Golgi targeting of human guanylate-binding protein-1 requires nucleotide binding, isoprenylation, and an IFN-γ-inducible cofactor

  1. Nir Modiano,
  2. Yanping E. Lu, and
  3. Peter Cresswell*
  1. Section of Immunobiology, Howard Hughes Medical Institute, Yale University School of Medicine, New Haven, CT 06520-8011

  1. Contributed by Peter Cresswell, April 20, 2005

 
Next Section

Abstract

Human guanylate-binding protein-1 (hGBP-1) is a large GTPase, similar in structure to the dynamins. Like many smaller GTPases of the Ras/Rab family, it is farnesylated, suggesting it may dock into membranes and perhaps play a role in intracellular trafficking. To date, however, hGBP-1 has never been associated with a specific intracellular compartment. Here we present evidence that hGBP-1 can associate with the Golgi apparatus. Redistribution from the cytosol to the Golgi was observed by immunofluorescence and subcellular fractionation after aluminum fluoride treatment, suggesting that it occurs when hGBP-1 is in its GTP-bound state. Relocalization was blocked by a farnesyl transferase inhibitor. The C589S mutant of hGBP-1, which cannot be farnesylated, and the previously uncharacterized R48P mutant, which cannot bind GTP, both failed to localize to the Golgi. These two mutants had a dominant-negative effect, preventing endogenous wild-type hGBP-1 from efficiently redistributing after aluminum fluoride treatment. Furthermore, hGBP-1 requires another IFN-γ-induced factor to be targeted to the Golgi, because constitutively expressed hGBP-1 remained cytosolic in cells treated with aluminum fluoride unless the cells were preincubated with IFN-γ. Finally, two nonhydrolyzing mutants of hGBP-1, corresponding to active mutants of Ras family proteins, failed to constitutively associate with the Golgi; we propose three possible explanations for this surprising result.

Guanylate-binding proteins (GBPs) belong to a family of large GTPases that includes dynamins and Mx proteins (1, 2). These proteins are characterized by their ability to oligomerize and can display oligomerization-dependent stimulation of GTP hydrolysis (3). Human GBP-1 (hGBP-1) may play a role in inhibiting endothelial cell function and proliferation (4-6) and perhaps in inhibiting viral replication (7). Mammals express a number of homologous GBPs, which may have overlapping functions. Most inbred mouse strains have a dysfunctional allele for murine GBP-1 that cannot be induced by IFNs (8, 9), yet they appear healthy, consistent with the idea that GBPs may compensate for one another.

We focus here on one of the five known human GBPs, hGBP-1, a 67-kDa farnesylated protein induced by IFN-γ and other inflammatory cytokines. Our approach is rooted in work done on smaller isoprenylated GTPases, the Ras/Rab-family proteins, and on the myristolated ADP-ribosylation factor GTPases. These GTPases are activated when bound to GTP and in this active state commonly anchor into a specific target membrane (10). A number of tools have been used to study these proteins, including aluminum fluoride (AlF). AlF can enter live cells and bind to GDP-bound proteins, producing a complex structure with the nucleotide resembling the GTP-bound form of the molecule (or the GTP hydrolysis transition state) (11-13). AlF can induce relocalization of some GTPases to their target membranes (14, 15). Membrane association requires a hydrophobic modification, such as a myristoyl or farnesyl group. Farnesyl transferase inhibitors were developed in attempts to block the known oncogenic properties of the Ras proteins (16-18). These drugs block the enzyme that adds the hydrophobic 15-carbon farnesyl group to proteins containing the CaaX box, or isoprenylation motif, at the C terminus of the proteins. They appear to inactivate Ras oncogenes and lead to tumor regression in mice. Mutagenesis has also been invaluable in the understanding of the functional characteristics of the small GTPases. Inactive mutants, which do not bind nucleotide, can have a dominant-negative effect on the endogenous protein (19). Active mutants, which bind GTP but do not hydrolyze it, remain GTP-bound and localize to their target membranes (20, 21). Mutation of the cysteine residue in the CaaX box prevents isoprenylation and anchoring into membranes (22).

Here we show that AlF treatment induces both endogenous and exogenously expressed wild-type hGBP-1 to redistribute from the cytosol to the Golgi in a manner that requires both isoprenylation and pretreatment with IFN-γ. We describe a mutant that cannot bind GTP, as well as two mutants that will bind but not hydrolyze GTP. These mutants may prove useful in examining the functional roles of hGBP-1.

Previous SectionNext Section

Materials and Methods

Cell Lines. HeLa M is a highly IFN-γ-inducible derivative of HeLa cervical carcinoma cells and was a gift from Peter Lengyel (Yale School of Medicine). HFFs are primary fibroblasts recovered from human foreskin (23). HeLa M cells were grown in Iscove's Modified Dulbecco's Medium (IMDM; GIBCOBRL) containing 10% bovine calf serum at 37°C. HFFs were grown at 37°C in IMDM containing 10% FBS, 100 units/ml of penicillin and 100 μg/ml of streptomycin (GIBCOBRL).

Antibodies, Inhibitors, and Modifying Reagents. The rabbit antiserum R.GBP-1 was raised against full-length recombinant human GBP-1 and efficiently recognized hGBP-1 by immunoprecipitation, immunofluorescence, and Western blot. It was affinity-purified to eliminate background immunofluorescence staining. The anti-HLA-A3 mAb, GAP.A3, was previously described (24). We obtained a mAb against the Golgi matrix protein GM130 from BD Transduction Laboratories and an anti-Myc mAb from Invitrogen. Secondary Abs for immunofluorescence were from Molecular Probes. Affinity-purified sheep antiserum to human TGN46 was from Serotec, anti-Rab5 (Clone 1) mAb from BD Transduction Laboratories, and rabbit anti-IκBα (C-21) from Santa Cruz Biotechnology. The farnesyl transferase inhibitor, L-744,832, was provided by Merck Research Laboratories (West Point, PA) and has been described (16).

Cloning and Mutagenesis. Full-length hGBP-1 was cloned by using standard PCR techniques from a cDNA library prepared from IFN-γ-treated primary human monocytes (25), and an N-terminal myc-tag was incorporated by PCR. Mutagenesis was achieved with the QuikChange system (Stratagene) by using a primer with an incorporated point mutation and a corresponding reverse complement primer for each mutant. Use of Pfu Turbo enzyme (Strategene) proved important for mutagenesis.

Vectors and Constructs. Myc-tagged hGBP-1 and its mutants were subcloned from PCR Blunt II Topo (Invitrogen) into pProEx HTa (GIBCO/BRL) by using EcoRI/SpeI I restriction sites with a partial EcoRI digest of PCR Blunt II Topo, such that the reverse orientation was corrected. pProEx HTa was used for bacterial expression, and purification was via an N-terminal 6×HIS tag present in the vector. For mammalian expression, myc-tagged hGBP-1 and its mutants were also subcloned from PCR Blunt II Topo into the retroviral vector pLNCX2 (Clontech) by using NotI/ClaI restriction sites.

Preparation of Recombinant Protein. Transformed Top10 cells were grown and induced for 4 h with isopropyl thiogalactoside under standard conditions. Frozen cells were resuspended at 4°C in 0.5 M NaCl/0.05 M Tris·HCl, pH 8.0, with 5.0 mM MgCl2/10 mM imidazole/10% glycerol/0.1% polyoxyethylene 9 lauryl ether (C12E9) containing 0.5 mM PMSF and 5 mM iodoacetamide. Sonicated lysate was clarified at 80,000 × g for 40 min at 4°C and applied to a 1-ml Ni-NTA Agarose (Qiagen, Valencia, CA) column preequilibrated in resuspension buffer. Beads were then washed with resuspension buffer containing 10 mM 2-mercaptoethanol. After additional high- and low-salt (0.15 M NaCl) washes, protein was eluted with 250 mM imidazole in buffer containing 0.15 M NaCl. After dialysis against PBS, samples were filter-sterilized and stored at 4°C.

Transduction and Generation of Stable Lines. HeLa M and HFF cells were retrovirally transduced by using the pLNCX2 constructs. Virus for transduction was produced by using the Pantropic Retroviral Expression System (Clontech), which features vesicular stomatitis virus (VSV)-G as an envelope protein. GP2-293 cells were transfected with the retroviral construct and the VSV-G construct by using Lipofectamine 2000 (Invitrogen). Virus was collected 36 h posttransfection. HeLa M and HFF cells were transduced as suggested by the manufacturer with the following modifications: cells were transduced at 30-40% confluency with 8 μg/ml Polybrene and incubated at 37°C for 4 h with occasional agitation; virus-containing medium was replaced with growth medium for 24 h before selection with 0.6 mg/ml G418 (Sigma Aldrich).

Immunofluorescence. IFN-γ-induced cells were treated with 200 units/ml IFN-γ (R & D Systems) for 18 h. Where indicated, the farnesyl transferase inhibitor L-744,832 was added to 10 μM at the same time as IFN-γ, and AlF was added 15 min before fixation as sodium fluoride (final concentration, 20 mM) and aluminum chloride (final concentration, 50 μM). Cells were fixed in 3.7% formaldehyde, permeabilized in saponin-based buffer, stained for indicated proteins by using fluorescein or Texas red-conjugated secondary antibodies, and examined by using an Axiophot 2 fluorescence microscope (Zeiss), all as described (26). Data were obtained by using openlab software (Improvision, Lexington, MA).

Metabolic Labeling and Immunoprecipitation. For the hGBP-1 induction study, HFFs were treated with IFN-γ for 1, 3, 4, or 24 h. At each time point, 7 × 105 trypsinized cells were suspended in 1 ml of methionine-free DMEM (GIBCO) containing 10 mM Hepes and 3% dialyzed FCS and labeled for 4 h with 0.25 mCi (1 Ci = 37 GBq) of [35S]-methionine (ICN). Cells were lysed in 1.0% Triton X-100 in 0.15 M NaCl/10 mM Tris, pH 7.4, with 0.5 mM PMSF and 5 mM iodoacetamide. After preclearing with normal rabbit serum and Protein A-Sepharose, extracts were immunoprecipitated with R.GBP-1 or control antibody and Protein A-Sepharose. Washed samples were eluted at 100°C in SDS sample buffer containing 2 mM DTT and analyzed by SDS/PAGE and autoradiography.

Subcellular Fractionation and Western Blotting. HFF cells (5 × 106) were treated or mock-treated with IFN-γ for 20 h, followed by 15-min incubation in AlF at 37°C where indicated. Cells were then trypsinized in the presence of AlF for 5 min at 37°C, washed twice with homogenization buffer (HB, 250 mM sucrose/2 mM MgCl2/10 mM Tris, pH 7.6), resuspended in 1.5 ml of HB buffer containing protease inhibitors, and disrupted by using a ball-bearing homogenizer (0.0011-inch clearance). Residual cells and debris were pelleted at 1,500 × g for 10 min at 4°C, and 0.9 ml of the postnuclear supernatant was layered onto a 12-ml sucrose gradient (27). After centrifugation in an SW41 rotor at 39,000 rpm (261,000 × g) for 4 h at 4°C, 1-ml fractions were isolated, and trichloroacetic acid-precipitated proteins were separated by 10.5% SDS/PAGE followed by Western blotting with R.GBP-1, as described (28).

GTP Binding and Hydrolysis. Recombinant hGBP-1 was dialyzed into GTPase reaction buffer (50 mM Tris/5 mM MgCl2/100 mM KCl, pH 8.0) (29). hGBP-1 (5 μg) was incubated in 10 μl of reaction buffer containing 10% glycerol, 0.1 mM DTT, and 6.6 mM (0.2 μCi) [α-32P]GTP (PerkinElmer) for 2 h at 37°C. SDS (2.5%) containing 10 mM EDTA was added to stop the reaction, and GTP, GDP, and GMP were separated by thin-layer chromatography on polyethyleneimine-cellulose sheets (J. T. Baker), as described (30). GTP binding was measured by using GTP-agarose beads (Sigma-Aldrich) equilibrated in GTPase reaction buffer: 25-μl beads were rotated at 4°C with 5 μg of hGBP-1 in a total volume of 100 μl for 30 min. After washing, bound hGBP-1 was eluted in SDS sample buffer and separated by SDS/PAGE. Gels were stained by using Coomassie Simply-Blue SafeStain (Invitrogen).

Previous SectionNext Section

Results

Intracellular Localization of hGBP-1. We first confirmed by radio-labeling, immunoprecipitation, and SDS/PAGE that hGBP-1 was rapidly induced by IFN-γ treatment of human fibroblasts. Expression increased progressively over 24 h (Fig. 1A), consistent with data obtained at the mRNA level (31, 32). Immunofluorescence staining of IFN-γ-treated HeLa M and HFF cells showed a diffuse cytosolic staining pattern with exclusion from the nucleus in both cell types (Fig. 1 B and C Upper). NMR and fluorescence spectrum analysis have shown that AlF can form a complex with GDP-bound hGBP-1 (33), as it does with other GTPases (34). After addition of AlF, a fraction of hGBP-1 rapidly relocalized to a perinuclear compartment in both HeLa M cells and HFFs (Fig. 1 B and C Lower). This compartment was identified as the Golgi apparatus by costaining with an antibody to the Golgi matrix protein GM130. Redistribution occurred within 10-15 min, and the cells were routinely fixed after 15-20 min of AlF exposure. hGBP-1 relocalization did not noticeably change the structure of the Golgi. To confirm that hGBP1 redistribution involves membrane attachment, we used a sucrose flotation equilibrium gradient to separate membranes from cytosol (27, 35). Fig. 2 shows that after AlF treatment, a portion of hGBP-1 floated to low-buoyant-density fractions 1 and 2, consistent with membrane association. The transGolgi marker TGN 46 and the endosomal marker Rab 5 were also found in these low-density fractions, as anticipated (27). In contrast, IκBα, a cytosolic protein, remained in the lower-gradient fractions.

Fig. 1.
View larger version:
Fig. 1.

Induction of hGBP-1 by IFN-γ. (A) hGBP-1 is rapidly induced in the hours after IFN-γ pulse, and the expression level continues to rise through 24 h of IFN-γ induction. Expression shown by immunoprecipitation from HFFs. Note that IFN-γ pulse is followed by 4-h radiolabel. hGBP-1 relocalizes to the Golgi apparatus after AlF treatment in both (B) HeLa M cells and (C) HFFs.


Fig. 2.
View larger version:
Fig. 2.

Golgi association correlates with membrane localization by cell fractionation. HFF cells were pretreated with IFN-γ for 20 h, then treated for 15 min with AlF at 37°C where indicated. Cells were lysed with a ball-bearing homogenizer and the postnuclear supernatant centrifuged on a sucrose gradient to float the membranes. Twelve fractions were collected and analyzed by SDS/PAGE and Western blotting. Earliest fractions (1-3) are the least dense and contain the membranes.


Golgi Localization Requires Active Farnesyl Transferase and IFN-γ Treatment. To examine the role of the farnesyl group of hGBP-1 in AlF-induced Golgi localization, we included the farnesyl transferase inhibitor L-744,832 during IFN-γ treatment. On subsequent addition of AlF, hGBP-1 remained cytosolic (Fig. 3A). We then examined hGBP-1 localization in the absence of IFN-γ induction. This was done in HFFs transduced with retroviral constructs encoding myc-tagged hGBP-1, detecting it with a monoclonal anti-myc antibody. AlF treatment of these cells did not induce relocalization of the exogenous hGBP-1 to the Golgi (Fig. 3B Upper). The great majority of total cellular hGBP-1 is exogenous in cells not treated with IFN-γ (data not shown). Consistent with this, total hGBP-1, stained by using the affinity-purified rabbit hGBP-1 antibody, also failed to redistribute from the cytosol. However, after overnight induction with IFN-γ and treatment with AlF, both total cellular hGBP-1 and exogenous (myc-tagged) hGBP-1 relocalized efficiently to the Golgi (Fig. 3B Lower), suggesting that an additional factor induced by IFN-γ is required.

Fig. 3.
View larger version:
Fig. 3.

hGBP-1 requires farnesylation for Golgi relocalization. Farnesylation requirement for Golgi relocalization (A). AlF-stimulated relocalization of hGBP-1 to the Golgi does not occur when hGBP-1 is induced in the presence of a farnesyl transferase inhibitor (FTI). Exogenously produced hGBP-1 cannot relocalize without treatment with IFN-γ before AlF exposure (B). Note that in B, Texas red fluorescence corresponds to anti-myc tag staining for exogenous hGBP-1.


Biochemical Characterization of hGBP-1 Mutants. Myc-tagged mutants of hGBP-1 were subcloned into a prokaryotic expression vector containing an N-terminal 6×His tag, expressed in Escherichia coli, and purified by Ni-agarose chromatography. Nucleotide-binding assays identified two mutants that would not bind GTP, one of them previously undescribed (R48P, Fig. 4A). In our hands, the published GTP nonbinding mutant (D184N) shows trace levels of GTP binding, whereas the R48P mutant exhibits no detectable GTP binding or hydrolysis (Fig. 4B). The R48P mutant behaves quite differently from a published mutant in which residue 48 is mutated to alanine. That mutant has a slightly higher affinity for the hydrolysis-resistant GTP analog GppNHp (guanosine 5′-[β, γ-imido] triphosphate) than wild-type hGBP-1 but loses its ability to hydrolyze GTP (36). We also generated two mutants, T75A and E99A, that bind but do not efficiently hydrolyze GTP (Fig. 4 A and B). The location of each mutant in the nucleotide-binding pocket of the hGBP-1 crystal structure is shown in Fig. 4C (image generated by using rasmol 2.6) (37).

Fig. 4.
View larger version:
Fig. 4.

Characterization of hGBP-1 mutants. (A) GTP-agarose binding for recombinant wild-type hGBP-1 and its mutants. (B) GTP hydrolysis activity of proteins examined in A.(C) Location of hGBP-1 mutants shown on nucleotide-bound crystal structure. R48P in red, T75A in blue, and E99A in green. C589S is not shown on the structure, because the structure is truncated at the C terminus. Its approximate location on the protein is indicated by *. Analysis was done on rasmol 2.6 software.


Elimination of hGBP-1 Farnesylation or GTP Binding Prevents Targeting to the Golgi. Treatment of cells with the farnesyl transferase inhibitor prevented hGBP-1 Golgi localization. However, farnesylation of another protein rather than specific farnesylation of hGBP-1 could be responsible for this. Mutagenesis of the acceptor cysteine of hGBP-1 to serine (C589S) allowed us to examine this question. Anti-myc staining showed that in HFFs (Fig. 5A) or HeLa M cells (data not shown), this mutant failed to redistribute to the Golgi upon IFN-γ treatment followed by addition of AlF. More strikingly, expression of the C589S mutant also prevented IFN-γ-induced endogenous hGBP-1 from relocalizing, as shown by total hGBP-1 staining. Endogenous hGBP-1 levels after IFN-γ induction are extremely high in HFFs, so that endogenous hGBP-1 would be clearly visible. The R48P mutant, which cannot bind GTP, was also expressed in both HFF (shown) and HeLa M cells. Once again, IFN-γ treatment followed by brief exposure to AlF did not induce the myc-tagged exogenous protein to localize to the Golgi (Fig. 5B). Furthermore, as with the farnesylation mutant, expression of R48P caused most if not all of the endogenous hGBP-1 to remain in the cytosol.

Fig. 5.
View larger version:
Fig. 5.

Molecular requirements for hGBP-1 recruitment to the Golgi. Neither the farnesyl mutant (A) nor the R48P GTP nonbinding mutant (B) of hGBP-1 can relocalize to the Golgi upon AlF treatment. Both mutants appear to act in a dominant-negative fashion, suppressing the ability of endogenous IFN-induced hGBP-1 to relocalize to the Golgi; fluorescein staining is with rabbit-anti-hGBP1 and reflects both exogenous and endogenous hGBP-1, whereas Texas red staining is for myc-tagged (exogenous) hGBP-1. In C, a mutant that can bind but not hydrolyze GTP (T75A) cannot relocalize to the Golgi without AlF treatment, even with IFN-γ pretreatment. It does relocalize to the Golgi efficiently in the presence of AlF and IFN-γ.


GTP Hydrolysis Is Not Required for Golgi Localization of hGBP-1 Nor Is Lack of Hydrolysis Sufficient to Induce It. The two mutants that bind GTP but do not hydrolyze it efficiently associated with the Golgi after IFN-γ induction and AlF treatment (E99A, data not shown, and T75A, Fig. 5C Bottom). However, neither of these hGBP-1 mutants localized to the Golgi in the absence of AlF, regardless of exposure of the cells to IFN-γ (Fig. 5C Top and Middle). This result differs from data obtained with other GTPases, such as the ADP-ribosylation factor family proteins, where similar GTP-binding nonhydrolytic mutants bind spontaneously to their target membranes.

Previous SectionNext Section

Discussion

The transcriptional regulation of hGBP-1 expression is among the best understood of any protein, and its investigation has played a central role in understanding IFN-γ-induced gene expression. However, the function of hGBP-1 has proven elusive for nearly two decades, although a second set of IFN-γ-induced GTPases, the p47 family, has recently been shown to have antimicrobial properties (38-40). In the last few years, some published work has begun to tackle the role of hGBP-1. One report suggested that hGBP-1 plays an antiviral role and inhibits vesicular stomatitis virus (VSV) and encephalomyocarditis virus (EMCV) replication (7). The strongest data in that study came from experiments where hGBP-1 expression was blocked by using antisense RNA. More recent work has proposed that murine GBP-2 (mGBP-2) is antiviral; however, cells expressing mGBP-2 showed decreases in VSV or EMCV yield of only ≈50%, which is not consistent with the log-order changes typical of antiviral proteins (41). We have attempted to corroborate the data for hGBP-1 by using alternative systems but have not been able to conclusively confirm an antiviral function (unpublished work). Recent studies have shown that hGBP-1 plays an important role in endothelial cell function. hGBP-1 induction was found predominantly in endothelial cells during inflammation in vivo (6). In an intriguing set of experiments, hGBP-1 interfered with angiogenesis and strongly inhibited the expression of matrix metalloproteinase-1, a collagenase necessary for leukocyte migration through the extracellular matrix (5). How hGBP-1 directs these actions in endothelial cells is not known.

Based on experiments with the smaller GTPases, e.g., Rabs, Racs, and ADP-ribosylation factors, we used AlF to simulate the GTP-bound state of hGBP-1. In cells induced with IFN-γ and treated with AlF, hGBP-1 redistributed from its diffuse cytosolic localization to the Golgi apparatus. The observed redistribution correlates with a shift to membranes upon subcellular fractionation and is reminiscent of the smaller GTPases, where relocalization depends on the presence of a hydrophobic lipid moiety, which anchors the protein to its target membrane. Prior work using the isoprenoid precursor [3H]-mevalonate showed that hGBP-1 is effectively farnesylated in cell culture (42) yet remains largely cytosolic. This contrasts with mGBP-2, which constitutively attaches to an unknown vesicular membrane in an isoprenylation-dependent manner (43). Although highly homologous to mGBP-2, mGBP-1 is diffusely cytosolic like hGBP-1. However, unlike hGBP-1, mGBP-1 is poorly isoprenylated in cell culture (44).

Using both a farnesyl transferase inhibitor and the C589S mutant of hGBP-1, we demonstrated that farnesylation is required for Golgi targeting, suggesting that in its GTP-bound form, hGBP-1 exposes its farnesyl group and anchors itself selectively to the cytosolic face of the Golgi. To eliminate the possibility that AlF can cause hGBP-1 redistribution through its actions on another protein rather than directly on hGBP-1, we used an hGBP-1 mutant (R48P) that does not bind GTP. This mutant did not relocalize to the Golgi upon AlF treatment. Thus, nucleotide binding by hGBP-1 is required, consistent with the hypothesis that the Golgi is the target membrane for the GTP-bound form. Two members of the p47 GTPase family, IIGP1 and LRG-47, are also associated with the Golgi but in a constitutive manner (45). Redistribution of LRG-47 to the plasma membrane and phagosomes follows phagocytosis and may be responsible for promoting its antimicrobial function.

Published data for other GTPases suggest that NaF treatment alone can have the same effect as the combination of NaF with AlCl3; this may be due to the presence of magnesium in culture media (46, 47) or a direct effect of fluoride on the protein (48). We indeed found that NaF alone produced the same effect as AlF, with efficient relocalization of hGBP-1 to the Golgi (data not shown). Because AlF is the commonly accepted treatment, and because hGBP-1 has been formally shown to form complexes with AlF (33), we elected to combine NaF and AlCl3 in our experiments.

In addition to being unable to relocalize to the Golgi themselves, both the C589S and the R48P mutant inhibited the redistribution of endogenous hGBP-1 induced by IFN-γ. Like dynamins and Mx proteins, hGBP-1 has a large C-terminal domain and appears to oligomerize. We speculate that hGBP-1 may function through higher-order structures, and that introduction of nonfunctional GBPs into these oligomers may interfere with the activity of the entire complex. Whether or not this is the case, that both inactive mutants interfere with Golgi relocalization of endogenous protein will make these mutants useful in the study of hGBP-1 function. If Golgi association is required for the biological role of hGBP-1, mutants that prevent it may eliminate its cellular function. Indeed, there are now published data that the D184N mutant of hGBP-1, which also cannot bind or hydrolyze GTP, cannot itself inhibit matrix metalloproteinase-1(MMP-1) expression in endothelial cells and successfully abolishes the effect of IL-1β-induced wild-type hGBP-1 on MMP-1 levels (5).

We also identified hGBP-1 mutants that bind but do not efficiently hydrolyze GTP, corresponding to the active mutants that have been described for Ras family proteins. The T75A and E99A mutants described here were recently identified independently by another group (36). Their studies showed that the T75A mutant is indeed a very poor catalyst of GTP hydrolysis, with a 1,500-fold decrease in specific activity relative to wild-type. The E99A mutant, on the other hand, showed only a 15-fold decrease in specific activity. Although E99A retained the ability to bind GTP-agarose under the conditions used in our experiments, the published study found a 25-fold decreased affinity for mant-GppNHp. Qualitatively, in our hands, both T75A and E99A show poor hydrolysis while retaining the ability to bind GTP. We initially expected these mutants to constitutively associate with the Golgi without AlF treatment. When they did not, we reasoned that the nonhydrolyzing mutants might require the unidentified IFN-γ-induced factor that was necessary for AlF-induced relocalization of wild-type hGBP-1. However, even after IFN-γ induction, these mutants remained diffusely cytosolic unless the cells were treated with AlF.

We propose three possible explanations for the behavior of the nonhydrolyzing mutants. First, IFN-γ also induces significant quantities of endogenous hGBP-1, which may have a dominant-negative effect on these mutants in the absence of AlF; like the R48P mutant that will not bind GTP, the wild-type protein is primarily not GTP-bound, and oligomerization with a GTP-bound mutant may prevent the mutant from localizing to the Golgi. Second, activation of an accessory GTPase may be necessary for hGBP-1 relocalization, such that global activation by AlF is effective, but activation of hGBP-1 alone is not sufficient. A third possibility is that these mutants are not active because the binding properties of hGBP-1 with nucleotides are very different from the small GTPases; hGBP-1 will bind to any guanylate nucleotide (GMP, GDP, and GTP) with similar affinity, unlike Ras proteins, which bind to GDP and GTP with an affinity orders of magnitude greater than binding to GMP. This difference reflects the fact that hGBP-1 binds very weakly to GDP and GTP, with a K d value in the micromolar range, whereas Ras-related proteins bind GDP and GTP with affinities in the picomolar range (1, 49). Furthermore, the T75A and E99A mutants have even lower affinities for GTP than wild-type hGBP-1 (36). Thus, nucleotide exchange may be a significant factor in the behavior of hGBP-1, negating the shift toward GTP binding when GTP hydrolysis is blocked. In vivo, the guanylate-binding state may be regulated by specific cytosolic conditions or signaling events, allowing a coordinated action to take place. Anchoring into the cytosolic face of the Golgi is likely to be a key component of that action.

Previous SectionNext Section

Conclusion

This work presents compelling evidence that hGBP-1 can target specifically to the Golgi membrane in its GTP-bound form. In so doing, it also introduces a rapid-outcome measure for hGBP-1 behavior, which we used to identify inactive mutants and dominant-negative activities. In addition, we examine two nonhydrolyzing mutants, which prove not to be constitutively active with respect to Golgi localization. Through examining the phenomenon of hGBP-1 relocalization to the Golgi apparatus, we have broadened our understanding of this protein while developing tools that can be applied to various aspects of its study.

Previous SectionNext Section

Acknowledgments

We thank Mary Pan for technical assistance and for preparation of this manuscript. This work was funded by the Howard Hughes Medical Institute and the Ellison Medical Foundation. N.M. was supported by a National Institutes of Health/National Institute of General Medical Sciences Medical Scientist Training Grant (GM07205) and Y.E.L. was supported by the Cancer Research Institute.

Previous SectionNext Section

Footnotes

  • ↵ * To whom correspondence should be addressed. E-mail: peter.cresswell{at}yale.edu.

  • Author contributions: P.C. designed research; N.M. and Y.E.L. performed research; N.M. analyzed data; and N.M. and P.C. wrote the paper.

  • Abbreviations: GBP, guanylate-binding protein; hGBP-1, human GBP; mGBP, murine GBP; AlF, aluminum fluoride; HFF, human foreskin fibroblast.

Previous Section
 

References

  1. ↵
    Prakash, B., Renault, L., Praefcke, G. J., Herrmann, C. & Wittinghofer, A. (2000) EMBO J. 19 , 4555-4564.pmid:10970849
    CrossRefMedline
  2. ↵
    Prakash, B., Praefcke, G. J., Renault, L., Wittinghofer, A. & Herrmann, C. (2000) Nature 403 , 567-571.pmid:10676968
    CrossRefMedline
  3. ↵
    Warnock, D. E., Hinshaw, J. E. & Schmid, S. L. (1996) J. Biol. Chem. 271 , 22310-22314.pmid:8798389
    Abstract/FREE Full Text
  4. ↵
    Guenzi, E., Topolt, K., Cornali, E., Lubeseder-Martellato, C., Jorg, A., Matzen, K., Zietz, C., Kremmer, E., Nappi, F., Schwemmle, M., et al. (2001) EMBO J. 20 , 5568-5577.pmid:11598000
    CrossRefMedlineISI
  5. ↵
    Guenzi, E., Topolt, K., Lubeseder-Martellato, C., Jorg, A., Naschberger, E., Benelli, R., Albini, A. & Sturzl, M. (2003) EMBO J. 22 , 3772-3782.pmid:12881412
    CrossRefMedline
  6. ↵
    Lubeseder-Martellato, C., Guenzi, E., Jorg, A., Topolt, K., Naschberger, E., Kremmer, E., Zietz, C., Tschachler, E., Hutzler, P., Schwemmle, M., et al. (2002) Am. J. Pathol. 161 , 1749-1759.pmid:12414522
    Abstract/FREE Full Text
  7. ↵
    Anderson, S. L., Carton, J. M., Lou, J., Xing, L. & Rubin, B. Y. (1999) Virology 256 , 8-14.pmid:10087221
    CrossRefMedlineISI
  8. ↵
    Staeheli, P., Prochazka, M., Steigmeier, P. A. & Haller, O. (1984) Virology 137 , 135-142.pmid:6089411
    CrossRefMedline
  9. ↵
    Prochazka, M., Staeheli, P., Holmes, R. S. & Haller, O. (1985) Virology 145 , 273-279.pmid:3927587
    CrossRefMedline
  10. ↵
    Chavrier, P. & Goud, B. (1999) Curr. Opin. Cell Biol. 11 , 466-475.pmid:10449335
    CrossRefMedlineISI
  11. ↵
    Wittinghofer, A. (1997) Curr. Biol. 7 , R682-R685.pmid:9382787
    CrossRefMedlineISI
  12. Hoffman, G. R., Nassar, N., Oswald, R. E. & Cerione, R. A. (1998) J. Biol. Chem. 273 , 4392-4399.pmid:9468490
    Abstract/FREE Full Text
  13. ↵
    Mittal, R., Ahmadian, M. R., Goody, R. S. & Wittinghofer, A. (1996) Science 273 , 115-117.pmid:8658179
    Abstract
  14. ↵
    Radhakrishna, H., Klausner, R. D. & Donaldson, J. G. (1996) J. Cell Biol. 134 , 935-947.pmid:8769418
    Abstract/FREE Full Text
  15. ↵
    Finazzi, D., Cassel, D., Donaldson, J. G. & Klausner, R. D. (1994) J. Biol. Chem. 269 , 13325-13330.pmid:8175763
    Abstract/FREE Full Text
  16. ↵
    Kohl, N. E., Omer, C. A., Conner, M. W., Anthony, N. J., Davide, J. P., deSolms, S. J., Giuliani, E. A., Gomez, R. P., Graham, S. L., Hamilton, K., et al. (1995) Nat. Med. 1 , 792-797.pmid:7585182
    CrossRefMedlineISI
  17. Gibbs, J. B., Oliff, A. & Kohl, N. E. (1994) Cell 77 , 175-178.pmid:8168127
    CrossRefMedlineISI
  18. ↵
    Prendergast, G. C. (2000) Curr. Opin. Cell Biol. 12 , 166-173.pmid:10712917
    CrossRefMedlineISI
  19. ↵
    Kaplan, J. B. (1994) Oncol. Res. 6 , 611-615.pmid:7787254
    Medline
  20. ↵
    Peters, P. J., Hsu, V. W., Ooi, C. E., Finazzi, D., Teal, S. B., Oorschot, V., Donaldson, J. G. & Klausner, R. D. (1995) J. Cell Biol. 128 , 1003-1017.pmid:7896867
    Abstract/FREE Full Text
  21. ↵
    Gomes, A. Q., Ali, B. R., Ramalho, J. S., Godfrey, R. F., Barral, D. C., Hume, A. N. & Seabra, M. C. (2003) Mol. Biol. Cell 14 , 1882-1899.pmid:12802062
    Abstract/FREE Full Text
  22. ↵
    Kato, K., Cox, A. D., Hisaka, M. M., Graham, S. M., Buss, J. E. & Der, C. J. (1992) Proc. Natl. Acad. Sci. USA 89 , 6403-6407.pmid:1631135
    Abstract/FREE Full Text
  23. ↵
    Howell, D. N., Machamer, C. E. & Cresswell, P. (1982) Hum. Immunol. 5 , 233-238.pmid:6983518
    Medline
  24. ↵
    Berger, A. E., Davis, J. E. & Cresswell, P. (1982) Hybridoma 1 , 87-90.pmid:6208124
    MedlineISI
  25. ↵
    Chin, K. C. & Cresswell, P. (2001) Proc. Natl. Acad. Sci. USA 98 , 15125-15130.pmid:11752458
    Abstract/FREE Full Text
  26. ↵
    Hammond, C., Denzin, L. K., Pan, M., Griffith, J. M., Geuze, H. J. & Cresswell, P. (1998) J. Immunol. 161 , 3282-3291.pmid:9759843
    Abstract/FREE Full Text
  27. ↵
    Harley, C. A., Dasgupta, A. & Wilson, D. W. (2001) J. Virol. 75 , 1236-1251.pmid:11152497
    Abstract/FREE Full Text
  28. ↵
    Dick, T. P., Bangia, N., Peaper, D. R. & Cresswell, P. (2002) Immunity 16 , 87-98.pmid:11825568
    CrossRefMedlineISI
  29. ↵
    Schwemmle, M. & Staeheli, P. (1994) J. Biol. Chem. 269 , 11299-11305.pmid:7512561
    Abstract/FREE Full Text
  30. ↵
    Liang, Z., Mather, T. & Li, G. (2000) Biochem. J. 346 , 501-508.pmid:10677372
    Medline
  31. ↵
    Lew, D. J., Decker, T. & Darnell, J. E., Jr. (1989) Mol. Cell. Biol. 9 , 5404-5411.pmid:2555698
    Abstract/FREE Full Text
  32. ↵
    Decker, T., Lew, D. J., Cheng, Y. S., Levy, D. E. & Darnell, J. E., Jr. (1989) EMBO J. 8 , 2009-2014.pmid:2507314
    MedlineISI
  33. ↵
    Praefcke, G. J., Geyer, M., Schwemmle, M., Robert Kalbitzer, H. & Herrmann, C. (1999) J. Mol. Biol. 292 , 321-332.pmid:10493878
    CrossRefMedline
  34. ↵
    Ahmadian, M. R., Mittal, R., Hall, A. & Wittinghofer, A. (1997) FEBS Lett. 408 , 315-318.pmid:9188784
    CrossRefMedline
  35. ↵
    Beckers, C. J., Keller, D. S. & Balch, W. E. (1987) Cell 50 , 523-534.pmid:3038335
    CrossRefMedlineISI
  36. ↵
    Praefcke, G. J., Kloep, S., Benscheid, U., Lilie, H., Prakash, B. & Herrmann, C. (2004) J. Mol. Biol. 344 , 257-269.pmid:15504415
    CrossRefMedline
  37. ↵
    Sayle, R. A. & Milner-White, E. J. (1995) Trends Biochem. Sci. 20 , 374.pmid:7482707
    CrossRefMedlineISI
  38. ↵
    MacMicking, J. D., Taylor, G. A. & McKinney, J. D. (2003) Science 302 , 654-659.pmid:14576437
    Abstract/FREE Full Text
  39. Taylor, G. A., Collazo, C. M., Yap, G. S., Nguyen, K., Gregorio, T. A., Taylor, L. S., Eagleson, B., Secrest, L., Southon, E. A., Reid, S. W., et al. (2000) Proc. Natl. Acad. Sci. USA 97 , 751-755.pmid:10639151
    Abstract/FREE Full Text
  40. ↵
    Taylor, G. A., Feng, C. G. & Sher, A. (2004) Nat. Rev. Immunol. 4 , 100-109.pmid:15040583
    CrossRefMedlineISI
  41. ↵
    Carter, C. C., Gorbacheva, V. Y. & Vestal, D. J. (2005) Arch. Virol., in press.
  42. ↵
    Nantais, D. E., Schwemmle, M., Stickney, J. T., Vestal, D. J. & Buss, J. E. (1996) J. Leukocyte Biol. 60 , 423-431.pmid:8830800
    Abstract
  43. ↵
    Vestal, D. J., Gorbacheva, V. Y. & Sen, G. C. (2000) J. Interferon Cytokine Res. 20 , 991-1000.pmid:11096456
    CrossRefMedline
  44. ↵
    Stickney, J. T. & Buss, J. E. (2000) Mol. Biol. Cell 11 , 2191-2200.pmid:10888661
    Abstract/FREE Full Text
  45. ↵
    Martens, S., Sabel, K., Lange, R., Uthaiah, R., Wolf, E. & Howard, J. C. (2004) J. Immunol. 173 , 2594-2606.pmid:15294976
    Abstract/FREE Full Text
  46. ↵
    De Boer, A. H., Van der Molen, G. W., Prins, H. B., Korthout, H. A. & Van der Hoeven, P. C. (1994) Eur. J. Biochem. 219 , 1023-1029.pmid:8112315
    Medline
  47. ↵
    Graham, D. L., Eccleston, J. F., Chung, C. W. & Lowe, P. N. (1999) Biochemistry 38 , 14981-14987.pmid:10555980
    CrossRefMedline
  48. ↵
    Vincent, S., Brouns, M., Hart, M. J. & Settleman, J. (1998) Proc. Natl. Acad. Sci. USA 95 , 2210-2215.pmid:9482864
    Abstract/FREE Full Text
  49. ↵
    John, J., Sohmen, R., Feuerstein, J., Linke, R., Wittinghofer, A. & Goody, R. S. (1990) Biochemistry 29 , 6058-6065.pmid:2200519
    CrossRefMedline
« Previous | Next Article »Table of Contents

This Article

  1. Published online before print June 3, 2005, doi: 10.1073/pnas.0503227102 PNAS June 14, 2005 vol. 102 no. 24 8680-8685
  1. AbstractFree
  2. Figures Only
  3. » Full Text
  4. Full Text (PDF)

Classifications

Social Bookmarking