Identification of a novel GTPase, the inducibly expressed GTPase, that accumulates in response to interferon gamma.

Interferon γ is a pleiotropic cytokine that regulates many immune functions. We have identified a novel protein, inducibly expressed GTPase (IGTP), whose expression was regulated by interferon γ in macrophages. In mouse RAW 264.7 macrophages, IGTP mRNA levels were almost undetectable but increased within 1 h of exposure to interferon γ, peaked at very high levels within 3 h, and remained at high levels to at least 48 h; pretreatment of the cells with cycloheximide blocked the majority of mRNA accumulation. In the mouse, the mRNA was highly expressed in thymus, spleen, lung, and small intestine. Using interspecific backcross analysis, the Igtp gene was mapped to mouse chromosome 11. The IGTP cDNA encoded a putative polypeptide of Mr 48,507 and pI 7.79 that contained three consensus GTP binding motifs, GXXXXGK(S/T), DXXG, and NTKXD. Both IGTP that had been immunoprecipitated from RAW cells and a glutathione S-transferase IGTP fusion protein were able to convert GTP to GDP in vitro. Subcellular protein fractionation and Western blotting localized IGTP to the cytosol of RAW cells. In addition, the protein was homologous to proteins encoded by three previously cloned cDNAs, IRG-47, TGTP/Mg21, and LRG-47, and thus may be representative of a new family of interferon γ-regulated GTPases.

Interferon ␥ is a pleiotropic cytokine that regulates many immune functions. We have identified a novel protein, inducibly expressed GTPase (IGTP), whose expression was regulated by interferon ␥ in macrophages. In mouse RAW 264.7 macrophages, IGTP mRNA levels were almost undetectable but increased within 1 h of exposure to interferon ␥, peaked at very high levels within 3 h, and remained at high levels to at least 48 h; pretreatment of the cells with cycloheximide blocked the majority of mRNA accumulation. In the mouse, the mRNA was highly expressed in thymus, spleen, lung, and small intestine. Using interspecific backcross analysis, the Igtp gene was mapped to mouse chromosome 11. The IGTP cDNA encoded a putative polypeptide of M r 48,507 and pI 7.79 that contained three consensus GTP binding motifs, GXXXXGK(S/T), DXXG, and NTKXD. Both IGTP that had been immunoprecipitated from RAW cells and a glutathione S-transferase IGTP fusion protein were able to convert GTP to GDP in vitro. Subcellular protein fractionation and Western blotting localized IGTP to the cytosol of RAW cells. In addition, the protein was homologous to proteins encoded by three previously cloned cDNAs, IRG-47, TGTP/Mg21, and LRG-47, and thus may be representative of a new family of interferon ␥-regulated GTPases.
Interferon ␥ (IFN␥) 1 is a 20-kDa protein that regulates a wide variety of immunological and inflammatory processes (for review see Ref. 1). While production of IFN␥ is limited to CD8 ϩ T cells, some CD4 ϩ T cell subsets (2), and natural killer cells (3), its receptor is found on almost all cell types where it elicits many diverse physiological responses. In macrophages for example, IFN␥ induces major histocompatibility class II expression (4), increases F c receptor-mediated phagocytosis (5), and mediates removal of neoplastic cells and virally and parasitically infected cells by initiating secretion of cytosidal compounds and tumor necrosis factor ␣ (6 -9). In endothelial cells on the other hand, IFN␥, in combination with tumor necrosis factor ␣, markedly increases expression of major histocompatibility class I molecules and the cell adhesion molecules, ICAM-1 and ELAM-1, thereby promoting recruitment of immune cells to areas of inflammation (10). These pleiotropic responses are initiated by binding of IFN␥ to its receptor, followed by receptor dimerization and a cascade of primary and secondary events (1). However, these molecular events have not been completely defined, and identification of proteins whose expression is regulated by IFN␥ is the subject of active investigation.
We report here the identification of a 48-kDa protein, designated IGTP, whose expression was rapidly and dramatically increased by IFN␥ in macrophages and fibroblasts. The protein was first recognized as the product of a differentially displayed cDNA in hepatocyte growth factor (HGF)-treated C127 mouse fibroblasts. In these cells, the IGTP mRNA levels showed a slight transient increase following HGF exposure; in addition, mRNA levels were constitutively elevated in C127 and NIH/ 3T3 cells that were transformed by coexpression of HGF and its receptor Met. These observations and their significance will be discussed elsewhere. In the present work, we characterize the IFN␥-induced IGTP gene expression, and we demonstrate that the protein product is a GTPase. Because of its expression pattern and its biochemical activity, the protein has been designated inducibly-expressed GTPase (IGTP). The protein had high sequence homology with the proteins encoded by three other cDNAs, IRG-47 (11), TGTP(12)/Mg21 (13), and LRG-47 (14), for which no biochemical or physiological function had been described previously, but which also contained putative GTP binding sequences. IGTP, therefore, may be representative of a new family of GTPases that could potentially mediate the effects of IFN␥ in macrophages and other cells. Northern Blotting-Total RNA was prepared from cultured cells (16) and from mouse tissues (17) using standard acidic phenol/chloroform extraction protocols. 15 g of RNA samples were separated on 1.2% agarose/formaldehyde gels and used for Northern blot analysis as de-scribed previously (18). The blots were probed with a 32 P-labeled human glyceraldehyde-phosphate dehydrogenase probe isolated as a 1.2-kb PstI fragment of pHcGAP (19), a mouse IP10 probe isolated as a 0.5-kb EcoRI fragment of the mouse C7-1 (ATCC 63135), a human ␤-actin cDNA probe (Clontech; supplied with product 7760 -1), or an IGTP probe isolated as a 0.28-kb EcoRI fragment of the IGTP cDNA (bases 1645-1927, GenBank accession U53219). The IGTP probe contained entirely 3Ј-untranslated sequences that were not conserved with the related cDNAs referred to in the text; thus the probe was specific for IGTP.
Differential Display Screening and Library Screening-C127/Met cells were exposed to 200 units/ml hepatocyte growth factor for various times, and total RNA was isolated and used for differential display screening with an RNAimage kit 1 (GenHunter Corp., Brookline, MA) according to the protocols supplied by the manufacturer. One 150-base cDNA fragment, designated A2c, was isolated using primers H-T 11 A and H-AP2; levels of A2c were slightly increased in cDNA generated from C127/Met cells exposed to HGF for 3-8 h. The cDNA fragment was used to screen a mouse spleen cDNA library (Stratagene, La Jolla, CA), from which a 1927-base cDNA clone was isolated and later designated as the IGTP cDNA. The clone was sequenced on both strands using the Applied Biosystems (Foster City, CA) Prism DyeDeoxy Sequencing System. One end of this cDNA corresponded to the 150-base A2c clone. Based on a partial poly(A) tail present in A2c, the IGTP cDNA was oriented, and a 1419-base open reading frame was identified. The DNA sequence and translated protein were analyzed using Genetics Computer Group (Madison, WI) software. 3 Plasmid Subcloning and Preparation of GST Fusion Proteins-The plasmid encoding GST-IGTP, pGEX/IGTP, was created by ligating a 1.6-kb NcoI-XhoI fragment of the IGTP CDNA into the NcoI-XhoI site of pGEX-KG (20); this produced an in-frame fusion of the GST gene and the IGTP cDNA. pGEX-KG and pGEX/IGTP were transformed into the Escherichia coli strain BL21/DE3, and the GST fusion proteins were purified using glutathione-Sepharose (Pharmacia) as described previously (21).
Interspecific Backcross Mapping-Interspecific backcross progeny were generated by mating (C57BL/6J ϫ Mus spretus) F 1 females and C57BL/6J males as described previously (22). A total of 205 N 2 mice were used to map the Igtp locus (see the text for details). DNA isolation, restriction enzyme digestion, agarose gel electrophoresis, and Southern blotting were performed as described previously (23). All blots were prepared with Hybond-N ϩ nylon membranes (Amersham Corp.). The probe, a 0.28-kb EcoRI fragment of the 3Ј-untranslated region of the IGTP cDNA, was labeled with an ␣-[ 32 P]dCTP random primer labeling system (number 300385, Stratagene, La Jolla, CA); washing was done to a final stringency of 0.25 ϫ SSCP, 0.1% (w/v) SDS, 65°C. A fragment of 2.5-kb was detected in SacI-digested C57BL/6J DNA, and a 3.0-or 4.5-kb fragment was detected in SacI-digested M. spretus DNA. The presence or absence of the 3.0-or 4.5-kb SacI M. spretus-specific fragment was followed in backcross mice.
A description of the probes and restriction fragment length polymorphisms (RFLPs) for Il3 and Trp53, two of the loci linked to Igtp, has been published (24). Recombination distances were calculated using the computer program SPRETUS MADNESS (25). Gene order was determined by minimizing the number of recombination events required to explain the allele distribution patterns.
Immunoprecipitation-Cells were plated on 60-mm plates and grown to confluence. In some experiments, the cells were labeled by washing them three times with Dulbecco's modified Eagle's medium without methionine (Life Technologies, Inc.) and then incubating them in the same medium supplemented with 10% fetal calf serum (v/v) and 0.5 mCi of [ 35 S]methionine (DuPont NEN) for 4 h. Cells were exposed to 100 units/ml mouse interferon ␥ (Boehringer Mannheim) for various times during the labeling period.
Total cellular protein lysates were prepared by washing the cells three times with ice-cold phosphate-buffered saline and then scraping them into 0.5-ml ice-cold lysis buffer (1%(v/v) Triton X-100, 0.1%(w/v) sodium dodecyl sulfate, 50 mM Tris, pH 7.5, 0.15 M NaCl, 5 mM EDTA) with plastic cell scrapers. Following sonication for 15 s in a cup-horn sonicator (Sonifier 450, Branson, Danbury, CT), the lysates were cleared by centrifugation in microcentrifuge tubes in a refrigerated microcentrifuge (MTX-150, Tomy Seiko, Tokyo) for 10 min at 15,000 ϫ g. The samples were matched for trichloroacetic acid-precipitated radioactive counts and then incubated with normal rabbit serum on a rotator at 4°C for 1 h, and with 6 mg of protein A-Sepharose (Pharmacia, Uppsala, Sweden) for an additional hour at 4°C. Following a 2-min centrifugation at 500 ϫ g to pellet the protein A-Sepharose, the supernatant was transfered to a new microcentrifuge tube and incubated for 3-4 h with anti-IGTP polyclonal rabbit serum that had been generated against the peptide CSLRKALKDSVLPPEIH-OH corresponding to the 16 carboxyl-terminal amino acids of IGTP. Next, 6 mg of protein A-Sepharose were added, and the slurry was incubated for an additional 1 h at 4°C. Finally, the protein A-Sepharose was washed four times with ice-cold lysis buffer and 0.1 ml of SDS buffer (24% (w/v) sucrose, 0.36 M dithiothreitol, 0.048 M EDTA, 4.8% (w/v) sodium dodecyl sulfate, 0.05% (w/v) bromphenol blue) that had been diluted 1:3 with PBS was added; the suspension was mixed vigorously and boiled for 5 min, and the resin was pelleted by centrifugation for 5 min at 15,000 ϫ g. The proteins in the SDS buffer were separated by 9% SDS-polyacrylamide gel electrophoresis.
GTP Hydrolysis Assays-Protein was immunoprecipitated from cell lysates as described above, except that the cells were lysed in a lysis buffer of 10 mM CHAPSO, 50 mM Tris, pH 8.0, and 0.15 M NaCl; immunoprecipitation using this buffer results in slightly higher GTPase activities for some GTP binding proteins (26). ␣-[ 32 P]GTP hydrolysis and separation of the nucleotide products by polyethyleneimine (PEI) chromatography were carried out as described previously (26).
Subcellular Fractionation and Western Blotting-Cells on 60-mm plates were washed three times with ice-cold PBS, gently scraped into 0.5 ml of PBS, and pelleted by centrifugation at 500 ϫ g for 5 min at 4°C. Next, cellular protein was fractionated by either of two methods. In one method, cells were gently resuspended in a hypotonic buffer (10 mM Tris, pH 7.5, 10 mM NaCl, 1.5 mM MgCl 2 , 0.06 g/ml aprotinin, 0.5 g/ml leupeptin, and 0.4 mM Pefabloc (Boehringer Mannheim)) and recentrifuged at 500 ϫ g for 5 min at 4°C. Then the cells were resuspended in 0.5 ml of hypotonic buffer and incubated on ice for 5 min. The cells were lysed by passing through a 25-gauge needle attached to a syringe 15-20 times. The broken cell suspension was centrifuged at 3000 ϫ g for 5 min at 4°C, producing a nuclear/particulate pellet and a cytosolic supernatant. The pellet was resuspended in 1 ml of hypotonic buffer, recentrifuged at 500 ϫ g for 5 min, and finally resuspended in 0.5 ml of hypotonic buffer; 0.25 volume of an SDS buffer (24% (w/v) sucrose, 0.36 M dithiothreitol, 4.8% (w/v) sodium dodecyl sulfate, 0.048 M EDTA, and 0.05% (w/v) bromphenol blue) was added to the suspension. The cytosolic fraction was centrifuged at 15,000 ϫ g to remove residual nuclei and particulates; an aliquot was removed for protein quantification using the Bio-Rad assay (Hercules, CA), and 0.25 volume of SDS buffer was added to the remainder.
Alternatively, cellular protein was fractionated by gently resuspending the pelleted cells in a Nonidet P-40 buffer (0.5% (v/v) Nonidet P-40, 0.06 g/ml aprotinin, 0.5 g/ml leupeptin, and 0.4 mM Pefabloc in PBS) and incubating them on ice for 5 min. The nuclear/particulate and cytosolic protein fractions were then separated by centrifugation as in the first fractionation method. With both methods, lysis of the cells and integrity of the nuclei were monitored by light microscopy.
25 g of cytosolic protein and the equivalent volume of the nuclear/ particulate protein fraction were separated on a 9% SDS-polyacrylamide gel; the gel was transfered to an Immobilon-P membrane (Millipore, Bedford, MA) with a TEX50 wet transfer apparatus (Hoefer, San Francisco, CA) using 25 mM Tris, 480 mM glycine, 0.1%(w/v) SDS, and 50%(v/v) methanol buffer. Western blotting and detection were carried out using the ECL Detection System (Amersham Int., Buckinghamshire, UK) according to the manufacturer's protocols. The rabbit polyclonal anti-IGTP antiserum and the mouse monoclonal anti-␤-tubulin antibody (Sigma) were used at a 1:1000 dilutions; the peroxidaseconjugated goat anti-mouse IgG and anti-rabbit IgG secondary antibodies (Boehringer Mannheim) were used at 1:10,000 dilutions.

RESULTS
The IGTP cDNA was first identified as a 150-base cDNA fragment that was differentially displayed in hepatocyte growth factor-treated C127 mouse fibroblasts (data not shown). However, it was observed during these experiments that IGTP mRNA levels were dramatically increased by 100 units/ml IFN␥, in both this cell line (data not shown) and in RAW 264.7 mouse macrophages (Fig. 1). While basal IGTP mRNA levels were almost undetectable, they were easily detectable within 1 h of IFN␥ exposure, peaking at very high levels at 3 h, and remaining at high levels to at least 48 h (Fig. 1). Longer exposures of the blot showed that trace amounts of IGTP mRNA were present before exposure to IFN␥ (data not shown). In RAW cells, IGTP mRNA levels were also induced by 1 g/ml lipopolysaccharide with similar kinetics but to a lesser extent (Fig. 1). With increased lipopolysaccharide exposure times, the mRNA showed a slightly increased mobility which could have resulted from shortening of the poly(A) tail; however, this possibility was not addressed experimentally.
The half-life of the IGTP mRNA was determined by exposing RAW cells to 100 units/ml IFN␥ for 3 h and then to the transcriptional inhibitor actinomycin D (4 M) for various times; the decay of the accumulated mRNA was then followed by Northern blotting. Under these conditions, the mRNA decayed with a half-life of about 4.5 h (data not shown).
Inducibly expressed genes are often classified as primary response genes if induction requires only previously translated transcription factors or as secondary response genes if newly translated factors are required (for review see Ref. 27). To determine if IGTP accumulation required protein synthesis, RAW cells were exposed to 0.1 mM cycloheximide for 30 min and then to 100 units/ml IFN␥; mRNA accumulation was followed by Northern blotting (Fig. 2). Inhibition of protein synthesis in this manner blocked the majority of the IGTP mRNA accumulation, although a small amount of mRNA did accumulate (Fig. 2). Conversely, under the same conditions, cycloheximide did not affect the accumulation of IP10 mRNA, which is a primary IFN␥ response gene in macrophages (28) (Fig. 2). Therefore, because IGTP mRNA levels did not show detectable increases until 1 h following IFN␥ exposure, and because full induction required protein synthesis, IGTP accumulation may be classified as a secondary IFN␥ response.
Hybridization of the IGTP cDNA to a total RNA blot containing RNA from several mouse tissues revealed very high expression in the thymus and slightly lower expression in the spleen, lung, and small intestine; expression in brain, heart, kidney, liver, skeletal muscle, and testes was very low or undetectable (Fig. 3). The tissue expression pattern suggested that IGTP may be highly expressed in immune cell populations.
This interspecific backcross mapping panel has been typed for over 2100 loci that are well distributed among the autosomes, as well as the X chromosome (22). C57BL/6J and M. spretus DNAs were digested with several enzymes and analyzed by Southern blot hybridization for informative restriction map polymorphisms (RFLPs) using a fragment from the 3Ј-untranslated region of the IGTP cDNA as probe. The 3.0-or 4.5-kb SacI M. spretus RFLP (see "Experimental Procedures") was used to follow the segregation of the Igtp locus in backcross mice. The mapping results indicated that Igtp was located in the central region of mouse chromosome 11 linked to Il3 and Trp53. One hundred and twenty-two mice were analyzed for every marker FIG. 1. Effects of interferon ␥ and lipopolysaccharide on IGTP mRNA accumulation in RAW 264.7 cells. Cells were exposed to 100 units/ml interferon ␥ or 1 g/ml lipopolysaccharide for the indicated times. The cells were then used for preparation of total cellular RNA and Northern blotting with an IGTP probe or a glyceraldehyde-3-phosphate dehydrogenase probe. The positions of the major ribosomal RNA species, as determined by the stained gel, are indicated. Other details are described in the text.

FIG. 2.
Effect of cycloheximide on interferon ␥-induced accumulation of IGTP mRNA in RAW 264.7 cells. Cells were exposed to 0.1 mM cycloheximide for the indicated times, to 100 units/ml interferon ␥ for the indicated times, or to 0.1 mM cycloheximide for 30 min and then to 100 units/ml interferon ␥ for the indicated times. The cells were then used for preparation of total cellular RNA and Northern blotting with an IGTP probe, an IP-10 probe, or a glyceraldehyde-3-phosphate dehydrogenase probe. The positions of the major ribosomal mRNA species, as determined by the stained gel, are indicated.

FIG. 3. Expression of IGTP mRNA in various mouse tissues.
Total cellular RNA was prepared from a 3-month-old male C57BL/6 mouse. 15-g samples were subjected to Northern blotting with an IGTP cDNA probe or a ␤-actin cDNA probe. The positions of the major ribosomal mRNA species, as determined by the stained gel, are indicated. and are shown in the segregation analysis (Fig. 4). Each locus was analyzed in pairwise combinations for recombination frequencies using the additional data. The ratios of the total number of mice exhibiting recombinant chromosomes to the total number of mice analyzed for each pair of loci and the most likely gene order were as follows: centromere-II3-4/122-Igtp-9/122-Trp53. The recombination frequencies (expressed as genetic distances in centimorgans (cM) Ϯ S.E.) were II3-3.3 Ϯ 6.6-Igtp-7.4 Ϯ 2.4-Trp53.
Based on the high IGTP mRNA levels in spleen, a spleen cDNA library was chosen for screening with the original 150base IGTP cDNA fragment. Several positive clones were obtained, one of which was 1927 bases and contained a 1419-base open reading frame between bases 204 and 1619 (GenBank accession U53219); the clone contained no other open reading frames longer than 200 bases. The first potential translation initiation codon within the long open reading frame occurred at base 348, and translation from this codon produced a 424amino acid polypeptide with a calculated M r of 48,507 and a pI of 7.79. The translated sequence showed substantial homology to the translated sequences of three previously cloned cDNAs: IRG-47 (11), TGTP (12)/Mg21 (13), and LRG-47 (14) (Fig. 5). These cDNAs were cloned as cDNAs whose expression was increased by IFN␥ in 70Z/3 pre-B cells (11), by T cell receptor cross-linking in CD4 ϩ 8 ϩ thymocytes (12) and IFN␥ in peritoneal macrophages (13), or by IFN␥ in RAW 264.7 macrophages (14), respectively. In IFN␥-treated RAW cells, levels of the three mRNAs increased markedly with kinetics similar to the increases in IGTP mRNA levels (data not shown). In addition, each mRNA was highly expressed in mouse thymus, spleen, lung, and small intestine (data not shown). Thus, expression of IGTP and these three related cDNAs was nearly identical in several cell types and tissues. Comparison of the amino acid sequences showed that in at least three of the four proteins 36% of the amino acids were absolutely conserved or were replaced with conservative substitutions (Fig. 5). Sequence homology was found throughout the proteins and was particularly high in the regions of three conserved GTP binding motifs, GXXXXGKS/T, DXXG, and N/TKXD (29); these motifs are found in many functionally diverse GTP-binding proteins including Ras (29) and dynamin (30). In the case of the proteins encoded by IRG-47, TGTP/Mg21, or LRG-47, neither GTP binding nor had any other biochemical or physiological function been determined.
To facilitate the study of IGTP, rabbit polyclonal antisera were generated that recognized the carboxyl-terminal 16 amino acids of IGTP (Fig. 5). This sequence was not shared with any of the three related proteins, and the antibody that was generated immunoprecipitated in vitro translated IGTP but not in vitro translated LRG-47 protein (data not shown). From the [ 35 S]methionine-labeled protein lysates of RAW 264.7 cells, the antibody immunoprecipitated a dominant band of about 48 kDa (Fig. 6) that probably represented IGTP. This band had the same relative migration as that of in vitro translated IGTP (data not shown) and was not precipitated by preimmune rabbit serum (Fig. 6). In addition, the band was weak in cells that had not been exposed to IFN␥, and its intensity increased markedly following IFN␥ stimulation, reaching peak levels in 3-4 h (Fig. 6). Two other bands of about 55-60 kDa were also immunoprecipitated by the IGTP antibody but not by preimmune serum. These bands could have been proteins that associated with IGTP and were coprecipitated; however, because the immunoprecipitation was done in the presence of SDS and because the intensities of the bands did not fluctuate with the intensity of the IGTP band, it seemed more likely that they were unrelated proteins that were also recognized by the antibody. In Western blotting experiments using the same antibody, the 48-kDa IGTP was not detectable in RAW cells that had not been exposed to IFN␥ but was easily detectable in cells that had been exposed to IFN␥ for 4 h or longer ( Fig. 9 and data not shown).
To determine whether IGTP was able to hydrolyze GTP to GDP, as suggested by the presence of putative GTP binding sites, protein that had been immunoprecipitated from RAW cells was incubated with ␣-[ 32 P]GTP for various times, and the radiolabeled nucleotides were separated by thin layer chromatography (Fig. 7). Very little ␣-[ 32 P]GTP was converted to GDP by protein immunoprecipitated by preimmune rabbit serum from either cells incubated under control conditions or cells exposed to IFN␥ for 4 h (Fig. 7). Similarly, very little GTPase activity was immunoprecipitated by the IGTP antibody from cells not exposed to IFN␥; however, greater than four times more GTPase activity was precipitated with the IGTP antibody from cells that had been exposed to IFN␥ for 4 h (Fig. 7). The increase in GTPase activity was probably due to the increased amounts of IGTP that are present in IFN␥-stimulated RAW cells (Fig. 6), which suggested that IGTP was a GTPase. Alternatively, it was also possible that another IFN␥-inducible factor that had GTPase activity, or a factor that was able to catalyze the IGTP GTPase activity, was also present in the immune pellet from IFN␥-stimulated cells. Similar experiments were performed using an antibody that immunoprecipi- tated the LRG-47 protein, and, again, increased GTPase activity was detected in immune pellets from IFN␥-treated cells, relative to control cells (data not shown). Taken together, these data suggest that IGTP, LRG-47, and possibly the other related proteins are GTPases.
In similar experiments, IGTP was expressed in bacteria as a glutathione S-transferase (GST) fusion protein and, following partial purification by glutathione affinity chromatography (data not shown), was used in ␣-[ 32 P]GTP hydrolysis assays (Fig. 8). The GST-IGTP fusion protein converted about six times more GTP to GDP than was converted by an equivalent amount of GST protein (Fig. 8). This result suggested that the IGTP GTPase activity was inherent to the protein and did not require other mammalian accessory proteins for basal activity.
To determine the subcellular distribution of IGTP, RAW cells were exposed to IFN␥ for 4 h to stimulate synthesis of the protein and then cellular protein was separated into cytosolic and nuclear/particulate fractions; the presence of IGTP and the cytosolic protein ␤-tubulin (31) was determined by Western blotting (Fig. 9). In some experiments, protein fractions were prepared by briefly incubating the cells in a hypotonic buffer, disrupting the cytosolic membranes by drawing the cells through a 25-gauge needle, and then separating nuclei and particulates from the cytosolic protein fraction by centrifugation. Using this procedure, IGTP was readily detected in the nuclear/particulate protein fraction of IFN␥-treated cells but not in the cytosolic protein fraction (Fig. 9). Conversely, when the cytosolic membranes were lysed by incubating the cells in an isotonic buffer containing 0.5% (v/v) Nonidet P-40, IGTP localized to the cytosolic protein fraction (Fig. 9). With both procedures, the nuclei appeared intact as determined by light microscopy (data not shown), and ␤-tubulin was detected only in the cytosolic protein fractions (Fig. 9). These data raised the possibility that IGTP may have been a cytosolic protein that was loosely associated with membranes in the nuclear/particulate fractions and that this association was disrupted by mild treatment with a nonionic detergent such as Nonidet P-40. Cytosolic localization of IGTP was also supported by preliminary results using the IGTP antibody to immunostain IFN␥treated RAW cells (data not shown). DISCUSSION We report here the identification of a 48-kDa GTPase, IGTP, whose expression was rapidly and markedly increased by IFN␥ in macrophages and fibroblasts. Its expression pattern suggested that IGTP could potentially mediate some IFN␥-induced responses in macrophages and fibroblasts, and its high expression in mouse thymus and spleen suggested that it may also function in other immune cell populations.
We mapped the Igtp gene to a location on mouse chromosome 11 between Il3 and Trp53. We compared our interspecific map of chromosome 11 with a composite mouse linkage map that reports the location of many uncloned mouse mutations 4 ; Igtp mapped in a region of the composite map that lacks mutations with a phenotype that might be expected for an alteration at this locus.
The central region of mouse chromosome 11 shares homology to regions of human chromosomes 5q and 17p, and thus, the FIG. 6. Time course of IGTP accumulation in RAW 264.7 cells exposed to interferon ␥. Cells were exposed to [ 35 S]methionine for 4 h and to 100 units/ml interferon ␥ for the indicated times during the labeling period. Cell lysates were prepared and used for immunoprecipitation with the IGTP antibody; the precipitated proteins were separated on a 9% SDS-polyacrylamide gel. The positions of selected molecular mass markers are shown at the left. Other details are described in the text.
human IGTP gene is likely to reside on 5q or 17p. Interestingly, monosomy for chromosome 5 or 5q deletions are frequently associated with myelodysplastic disorders, myeloproliferative syndromes, and acute myeloid leukemia (32,33). Because Igtp is expressed in myeloid cells, it will be of interest to determine if the human IGTP gene maps to the region on chromosome 5 thought to contain this important myeloid tumor suppressor gene (32).
The IGTP protein contained three GTP-binding motifs, GXXXXGKS/T, DXXG, and NKXD (29), and the protein had inherent GTPase activity. Although many other GTP-binding proteins have been identified, they have diverse cellular functions, and the presence of this biochemical activity in itself does not suggest a cellular function for IGTP. However, the extensive information gained in the biochemical studies of Ras and other GTP-binding proteins may facilitate study of IGTP. For instance, the x-ray crystal structure of Ras (34,35) has shown that the first and second GTP-binding motifs both contact the phosphates of the GTP molecule and a Mg 2ϩ cofactor, whereas the third motif contacts the purine ring. Mutation of serine 17 in the first GTP-binding motif to asparagine blocks the ability of Ras to hydrolyze GTP, and expression of this Ras mutant in cells blocks the function of wild-type Ras (36). The analogous mutation in dynamin, another GTP-binding protein, also has a dominant-negative effect (37,38), and this may well be true of FIG. 7. GTPase activity of immunoprecipitated IGTP. RAW 264.7 cells were exposed to control conditions or to 100 units/ml interferon ␥ (IFN␥) for 4 h, and the cells were then used for immunoprecipitation with preimmune serum or anti-IGTP serum. ␣-[ 32 P]GTP was incubated with the immunoprecipitated proteins or with buffer at 37°C for the indicated times, and the nucleotide products were separated by polyethyleneimine thin layer chromatography. A, the chromatogram was used for autoradiography, and the positions of GTP and GDP were determined using the fluorescent indicator in the PEI-cellulose plates. B, the amounts of radioactive GTP and GDP were determined with a phosphorimager, and the ratio of GDP to GTP was plotted as a function of time. Other details are described in the text. Ⅺ, buffer; छ, control, preimmune; E, control, anti-IGTP; Ç, IFN, preimmune; µ, IFN, anti-IGTP.

FIG. 8. GTPase activity of a GST-IGTP fusion protein.
Bacterially expressed GST and GST-IGTP fusion proteins were partially purified using glutathione affinity chromatography. ␣-[ 32 P]GTP was incubated with 10 g of the purified proteins or bovine serum albumin (BSA) at 37°C for the indicated times, and the nucleotide products were separated by polyethyleneimine thin layer chromatography. A, the chromatogram was used for autoradiography, and the positions of GTP and GDP were determined using the fluorescent indicator in the PEIcellulose plates. B, the amounts of radioactive GTP and GDP were determined with a phosphorimager, and the ratio of GDP to GTP was plotted as a function of time. Other details are described in the text. Ⅺ, BSA; छ, GST; Q, GST-IGTP. IGTP. Such a dominant negative mutant would be useful in determining the cellular function of IGTP.
The IGTP amino acid sequence was very similar to the translated sequences of three previously cloned cDNAs, IRG-47 (11), TGTP (12)/Mg21 (13), and LRG-47 (14), each of which contained three GTP-binding motifs and had molecular masses of about 47 kDa. Moreover, the increases in IRG-47, TGTP/Mg21, and LRG-47 mRNA levels in IFN␥-treated RAW macrophages, and their expression patterns in mouse tissues, paralleled those of IGTP mRNA. Therefore, it seems likely that IGTP is representative of a new family of proteins whose expression is regulated similarly and whose biochemical functions are similar. The levels at which protein expression and biochemical function differ will be the subject of future studies.
At least two other groups of interferon-induced GTP-binding proteins exist. The first of these is comprised of the Mx proteins, a group of 72-kDa proteins involved in the interferoninduced inhibition of viral proliferation (39). The second group includes several 65-67-kDa proteins that bind to GTP-agarose but whose function is unknown (40,41). However, because little sequence similarity exists between IGTP and either of these groups of proteins, other than that found in the GTPbinding motifs, no functional comparisons can be drawn.
Future experiments will focus on determining the physiological function of IGTP. Among the many possible functions of the protein are mediation of IFN␥ signal transduction at some level and regulation of some of the multiple IFN␥-induced phenotypic responses. It seems unlikely that IGTP is involved in any of the immediate signal transduction events stimulated by IFN␥, given that its expression is not detectable until 1 h following stimulation. In addition, it seems unlikely that IGTP could have any direct effects on gene expression, as it is probably not located in the cell nucleus. However, it is tempting to speculate that because of the large increases in protein processing associated with cytokine secretion and cell surface protein expression in IFN␥-treated cells (1), IGTP might be involved in protein processing. One GTP-binding protein that is involved in protein processing is SRP54, the 54-kDa component of the signal recognition particle (SRP) that targets ribosomes synthesizing proteins with a signal sequence to the endoplasmic reticulum (for review see Ref. 42). One model of SRP54 function (43) suggests that while in the GTP-bound state, the SRP54 component of SRP recognizes a ribosome from which the nascent signal sequence of a presecretory protein has just emerged, then binds this complex, and targets it to the signal recognition receptor complex on the surface of the endoplasmic reticulum. Once transfer to the endoplasmic reticulum is complete, the GTP is somehow hydrolyzed catalyzing transfer of the signal sequence to the translocon and release of the SRP with the SRP54 component in the GDP-bound state (43). Because IGTP shows some similarities to mouse SRP54 (44), it is possible that IGTP could have a similar function. At the amino acid level, the two proteins are about 15% identical and 43% similar; this includes residues in the GTP-binding motifs and others scattered throughout the proteins. In addition, both proteins contain several methionines in the carboxyl-terminal portion of the protein, which in the case of SRP54 may be involved in binding the signal sequence of nascent polypeptides (45). Moreover, we have shown that IGTP might associate with cellular membranes, as SRP54 does. Future studies should address the possible involvement of IGTP in protein processing.