conjugated antibody diluted 1:1000 (Boehringer, Mannheim). At the end of the reaction, cells were washed three times for 15 min each at room temperature with 150 mM NaCl and 100 mM Tris, pH 7.5, equilibrated for 5 min in buffer containing 100 mM Tris, pH 9.5, 100 mM NaCl, and 50 mM MgCl 2 and developed with nitro blue tetrazolium/bromochloroindolyl phosphate solution (0.34 mg/ml nitro blue tetrazolium and 0.175 mg/ml bromochloroindolyl phosphate in the same buffer) at room temperature. Color development was monitored by bright-field microscopy and stopped by the addition of 10 mM Tris, pH 7.5, and 1 mM EDTA (15-90 min).
Northern Blot Analysis of Rat Brain TGase-Total RNA was prepared by the RNAzol B method with the aid of a Biotec Lab Kit. The procedure was performed at 4°C or on ice. Poly(A) ϩ RNA was purified by oligo(dT) magnetic beads with the aid of the Dynabeads Kit (Dynal), and quantified by absorbance at 260/280 nm. The poly(A) ϩ RNA was denatured at 65°C for 5 min in denaturing buffer. RNA was then size-fractionated by electrophoresis through a 1% agarose-formaldehyde gel and transferred to Gene Screen Plus membranes (DuPont). Gels were blotted in 10 ϫ SSC and left overnight. Blots were baked for 1 h at 80°C in a vacuum. The relative amounts and quality of the transferred RNA were assessed by hybridization with a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe. Prehybridization took place at 42°C for 24 h in buffer containing 50% formamide, 5 ϫ SSC, pH 7.2, 5 ϫ Denhardt's buffer (0.1% polyvinylpyrrolidine, 0.1% Ficoll, and 0.1% bovine serum albumin), 1% sodium dodecyl sulfate (SDS), 50 mM Na 2 HPO 4 /NaH 2 PO 4 , pH 7.0, and 300 mg/ml denatured herring sperm DNA. The blots were hybridized at 42°C for 24 h in buffer containing the same materials but with a smaller proportion of Denhardt's buffer (1 ϫ) and Na 2 HPO 4 /NaH 2 PO 4 , pH 7.0 (20 mM), in the presence of the 32 P-labeled rat astrocyte PCR product. Blots were washed twice at room temperature in 2 ϫ SSC, 0.1% SDS for 10 min each, and in 0.1 ϫ SSC, 0.1% SDS at 60°C for 60 min. The blots were then exposed to x-ray film.
In Vitro Transcription-The plasmid containing TGase cDNA was linearized at the 3Ј-end of the region to be transcribed. Complete digestion was determined by electrophoresis on a 1% agarose gel. The DNA was extracted with phenol/chloroform and precipitated in the presence of 0.3 M sodium acetate, pH 5.2, and 70% ethanol. The DNA was then dissolved at a concentration of 1 g/l. The transcription reaction mixture consisted of 18 l of sterile double-distilled water, 10 l of 5 ϫ transcription buffer (200 mM Tris-HCl, pH 7.5, 30 mM MgCl 2 , 10 mM spermidine, 50 mM NaCl), 5 l of 0.1 M DTT, 2 l of 20 g/ml RNasin, 10 l of 2.5 mM ribonucleotide mix, 5 l of 1 g/l template DNA, and 1 l of T 7 or T 3 polymerase (70 units/l). The reaction was incubated for 2 h at 37°C, and 2 mg/ml RNase-free DNase was added for an additional 15 min at 37°C. The RNA was extracted as described and dissolved in 30 l of sterile double-distilled water. The RNA sample was analyzed for quantity and quality by electrophoresis on 1% agarose.
Southern Blot Analysis-Rat DNA (15 g) prepared from the spleen was cleaved with various restriction enzymes and then fractionated by electrophoresis on a 0.8% agarose gel. The gel was denatured with 1.5 M NaCl, 0.5 N NaOH for 45 min at room temperature and then neutralized with 0.5 M Tris, pH 7.4, 1.5 M NaCl for 45 min. Southern blots were prepared, transferred onto Gene Screen paper in the presence of 10 ϫ SSC, and kept overnight at room temperature. The Gene Screen paper was then rinsed in 0.4 N NaOH for 1 min, washed with 0.2 M Tris, pH 7.5, 2 ϫ SSC and dried at room temperature. Prehybridization was carried out in 1% SDS, 1 N NaCl, and 10% dextran sulfate for 3 h at 65°C, followed by overnight hybridization in the presence of 200 g/ml calf thymus DNA and 32 P-labeled DNA probe. The blot was then washed twice with 2 ϫ SSC, 1% SDS for 30 min at 65°C and twice with 0.1 ϫ SSC at room temperature, then exposed to x-ray film.
Human Embryonic Kidney (HEK)-293 Cell Transfection and Fractionation-HEK-293 cells were transfected with CaPO 4 as described elsewhere (27), with several modifications. Briefly, cells were grown as a monolayer in Dulbecco's modified Eagle's medium/10% inactivated fetal calf serum, 100 units/ml penicillin-streptomycin at 37°C, 5% CO 2 , and split 1:5 24 h prior to transfection. Plasmid DNA (10 -20 g) was mixed with 50 l of 2.5 M CaCl 2 , and double-distilled water was added to a final volume of 0.5 ml. To the DNA solution 2 ϫ HBS was gently added until a precipitate could be seen (0.4 -0.5 ml), and the mixture was immediately poured into a 10-cm Petri dish for overnight incubation at 37°C, 5% CO 2 . Half of the medium was then replaced. After incubation for 24 h at 37°C, the cells were washed with PBS, 0.5 mM EDTA, 1:100 protease inhibitors (1 mM spermidine, 25 mg/ml aprotinin, 25 mg/ml leupeptin, 5 mg/ml pepstatin) and scraped from the Petri dish on ice. Cells were centrifuged (400 ϫ g, 4°C, 10 min), and the pellet was resuspended with 50 mM Tris acetate, pH 7.2, and 1:100 protease inhibitors. Cells were homogenized on ice with a Teflon glass homogenizer for 5 min and centrifuged (800 ϫ g, 4°C). The resulting soluble fraction was removed and further centrifuged (14,000 ϫ g, 4°C, 45 min). The soluble fraction was removed and stored at Ϫ70°C. The particulate fraction was washed twice and resuspended in 50 mM Tris acetate, pH 7.2, and then homogenized on ice for 5 min. Membrane particles were precipitated by centrifugation (45 min, 14,000 ϫ g) and resuspended with SDS-polyacrylamide gel electrophoresis loading buffer.
TGase Assay-The kinetics of TGase activity were assayed by measuring incorporation of putrescine into N,N-dimethylated casein. The reaction mixture contained 50 mM Tris-HCl, pH 8.0, 5 mM dithiothreitol, 5 mM CaCl 2 , 0.37 mM putrescine (1:5, [ 3 H]putrescine:putrescine), increasing amounts of N,N-dimethylated casein (0.057 Ϯ 5.7 mg/ml) and samples of HEK-293 cytosolic fraction. The reaction mixture was incubated for 1 h at 37°C and transferred onto ice. In the GTP experiments, 0.075 mM putrescine was used. Cold trichloroacetic acid was added to a final concentration of 5% for 15 min. Samples were centrifuged (14,000 ϫ g for 5 min at room temperature), and the pellet was washed twice with 1 ml of 5% trichloroacetic acid and once with 100% ethanol. Samples were dried and resuspended in 200 l of 0.1 NaOH. Radioactivity was measured in 10 ml of scintillation liquid (40% Lumax, 60% xylene).
Preparation of Polyclonal Antibodies and Affinity Purification-Two cDNA clones encoding for rat brain TGase were cloned and sequenced. Synthetic peptides from the C terminus of the predicted amino acid sequence were synthesized in the Laboratory of Peptide Synthesis at the Weizmann Institute of Science, by the Merrifield solid-phase method. The sequences of the chosen peptides are YPQIMSDDCTPF-GLT from the short TGase nerve (s-TGN) protein and NFQCDKLKS-VKGYRN from the long TGase nerve (l-TGN) protein. The peptides were dissolved in 2 ml of PBS and conjugated to keyhole limpet hemocyanin. New Zealand White rabbits were immunized by intradermal and subcutaneous injections of conjugated peptide (1 mg) in complete Freund's adjuvant followed by three additional injections (1 mg) in incomplete Freund's adjuvant. The injections were administered as above, at 2-week intervals. Sera from immunized rabbits were checked by the enzyme-linked immunosorbent assay.
The synthetic peptides of l-TGN and s-TGN were bound to Affi-Gel 10 and Affi-Gel 15 (Bio-Rad), respectively (10 mg of peptide/ml of gel), by mixing overnight at 4°C in PBS, followed by stripping with 0.2 M glycine in PBS. The column was washed with elution buffer (PBS containing 0.2 M glycine, pH 2.7) and then equilibrated with PBS, pH 7.0. Serum immunoglobulins were loaded on the column after ammonium sulfate precipitation. The bound antibodies were eluted with elution buffer, and the eluate was neutralized with Tris, pH 8.0. The antibodies were dialyzed against PBS and stored at Ϫ20°C in the presence of glycerol.
Western Blot Analysis-The various preparations (detailed under "Results") were separated by SDS-polyacrylamide gel electrophoresis on 10% acrylamide slab gels. Following electrophoresis, proteins were transferred to a nitrocellulose filter for 2.5 h at 200 mA (in Tris-glycine).
The immune reaction was carried out as follows; the blot was incubated overnight in PBS containing 5% milk at 4°C and then with the affinitypurified antibodies in PBS containing 5% milk for 2 h at 37°C. This was followed by several washings with PBS containing 0.05% Tween 20, incubation with alkaline phosphatase-conjugated goat anti-rabbit IgG (Jackson) for 1 h at room temperature, several washings with PBS containing 0.05% Tween 20, and development with an ECL detection system (Amersham) for 1 min.
Expression in Baculovirus and Purification of 6 ϫ His Fusion Enzymes-For expression and purification of the enzymes we used the baculovirus 6 ϫ His expression and purification system (Pharmingen). The l-TGN cDNA was subcloned into pAcHLT plasmid containing a His tag-coding region in the 5Ј-region. The plasmid was then cotransfected with the linearized Baculogold DNA to Sf9 cell line to construct a recombinant baculovirus. The Sf9 cells were then infected with the recombinant baculovirus for 72 h. The cells were harvested, lysed, and centrifuged for 1 h at 14,000 ϫ g. The His-tagged enzyme was affinitypurified using Ni-NTA-agarose.

Expression of TGase mRNA in Mammalian Central Nervous
System-PCR analysis of RNA from whole brains and astrocytes of 2-day-old rats was performed using primers derived from the conserved active site region (see "Experimental Procedures"). PCR products of 300 bp were observed in both brain and astrocytes (Fig. 1A, lanes 1 and 2, respectively). These products were subcloned in a Bluescript plasmid, and their sequences were determined. The partial cDNA clones were found to be identical and showed 94% homology with mouse macrophage tissue-type TGase cDNA. Rat liver was also analyzed under the same experimental conditions and showed a PCR product of identical size.
The astrocytic origin of the observed TGase PCR product was further confirmed by in situ hybridization. Astrocyte primary cultures were prepared and seeded on 24-mm coverslips. Cells were fixed with paraformaldehyde and prepared for hybridization, as described under "Experimental Procedures." The cloned rat PCR (300 bp) product was transcribed in vitro, using digoxigenin-uridine triphosphate (DIG-UTP) in the ribonucleotide mixture, and the RNA probe was hybridized to the fixed astrocytes. As shown in Fig. 1B, a, high levels of astrocyte TGase mRNA were localized in the cytoplasm of the cell body and the processes. A sense RNA probe with the same specific activity as the antisense RNA probe showed no hybridization signal (Fig. 1B, b).
FIG. 1. PCR analysis of TGase in neonatal rat brain and primary rat brain astrocytes. The PCR procedure, conditions, and primers are described under "Experimental Procedures." A, total RNA was subjected to PCR using primers derived from the conserved active site region of TGase. The PCR product was analyzed on a 1% agarose gel (1 g/ml ethidium bromide) and visualized by UV light. Lane 1, rat brain; lane 2, rat astrocytes. HindIII-cleaved -DNA was used as a marker. B, the PCR product (lane 1) was used for in situ hybridization. Cultures of astrocytes were prepared as described under "Experimental Procedures" and processed for in situ hybridization, which was performed using the probe DIG-UTP RNA (derived from the PCR product) and alkaline phosphate anti-DIG antibodies for detection. In situ hybridization with antisense is shown in a and with sense in b.

FIG. 2.
Up-regulation of TGase mRNA in IL-1␤-treated astrocytes. Primary cultures of neonatal rat brain astrocytes were prepared. For Northern blot analysis, mRNA preparations from astrocytes treated with 800 units/ml recombinant human IL-1␤ or untreated astrocytes were blotted using the TGase 300-bp PCR product (Fig. 1) as the hybridizing probe. Arrows indicate long transcripts encoding for l-TGN and short transcripts encoding for s-TGN. GAPDH was used as a control for RNA quantity. Ribosomal RNA was used as a size marker.

IL-1␤ Up-regulates Transcription of TGase mRNA in Rat
Astrocytes-Northern blot analysis, using the 300-bp PCR product as a probe, revealed the presence of a 3.7-kb transcript encoding for TGase in primary cultures of rat brain astrocytes. We then examined the expression of TGase in rat astrocyte primary cultures treated for 48 h with IL-1␤. Untreated astrocytes served as a control. As shown in Fig. 2, up-regulated transcription of the 3.7-kb TGase mRNA occurred in the treated astrocytes; in addition, a smaller transcript of 2.4 kb appeared, which could not be detected in the control. Under similar experimental conditions, TNF-␣ showed a moderate effect, although in some cases the low molecular weight form was below detection level (data not shown).
Cloning of TGase from a cDNA Library of IL-1␤-treated Astrocytes-Because TGase mRNA was found to be up-regulated by the cytokines IL-1␤ and TNF-␣, and was more pronounced with IL-1␤, we constructed a cDNA library of mRNA preparations derived from IL-1␤-treated astrocytes and used it FIG. 3. Alignment analysis of rat astrocyte TGase cDNAs. A rat astrocyte cDNA library was prepared from primary cultures of rat brain astrocytes preincubated with IL-1␤. The library was screened with the rat astrocyte TGase PCR product as a probe, as described under "Experimental Procedures." Two TGase cDNAs were cloned, one of 3468 bp to clone the rat astrocyte TGase cDNAs. Use of a 300-bp TGase PCR product as a screening probe yielded two TGase cDNAs, a longer transcript (l-TGN) consisting of 3468 bp and a shorter one (s-TGN) of 2153 bp (Fig. 3A). The long cDNA clone contains a consensus polyadenylation signal (AAATAAA) that starts at nucleotide 3443, whereas the short clone, although polyadenylated, contains no such signal.
The two clones are identical except for differences in the 3Ј-coding and noncoding regions (Fig. 3, A and B). The s-TGN cDNA shows marked homology with l-TGN cDNA, except for a region in l-TGN between nucleotides 1882 and 2092, immedi-ately after the sequence CTTCAC (Fig. 3, A and B). In addition, here the long transcript contains a noncoding region, which in the short transcript becomes a partial coding sequence (nucleotides 1893-1987) with a different termination codon (UGA). Another deletion occurs in the 3Ј-untranslated region of l-TGN, from nucleotide 2336 to the end of l-TGN, resulting in a relatively short untranslated region in s-TGN.
To verify that the two clones are related to the two transcripts that we observed by Northern blotting using a probe directed to the active site region (Fig. 2), we carried out additional Northern blot analyses using a hybridization probe corresponding to the 3Ј-region exclusive to l-TGN cDNA (bp 1902-2071) and a probe common to both l-TGN and s-TGN (corresponding to bp 1780 -2090 in l-TGN). The results showed that the TNF-␣or IL-1␤-treated astrocytes which had exhibited two transcripts with the active site-related probe showed one transcript with the l-TGN-exclusive probe (Fig. 4A, a) and two with the common probe (Fig. 4A, b). Interestingly, astrocytes treated with cytokines other than IL-1␤ and TNF-␣ did not show induction of either l-TGN or s-TGN (Fig. 4A, a and b). Fig. 4B shows one band obtained by Southern analysis as a result of DNA fragmentation with the restriction enzymes ClaI, EcoRI, KpnI and SpeI, as well as by high stringency hybridization, suggesting that the two TGase clones are derived from a single gene.
Detection of the short form of tissue-type TGase in astrocytes raised the possibility of its expression in other tissues as well. This was investigated by subjecting RNA preparations of rat liver and brain to PCR analysis using primers specific for tissue-type TGase. These primers, CTCATCAAGGTGCGGG-GTCT and GAGGCTCTAGAGACAGTTCA, should lead to synthesis of the long and/or the short TGase forms, provided that the corresponding messages are expressed by the tissues. As a control for each form, we used mRNA derived from HEK-293 cells transfected with l-TGN or s-TGN. Fig. 4C shows that both brain and liver express one PCR product of the same size as the l-TGN fragment, but neither possesses a PCR product corresponding to s-TGN. These results further argue against the constitutive expression of s-TGN in tissues containing tissuetype TGase. One cannot, however, rule out its possible expression in these tissues, as in astrocytes, given the appropriate conditions of induction.
The predicted amino acid sequences indicated that the long cDNA encodes for a l-TGN of 686 amino acids with a molecular mass of 77 kDa, while the short cDNA encodes for a s-TGN of 652 amino acids with a molecular mass of 73 kDa. The predicted isoelectric points of the putative amino acid sequences of l-TGN and s-TGN are 4.79 and 4.64, respectively.
We raised antibodies to a peptide corresponding to the Cterminal region of l-TGN and to a peptide corresponding to the C-terminal region of s-TGN where no homology exists between them, and used them to analyze the proteins expressed in the cytokine-treated astrocytes. In untreated astrocytes, only l-TGN immunoreactivity was found. In IL-1␤-treated astrocytes, , and a GAPDH probe for quantitative analysis (c). Lanes 1, 2, and 3 correspond to astrocytes treated with tumor growth factor-␤1, basic fibroblast growth factor, and TNF-␣, respectively. Lane 4 relates to untreated astrocytes. B, Southern blot analysis of rat genomic DNA. Rat genomic DNA was purified from spleen cells and subjected to restriction enzyme fragmentation followed by electrophoresis on a 0.8% agarose gel, as described under "Experimental Procedures." A partial fragment of l-TGN cDNA (bp 1780 -2230) was used as a probe for Southern blot hybridization. HindIII-cleaved cDNA was used as a marker. C, PCR analysis of mRNA derived from rat brain and liver, using primers encoding for the C terminus of tissue-type TGase. mRNA preparations of rat liver, brain, and HEK-293 cells transfected with l-TGN or s-TGN were subjected to PCR analysis. The primers were chosen in such a way as to lead to synthesis of distinct C-terminal fragments corresponding to either s-TGN or l-TGN. Lanes 1, 2, 3, and 4 correspond to liver, brain, l-TGN-, and s-TGN-transfected cells, respectively. Note the absence of s-TGN fragments from rat brain and liver.

FIG. 5. Western blot analysis of TGase expression in IL-1␤induced astrocytes.
Astrocytes were cultured and treated with IL-1␤ or TNF-␣, as described in Fig. 1. The cultures were then harvested, and cytosolic fractions were prepared and subjected to Western blot analysis using antibodies recognizing s-TGN or l-TGN exclusively. The arrow points to s-TGN-immunoreactive protein.
both l-TGN-and s-TGN-immunoreactive proteins were found, but the levels of s-TGN-immunoreactive protein were much lower (Fig. 5). In a similar experiment, s-TGN-immunoreactive protein was below detection level in astrocytes treated with TNF-␣. The additional immunoreactive bands, smaller than both the l-TGN and the s-TGN bands, seen in these Western blots might be a result of degradation, since they appeared only in the cytokine-treated astrocytes. Since the antibodies were raised against C-terminal regions, the degradation seems to be N-terminal; this might explain why the size difference between the degradation product and the nondegraded form is the same for s-TGN and l-TGN. The degraded form in the case of s-TGNimmunoreactive protein seemed to be of higher proportion relative to the nondegraded form than in the case of l-TGN. This is consistent with the substantially lower levels of s-TGN obtained in the cytokine-treated astrocytes or in transfected/infected cells. It should be noted that the same pattern was also observed in the HEK-293 l-TGN-and s-TGN-transfected cells; here, too, no immunoreactivity was observed in nontransfected cells (Fig. 8).
Alignment analysis of the amino acid sequences revealed that rat astrocyte TGase shows marked homology with mouse macrophage tissue-type TGase but less homology with human endothelial tissue-type TGase. The overall homology among tissue-type TGases in different species is high, whereas the homology between tissue-type and human factor XIII or human keratinocyte TGases is mainly in the active site region and is lower toward the C-terminal region (Fig. 6).
Kinetic Parameters of l-TGN and s-TGN-The predicted amino acid sequence of the unique coding region in s-TGN cDNA consists of a novel sequence with a high frequency of glutamine residues and no significant secondary structure. The s-TGN form appears to be expressed in IL-1␤or TNF-␣-treated astrocytes, although at considerably lower levels than l-TGN.
These findings suggested that s-TGN and l-TGN might be differently regulated. We were therefore interested in finding out whether s-TGN encodes for an active protein and, if so, in comparing its activity with that of the protein encoded by l-TGN cDNA. To this end HEK-293 cells were transfected with l-TGN cDNA or s-TGN cDNA or both, driven by a cytomegalovirus promoter for high transcription efficiency.
Cytosolic fractions of HEK-293 cells were prepared 48 h after transfection, as described under "Experimental Procedures," and their TGase activities were measured using N,N-dimethylated casein as a substrate and an excess of putrescine. Saturation curves of l-TGN and s-TGN showed that the V max was significantly lower in cells transfected with s-TGN cDNA (64 pmol/h/mg of protein) than in those transfected with l-TGN cDNA (2000 pmol/h/mg of protein) (Fig. 7, a and b). The K m values of l-TGN and s-TGN were derived from reciprocal plots, which were calculated from the linear part of the curve in Fig.  7, a (0 -0.02 mM casein) and b (0 -0.02 mM casein), and found to be similar (4 and 3.8 M, respectively) (Fig. 7, c and d). The lower V max could result either from a lower binding affinity of the s-TGN or from lower enzyme expression in the transfected cells (despite the introduction of the same amounts of cDNA). To distinguish between these two possibilities, Western blot analysis was performed using anti-l-TGN and anti-s-TGN antibodies that were selectively raised against each of the expressed proteins. As shown in Fig. 8A, the two types of antibodies raised against s-TGN and l-TGN did not cross-react immunologically. Both s-TGN and l-TGN were found in the cytosolic and particulate fractions of the transfected cells. However, lower levels of s-TGN than of l-TGN were expressed, both when the cells were transfected with each of the cDNAs separately and when they were cotransfected. These results support the suggestion that the difference in V max values between s-TGN and l-TGN, observed when equal amounts of cytosolic fractions from the transfected cells were examined, was due to differences in levels of the enzymes in the preparations and not in their affinities. Northern blot analysis, in contrast to Western blot, indicated that similar yields of mRNA were obtained when the same amounts of l-TGN and s-TGN cDNAs were used under identical transfection conditions (Fig. 8B). The results of immunoreactivity assays were confirmed by assays of enzymatic activities of the cytosolic fractions derived from l-TGNand s-TGN-transfected cells; activity in the l-TGN-transfected cells was higher than in the s-TGN-transfected cells (Fig. 8C). This quantitative relationship between s-TGN and l-TGN at the transcript and the protein levels was observed in several transfections using different preparations of l-TGN and s-TGN cDNAs, and in cotransfections. These results suggest that the two enzymes have similar TGase activities but different regulatory mechanisms for their expression, and that they probably play distinct roles.
Role of GTP in TGase Activity of l-TGN and s-TGN-Tissuetype TGases were recently found to act as G-proteins in signal transduction. The short form (s-TGN), found in the present study to be expressed in astrocytes, lacks the C-terminal region which includes the site previously suggested to be involved in phospholipase C activation in response to interaction with GTP (1, 2). It was therefore important to determine whether its activity as a TGase is affected by GTP. This was done using the baculovirus expression system. The two enzymes were purified , and cells cotransfected with both cDNAs using 10 g of cDNA for each (sϩl). Equal amounts of total protein or mRNA were subjected to Western (A) or Northern (B) blotting, respectively. In A the blots were exposed either to antibodies directed against s-TGN (␣-s-TGN) or to antibodies directed against l-TGN (␣-l-TGN). In B mRNA were prepared from the same transfected cells used in A and hybridized with the 300-bp PCR product. GAPDH (G) hybridization was used as control for quantitative analysis. C, enzymatic activity of [ 3 H]putrescine incorporation into casein in the presence of 1 mM of CaCl 2 , using 25 g of total protein of each cytosolic fraction, as described under "Experimental Procedures." with a His tag at the N terminus (see "Experimental Procedures"). Their TGase activity was assayed in the presence of 0.5 mM GTP and increasing concentrations of Ca 2ϩ up to 5 mM, using similar amounts of the two enzymes (0.5 g of l-TGN and 0.4 g of s-TGN). As shown in Fig. 9, GTP exerts marked inhibition of l-TGN (upper panel) but had a very limited effect on s-TGN (lower panel), even at low Ca 2ϩ concentrations (Fig.  9). It thus seems that loss of the C-terminal region results in a significantly reduced affinity for GTP binding. The inhibition of l-TGN by GTP appears to be noncompetitive and dominant, especially at physiological concentrations of GTP and Ca 2ϩ . These findings might suggest a regulatory role for GTP in the intracellular TGase activity of l-TGN but not of s-TGN; thus, partial loss of the C-terminal domain in s-TGN may result in a loss of the activity of guanosine triphosphatase (GTPase) but not of TGase. Interestingly, the purified l-TGN was found to be active even without Ca 2ϩ and reached saturation at a very low Ca 2ϩ concentration, while the purified s-TGN was not active in the absence of Ca 2ϩ and reached saturation only at a relatively high Ca 2ϩ concentration. DISCUSSION In this study we showed that the inflammation-associated cytokines, IL-1␤ and TNF-␣, up-regulate TGase expression in rat astrocytes. This up-regulation was shown to be associated with the expression of two transcripts. Cloning and sequencing of the cDNAs encoding for these transcripts revealed that they are members of the tissue-type TGase family and might be a product of the same gene. The shorter form encodes for an enzyme that lacks a C-terminal region, which may be needed for phospholipase C activation when TGase acts as a G␣ h protein (2), and it contains a novel TGase peptide sequence in the C terminus. The cDNA of the long form, when transfected or infected in vitro into HEK-293 cells or baculovirus, respectively, produces a high level of enzyme expression in both the soluble and the membrane-bound form, which is noncompetitively inhibited by GTP. Expression of the cDNA of the short form was significantly lower, and the expressed protein was hardly inhibited by GTP.
The TGases constitute a family of closely related enzymes that probably arose from a common ancestral gene. Sequence differences among the various TGases may be related to differences in their function and/or specificity (28). It has been suggested that the TGase superfamily arose by repeated incorporation of new amino acid sequences at the N terminus after gene duplication (29). Phillips et al. (29) proposed that evolution of the TGase genes included recruitment of new exons to modify the enzymatic action. The main function of the TGases appears to be catalysis of the Ca 2ϩ -dependent covalent crosslinking of peptide-bound glutamine residues of proteins to primary amines, either lysine residues of other proteins or polyamine substrates.
The TGase found in rat brain astrocytes was shown here to be highly homologous with tissue-type TGase. The latter enzymes play a number of roles in many different tissues, including axonal growth (15), differentiation (16), and cell death (30 -32). They may participate in the Ca 2ϩ -dependent crosslinking of cell membranes and cytoskeletal proteins (28). They also act as GTP-binding proteins (1,28).
We recently showed that TGases may play a key role in the central nervous system response to injury and regeneration. Expression of TGase was found to be elevated in injured fish optic nerve, a regenerating system, and the enzyme is presumably localized in the extracellular milieu (24). Furthermore, when this TGase was purified from medium conditioned by fish optic nerve and applied locally to injured rat optic nerve, a significant improvement was observed in the visual evoked potential response to light (25). These results suggested the possibility that expression or activation of TGase in injured mammalian central nervous system is inhibited. Recently, we demonstrated that astrocytes exposed ex vivo to IL-1␤ and TNF-␣ facilitate recovery upon their subsequent transplantation into transected adult rat optic nerve. We suggested that this effect might be due to increased or altered expression of regeneration-associated components, including TGases, in the treated astrocytes. In the present study we show that TGase is expressed by rat astrocytes; moreover, we show that the inflammation-associated cytokines IL-1␤ and, to a lesser extent, TNF-␣ cause up-regulation of TGase expression and also induce expression of an additional form of TGase not detectable in untreated cells. IL-1␤ and TNF-␣ have also been shown to cause elevated expression of other regeneration-supportive el- ements (33), such as basic fibroblast growth factor (34), nerve growth factor (35)(36)(37), and cell adhesion molecules. These cytokines may be deficient in the central nervous system due to inappropriate signaling for macrophage recruitment and activation (38 -40).
In this study we detected two TGase transcripts in cytokinetreated astrocytes, one of 3.7 kb (l-TGN) and the other of 2.4 kb (s-TGN). Cloning and sequencing of the cDNAs revealed that l-TGN cDNA encodes for a long polypeptide of 686 amino acid residues with a molecular mass of 77 kDa, whereas s-TGN cDNA encodes for a slightly shorter polypeptide of 652 amino acid residues and a molecular mass of 73 kDa. Both polypeptides show marked homology with tissue-type TGase. The two polypeptides are identical except for their C termini, where a deletion in l-TGN results in a shift in the coding region and replacement of the termination codon in s-TGN. In the shorter enzyme, the region that encodes for amino acids 623-653 corresponds to the noncoding region in l-TGN cDNA. Alternative splicing has been suggested previously as a mechanism for creating variant forms in the TGase gene family (28,41,42). Southern blot analysis, in which only one band was detected after digestion with several restriction enzymes, may point to alternative splicing as the mechanism for creating these two mRNAs. A CTTCAC sequence, which is in consensus with the preexon sequence CTTCAGG (43) and appears in the flanking region of the deleted part (see Fig. 3), might be related to such an alternative splicing mechanism. The final confirmation of this issue, however, must await full genomic sequencing of the tissue-type TGase and analysis of consensus and new bordering sequences of alternative splicing (43,44).
Expression of s-TGN in this study occurred during overexpression of TGase in astrocytes and has not been described in any other tissue. It is possible that other tissues can also be induced to express the two forms of the enzymes under conditions yet to be discovered. If this is so, the induction conditions should be specified. It was shown, for example, that mouse macrophage tissue-type TGase is overexpressed when induced with retinoic acid, but although it contains the signal sequences (45,46), no other tissue-type TGase mRNA was observed.
The C terminus of s-TGN lacks a sequence that was previously proposed to be involved in the activity of TGase when it acts as a GTP-binding G␣ h protein (2). Instead, it contains a unique C-terminal sequence with a high frequency of glutamine residues. We suggest that the change in the C-terminal sequence of s-TGN might also result in structural and functional differences that would affect the GTP binding. Binding of GTP was assayed in terms of the rate of inhibition of TGase activity, since such activity is blocked by binding of GTP. The results confirmed that GTP binds to l-TGN with substantially higher affinity than to s-TGN, which means that the modification in s-TGN includes functional elements that affect GTP binding.
Tissue-type TGases are probably active both as TGases and as GTPases, depending on the extracellular signals and the intracellular GTP/Ca 2ϩ ratio. It seems likely, however, that GTP serves as an intracellular regulator of TGase activity, since it inhibits noncompetitively the enzyme activity even in the presence of Ca 2ϩ . Excessive elevation of Ca 2ϩ concentration, as in apoptosis, might promote cross-linking activity, and could account for the down-regulation of s-TGN levels, which possess GTP-independent TGase activity. The s-TGN enzyme showed a higher Ca 2ϩ dependence, which might serve as a safety mechanism for its GTP-independent cross-linking activity.