Regulation of Immature Dendritic Cell Migration by RhoA Guanine Nucleotide Exchange Factor Arhgef5*

There are a large number of Rho guanine nucleotide exchange factors, most of which have no known functions. Here, we carried out a short hairpin RNA-based functional screen of Rho-GEFs for their roles in leukocyte chemotaxis and identified Arhgef5 as an important factor in chemotaxis of a macrophage phage-like RAW264.7 cell line. Arhgef5 can strongly activate RhoA and RhoB and weakly RhoC and RhoG, but not Rac1, RhoQ, RhoD, or RhoV, in transfected human embryonic kidney 293 cells. In addition, Gβγ interacts with Arhgef5 and can stimulate Arhgef5-mediated activation of RhoA in an in vitro assay. In vivo roles of Arhgef5 were investigated using an Arhgef-5-null mouse line. Arhgef5 deficiency did not affect chemotaxis of mouse macrophages, T and B lymphocytes, and bone marrow-derived mature dendritic cells (DC), but it abrogated MIP1α-induced chemotaxis of immature DCs and impaired migration of DCs from the skin to lymph node. In addition, Arhgef5 deficiency attenuated allergic airway inflammation. Therefore, this study provides new insights into signaling mechanisms for DC migration regulation.

Leukocyte chemotaxis underlies leukocyte migration, infiltration, trafficking, and homing that are not only important for normal leukocyte functions, but also have a important role in inflammation-related diseases. Leukocyte chemotaxis is regulated by leukocyte chemoattractants that include bacterial byproducts such as formylmethionylleucylphenylalanine, complement proteolytic fragments such as C5a, and the superfamily of chemotactic cytokines, chemokines. These chemoattractants bind to their specific cell G protein-coupled receptors and are primarily coupled to the G i family of G proteins to regulate leukocyte chemotaxis. Previous studies have established that the Rho family of small GTPases regulates leukocyte migration (1,2). Rac, Cdc42, and RhoA are the three best studied Rho small GTPases. In myeloid cells, Cdc42 regulates directionality by directing where F-actin and lamellipodia are formed, and Rac regulates F-actin formation in the lamellipodia, which provides a driving force for cell motility (3)(4)(5)(6). On the other hand, RhoA regulates the formation and contractility of the actomyosin structure at the back that provides a pushing force (5,7). Rho guanine nucleotide exchange factors (GEF) 3 are key regulators for the activity of these small GTPases. GEFs activate small GTPases by promoting the loading of GTP to the small GTPases, a rate-limiting step in GTPase regulation (8 -11). Previous biochemical and genetic studies have revealed how Cdc42 and Rac may be regulated by chemokine receptors in leukocytes. Chemokine receptors can regulate Cdc42 via a Rho-GEF PIX␣, which is regulated by G␤␥ from the G i proteins via the interactions between G␤␥ and Pak1 and between Pak1 and PIX␣ in myeloid cells 12. On the other hand, in neutrophils chemokine receptors regulate Rac2 via another Rho-GEF P-Rex1, which is directly regulated by G␤␥ (13)(14)(15). Two Rho-GEFs have been implicated in regulation of RhoA in neutrophils. GEF115 was found in the leading edges of polarized mouse neutrophils, whereas PDZ Rho-GEF was found in the uropods of differentiated HL-60 cells. Both Rho-GEFs were believed to mediate pertussis toxin-resistant activation of RhoA in these cells. However, a significant portion of RhoA activity in leukocytes are pertussis toxin-sensitive, which is presumably regulated by the ␣ and/or ␤␥ subunits from the G i proteins. The signaling mechanism for this pertussis toxin-sensitive RhoA regulation by chemokine receptors remains largely elusive.
Molecular cloning and genomic sequencing have identified more than 70 Rho-GEFs in mammals (16 -20). Many of these Rho-GEFs have been shown to activate RhoA in in vitro and overexpression assays (16 -20). However, it is not known if any of them regulate RhoA in vivo, we have found that PIX␣ is a specific GEF for Cdcd42 in neutrophils (12) despite its potent activity on Rac in in vitro and overexpression assays (21,22). Therefore, we used a siRNA-based loss of function screen in an attempt to identify the GEFs that regulate myeloid cell migration and RhoA activity. One of the candidates, Arhgef5, was found to be directly activated by G␤␥ to regulate RhoA and has an important role in immature DC migration. In addition, Arhgef5 deficiency attenuated allergic airway inflammation in a mouse model.

MATERIALS AND METHODS
Plasmids, Protein Preparation, Cell Culture, and Transfection-The full-length Arhgef5 cDNA and its truncated mutants were amplified and subcloned into a mammalian expression vector carrying a FLAG tag. The exchange activitydeficient Arhgef5 mutant Arhgef5DH was generated by substituting Ala residues for Leu 245 and Leu 246 . For the in vitro binding assay, Arhgef5 cDNA was subcloned into the pET21-His vector. His-tagged Arhgef5 was expressed in BL21(DE3)-competent cells and subsequently purified with nickel-nitrilotriacetic acid (Qiagen). Recombinant RhoA, Rhoteckin-RBD, and PAK-PBD, and Elmo were prepared from bacteria as GST fusion proteins. The G␤1␥2 protein was prepared as previously described (12). The cDNAs of RhoB, RhoC, RhoD, RhoG, and RhoQ as well as the dominant negative forms of RhoC, RhoD, and RhoF were acquired from the Missouri S&T cDNA Resources Center carrying an HA tag at their N termini. DNA sequences of all expression constructs were verified by sequencing. Antibodies specific to HA, His, and FLAG were acquired from Covance.
The method to culture immature and mature dendritic cells from mouse bone marrow was described in detail (44). Briefly, mouse bone marrow cells were collected by flushing the femurs. After lysis of the red blood cells, the cells were cultured in the Dulbecco's modified Eagle's medium supplemented with 20 ng/ml mouse recombinant granulocyte macrophage colonystimulating factor (PeproTech). After 5 days, cultured cells were collected and used for immature DC migration assays. Alternatively, the immature DCs were treated with 1 g/ml lipopolysaccharide (Sigma) for 24 h and used for mature DC migration assays.
shRNAs Screening and Chemotaxis Assays-The shRNA vector, which was named pAS, was modified based on pSuper (23) by incorporating a GFP-luciferase fusion protein expression unit. The sequences for these shRNAs are shown in supplemental Table S1. For screening, the shRNAs and control vector were transfected into Raw264.7 cells using Lipofectamine Plus. The cells were collected 48 h later by trypsinization and resuspended in Dulbecco's modified Eagle's medium containing 1% fetal bovine serum. They were then loaded into the upper chambers of 24-well transwell plates (Costar, 5 M pore size). The lower chambers were filled with the same medium, but supplemented with 10 nM C5a (Sigma). The plates were incubated at 37°C for 4 h. Migrated cells were detached from the lower surface of the transwell inserts by trypsin and EDTA and lysed for luciferase assays. The chemotactic indices were calculated by dividing the luciferase activity of migrated cells in the presence of C5a by that of its absence.
For DC migration assays, BM-derived immature DCs or mature DCs were collected and resuspended in Dulbecco's modified Eagle's medium containing 1% fetal bovine serum and loaded onto the upper chambers of transwell plates. Immature DCs were stimulated with 300 ng/ml MIP-1␣ (PeproTech), whereas mature DCs were stimulated with 60 ng/ml CCL-19 (PeproTech) for 4 h at 37°C. Migrated cells were then counted and stained with CD11c for flow cytometric analysis. The chemotaxis indices were calculated by dividing the number of migrated immature (CD11c mid ) or mature (CD11c high ) DCs in the presence of a chemotactic ligand by that in its absence.
To evaluate the role of Arhgef5 and -15 in mature DC chemotaxis, bone marrow-derived immature DCs were collected and transfected with synthetic siRNA duplex oligos of Arhgef5 and Arhgef15 using the Amaxa Nucleofector system (Amaxa Inc.). Twenty-four hours after transfection, lipopolysaccharide was added into the culture to induce DC maturation. Mature DCs were collected for the transwell migration assay as described above 24 h after the induction. The target sequences for the Arhgef5 and Arhgef15 siRNAs were CAGGAGGAA-TTTAATAATACA and AAGTATTAAATTAATCTAATA, respectively.
For evaluating lymphocyte migration, splenocytes were used in the transwell assay in response to 10 nM SDF-1. After 3 h incubation at 37°C, migrated cells were counted and stained with anti-CD3-FITC and anti-B220-R-phycoerythrin. The numbers of migrated T and B cells were determined based on the cell number and their relative percentages, and chemotactic indices were computed as described above. For macrophage chemotaxis assay, macrophages were obtained from the mouse peritonea elicited by thioglycolate. Similar transwell assays were carried out with 10 nM C5a (Sigma) as the stimulus.
GTPase Pulldown Assays-The pulldown assays for determining the activity of small GTPases were carried essentially as previously described. In brief, HEK293T cells were cotransfected with Arhgef5 or its inactive mutant Arhgef5DH with one of the HA-tagged small GTPases. After 24 h transfection, cells were lysed with the lysis buffer (50 mM Tris-HCl (pH 7.3), 10 mM MgCl 2 and 0.2 M NaCl, 2% Nonidet P-40, 10% glycerol, 2 mM orthovanadate) containing recombinant GST-Rhoteckin-RBD for RhoA, RhoB, and RhoC pulldown, GST-PAK-CRIB for Rac, RhoD, and RhoQ pulldown, or GST-Elmo for RhoG pulldown. Bound GTPases were detected by Western analysis with an anti-HA antibody.
RBD Binding Assay-Cells were stimulated with or without 30 ng/ml SDF-1 for 15 s before they were fixed with 4% paraformaldehyde. The cells were then washed and permeabilized by 0.1% Triton X-100 at room temperature for 5 min. After washing, cells were blocked by 1% bovine serum albumin and incubated with purified GST-RBD at room temperature for 1 h followed by Alexa 633-conjugated anti-GST antibody (Molecular Probes, Inc.). After washing, cells were analyzed by a flow cytometer, and the mean fluorescence intensity (MFI) of GFP-positive populations that represent cells carrying the pAS vector was determined.
Guanine Nucleotide Loading Assays-The exchange activity of Arhgef5 was determine for its ability to promote the loading of N-methylanthraniloyl-GTP (Mant-GTP, from JENA Bioscience) to recombinant RhoA as previously described (45). In brief, purified RhoA (0.12 M) was incubated in an assay buffer containing 20 mM Tris (pH 7.5), 150 mM NaCl, 5 mM MgCl 2 , and 1 mM dithiothreitol with 10 M MANT-GTP and recombinant proteins of 0.85 M Arhgef5 and/or 1 M G␤␥. Immediately after mixing, fluorescence intensity was determined by a fluorometer (Wallac Vector, 1420 multilabel counter) with excitation wavelength of 360 nm and emission wavelength of 440 nm.
Luciferase Reporter Assays-HEK293T cells were seeded in 24-well culture plates and transfected with the luciferase reporter construct SRE-luc, normalization plasmid GFP, and other plasmids shown in the figures by using Lipofectamine Plus. After transfection, cells were cultured in serum-free medium for 24 h before the GFP intensity was measured by a fluorometer. The cells were then lysed, and their luciferase activities were determined by a luminometer. Data are presented after the luciferase activity was normalized against the GFP intensity.
Immunoprecipitation-HEK293T cells were cotransfected with G␤1␥2 and FLAG-tagged Arhgef5 or its mutants for 24 h. The cells were lysed with a lysis buffer (1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS, 0.15 M NaCl, 0.01 M sodium phosphate, 50 mM sodium fluoride, 2 mM EDTA, pH 7.2). Immunoprecipitation was then carried out with an anti-FLAG antibody and Protein A/G beads at 4°C for 1 h. The immunocomplexes were subjected to Western analysis with anti-G␤1 antibody.
RT-PCR-Total RNAs were extracted from bone marrowderived immature and mature DCs using the TRIzol reagent (Invitrogen) according to the manufacturer's instructions. RNA was reverse-transcribed using the iScript cDNA synthesis kit (Bio-Rad). RT-PCR was performed. The sense oligo for Ephexin is GAACTGATCGCACAGTTGGA, and the antisense oligo is ATCTTCCGGACACCCTCATT. The sense oligo for Arhgef15 is ATCACTCAGCCCA-AGAGTGG, and the antisense oligo is AGATGGTGTCTGGGGAACAG. The sense oligo for Arhgef5 is TATGTCACCAACCAGACC, and the antisense oligo is ACCTGACTGATGAAGTTCCT.
In Vivo DC Migration Assay-The migration of DCs from the skin to lymph node was determined by FITC skin painting as previously described (46). In brief, FITC (Sigma) was dissolved in a 50:50 (v/v) acetone/dibutylphthalate mixture at a concentration of 5 mg/ml. Mice were anesthetized and their abdominal fur shaved. The FITC solution (0.25 ml/animal) was applied to the shaved skin. Twenty-four hours later, inguinal lymph nodes were harvested and treated with collagenase D (1 mg/ml, Roche) for 20 min at 37°C. The lymph nodes were then smashed onto 70 M cell strainers to produce cell suspensions. The cells were collected and stained with PE-CD11c and analyzed by a flow cytometer.
Generation of Arhgef5-null Mice-A BAC clone that contains the Arhgef5 gene was acquired from the BACPAC Resources Center at the Children's Hospital Oakland Research Institute, Oakland, CA. Exons 5-9, which encode residues Ala 1153 -Lys 1355 in the DH domain of Arhgef5, were floxed with the LoxP sequences in the gene-targeting construct. The Arhgef5 mutant mouse line was generated at the Gene Targeting and Transgenic Facility of the University of Connecticut Health Center using the ES cell line 129S6 derived from 129SvEvTac/ C57BL/6J F1 blastocysts. The chimeric mice were crossed with 129S1-Hprt1-Cre from JAX to produce germline excision of the sequences between the two LoxP sites. Finally, mice heterozygous for the disrupted Arhgef5 gene were interbred to produce homozygous mice. Animals from F1 to F3 were used in this study.
Ovalbumin-induced Asthma Model-Mice were sensitized by intraperitoneal injection with 200 l of OVA-alum suspension (0.5 mg/ml ovalbumin (OVA, Sigma) in phosphate-buffered saline mixed with an equal volume of 20 mg/ml aluminum hydroxide). Eight days after immunization, mice were challenged with an aerosol of 1% FITC-OVA in phosphate-buffered saline, delivered by an ultrasonic nebulizer (OMRON, Compair) for 20 min. FITC-OVA was prepared by mixing 2 mg/ml of FITC solution in carbonate buffer (220 mM, pH 9.6) with OVA at 10 mg/ml. The mixture was gently rotated overnight at 4°C in the dark. Unbound FITC was removed by ultrafiltration using a 10-kDa molecular mass cut-off membrane in a 15-ml filtration cell (Amicon). One day after challenge, the mice were anesthetized. Their trachea were cannulated, and their lungs lavaged five times with 1 ml of pre-chilled phosphate-buffered saline. The bronchial lymph nodes were then collected. Cells in the bronchoalveolar lavage (BAL) fluid were collected by centrifugation at 4000 ϫ g for 5 min. Cytospin preparations of cells were stained with Diff-Quik (Dade Behring) and differentials were performed on 200 cells based on morphology and staining characteristics. The supernatant of BAL fluids were analyzed for IL-4 levels using a mouse IL-4 ELISA kit (Endogen). Lymph nodes were passed through a cell strainer (BD Falcon), and cells were counted and stained with R-phycoerythrin-CD11c (BD Pharmingen) and analyzed by a flow cytometry (Caliber, BD Biosciences).
Intravaginal Infection of Herpes Simplex Virus Type 2 (HSV-2)-The thymidine kinase mutant HSV-2 viruses were prepared and inoculated intravaginally into Arhgef5-null and wild type littermates (1 ϫ 10 6 plaque-forming units of HSV-2) as previously described (47). Vaginal washes were collected daily, and the levels of IFN-␥ and IL-12 were determined by ELISA. On day 6, mice were euthanized, and CD4 ϩ or CD8 ϩ T cells were isolated from iliac and inguinal lymph nodes. The T cells were co-cultured with naïve wild type splenocytes and heat-inactivated HSV-2 of varying plaque-forming units (for CD4 ϩ T cells) or 1 g/ml gB peptide (for CD8 ϩ T cells) for 3 days. The levels of IFN-␥ in the conditioned media were determined by ELISA.
Statistical Analysis-Statistical comparisons between different groups or treatments were performed by unpaired twotailed Student's t test and p Ͻ 0.05 was considered statistically significant.

RESULTS
To investigate the roles of Rho-GEFs in leukocyte chemotaxis, we carried out a siRNA-based functional screen for Rho-GEFs that may play a role in migration of macrophage-like Raw264.7 cells. We generated a mini vector-based shRNA library targeting 38 Rho-GEFs, whose expression could be detected by RT-PCR in RAW264.7 cells (data not shown). The shRNA vector was modified based on pSuper (23) by incorporating a GFP-luciferase fusion protein expression unit; thus, cells producing shRNA can be monitored by the expression of GFP and/or luciferase. This vector was named pAS. All of the GEF shRNAs were validated based their ability to knock down endogenous GEF expression detected by quantitative RT-PCR and/or coexpressed cDNAs if available (data not shown). To test the effects of these shRNAs on RAW264.7 cell migration, cells were transfected with one of the shRNAs, and the empty pAS was used as a negative control. Two days after transfection, transwell migration assays were carried out, and the migratory ability of the cells expressing a GEF shRNA was compared with that of the control cells. Of 38 GEF shRNAs we screened, 6 shRNAs showed more than 50% inhibition ( Fig. 1A and supplemental Table S1). Because siRNAs are known to have off-target effects, we tried to validate the effects by constructing a second shRNA expression plasmid that has a different targeting sequence for these 6 putative hits. We successfully generated the second shRNAs for 5 of these 6 putative hits. Among these 5 shRNAs, only Arhgef5 shRNA showed more than 50% inhibition of RAW264.7 cell migration (supplemental Table S1 and Fig. 1A). To further validate siRNA specificity, we coexpressed an Arhgef5 expression plasmid together with its shRNA plas-mid to determine whether effects of shRNA can be rescued. To prevent the silencing effect of the shRNA on expression of exogenous Arhgef5, we introduced silent mutations at the shRNA targeting sequence of the Arhgef5 cDNA. As shown in Fig. 1B, expression of Arhgef5 effectively rescued the effect of Arhgef5 shRNA on RAW264.7 cell migration. We also tested the effect of Arhgef5 shRNA on the migration of J774 cells; the shRNA could also effectively inhibit its migration, which could be rescued by expression of the silently mutated Arhgef5 (Fig.  1C). The efficiency of this Arhgef5 siRNA was validated as shown in Fig. 1D. Putting all of these results together, we believe that Arhgef5 may have an important role in the migration of these two myeloid cell lines.
Arhgef5, as a member of the Rho-GEF superfamily, possesses a DH-PH tandem domain and a C-terminal SH3 domain. Some members of the subfamily, including Arhgef15 (24), Ephexin-1 (25), and Arhgef5 (26), were shown to activate RhoA, whereas the other (SGEF) was shown to activate RhoG (27). We tested the effects of Arhgef5 on a number of Rho small GTPases. Previous studies have shown that the RhoA-binding domain of Rhotekin (RBD) has a high affinity for a subset of GTP-bound small Rho GTPases that include RhoA, RhoB, and RhoC (28,29), whereas the p21binding domain of PAK1 (PBD) has a high affinity for Rac, RhoQ, RhoD, and RhoV (29,30). Active RhoG was found to be bound to Elmo (31). These interactions have been used to determine the levels of active small GTPases in pulldown assays (28 -30). As shown in Fig. 2A, expression of Arhgef5, but not an Arhgef5 mutant with its DH domain mutated, led to a strong increase in the levels of active RhoA and RhoB, resulting in weak activation of RhoC and RhoG. Expression of Arhgef5 had no effect on the level of active RhoD, RhoV, RhoQ, or Rac1 ( Fig. 2A). Previous studies also showed that GEFs generally exhibit high affinities for the GTP-free mutant forms of small GTPases they regulate (32,33). Consistent with the pulldown assay results, Arhgef5 co-immunoprecipitated with RhoA-N 19 , but not RhoC-N 19 or RhoD-N 31 (Fig. 2B). In addition, RhoF-N 33 did not show detectable interaction with Arhgef5 (Fig. 2B), suggesting that Arhgef5 may not regulate RhoF. These results indicate that Arhgef5, Ephexin-1, and Arhgef15, which are more homologous in amino acid sequences than SGEF, belong to a subgroup that potently activates RhoA rather than RhoG.
Next, we wanted to assess the significance of Arhgef5 in chemoattractant-induced small GTPase activation. Because of the relative low transfection efficiency for leukocytes and the large number of cells required for the pulldown assays particularly for detection of endogenous proteins, we developed a flow cytometry-based approach to assess the GTPase activities. J774 cells expressing Arhgef5 shRNA (G5pAS) or the control vector (pAS) were stimulated with SDF-1. They were then fixed, permeabilized, and incubated with purified GST-RBD and Alexa 633-conjugated anti-GST antibody. The MFI of GFP-positive populations that represent cells carrying the pAS vector was determined by a flow cytometer. As shown in Fig. 2C, SDF-1 treatment resulted in a marked right-shift in MFI in cells expressing the control vector, suggesting that SDF-1 stimulates the activation of small GTPase that can bind to RBD in these cells. However, in cells expressing Arhgef5 shRNA, the SDF-1induced MFI increase was blunted (Fig. 2C). This result suggests that Arhgef5 has a significant role in regulation of endogenous Rho GTPases.
Because RhoA and its close homolog RhoB can be potently activated by Arhgef5, we interpret the inhibitory effect of the Arhgef5 shRNA on RBD binding to suggest that Arhgef5 may be involved in chemoattractant-mediated activation of RhoA and/or RhoB. In leukocytes, G␤␥ subunits have been shown to mediate many of the chemoattractant signaling events. Thus, we examined if G␤␥ regulates Arhgef5. We first performed the SRE.L-luciferase reporter gene assay, which was previously shown to be specifically activated by Rho GTPases (15, 34 -36). As shown in Fig. 2D, coexpression of G␤1␥2 and Arhgef5 led to synergistic activation of the reporter gene activity, suggesting that G␤␥ and Arhgef5 may function in the same signaling pathway. This conclusion is supported by the observation that coexpression of Arhgef5, G␤1␥2, and RhoA led to the highest levels of active RhoA (Fig. 2E).
To further investigate the relationship between G␤␥ and Arhgef5, we examined if G␤␥ can interact with Arhgef5 in an immunoprecipitation assay. We found that G␤␥ and Arhgef5 coimmunoprecipitated in HEK293 cells expressing both proteins (Fig. 3, A and B). We went on to delineate the sequences on Arhgef5 that are required for its interaction with G␤␥. A series of Arhgef5 deletion mutants were generated as depicted in Fig. 3A. These Arhgef5 mutants were tested for their ability to coimmunoprecipitate with G␤1␥2 in HEK293 cells. We found that the sequence encompassing residues Val 281 -Thr 395 is minimal for retaining the full ability to interact with G␤␥ (Fig. 3B). The fact that the sequences encompassing residues Val 281 -Arg 311 and Trp 312 -Glu 424 are also able to bind to G␤␥ (Fig. 3B) support the idea that the G␤␥-binding site is located in and near the PH domain.
To determine whether the interaction between G␤␥ and Arhgef5 is direct, we carried out an in vitro pulldown assay using recombinant Arhgef5 protein prepared from a bacterial expression system and G␤1␥2 protein prepared from a baculoviral expression system. The pulldown assay shows that these two proteins can interact directly (Fig. 3C).
The aforementioned results together suggest that G␤␥ binds to and may activate Arhgef5, which in turn activates RhoA. To demonstrate that G␤␥ can activate RhoA via Arhgef5 in a biochemically defined system, we carried out an in vitro GTP loading assay. In this assay, we tested if recombinant G␤1␥2 and Arhgef5 proteins were able to stimulate the loading of a fluorogenic GTP analog, Mant-GTP (37), to recombinant RhoA, which represents the activation of the small GTPase. We found that, whereas Arhgef5 alone stimulated GTP loading to RhoA as expected, addition of G␤1␥2 resulted in a further increase in GTP loading (Fig. 3D) and that G␤1␥2 regulated this RhoA activity in a dose-dependent manner (Fig. 3E). These results demonstrate that G␤␥ is able to directly activate Arhgef5, which in turn activates RhoA.
Next we investigated the roles that Arhgef5 plays in vivo by generating a mouse line in which the Arhgef5 gene was disrupted (supplemental Fig. S1). Because siRNA-mediated knocking down of Arhgef5 expression led to impaired chemotaxis of a macrophage, we first examined whether Arhgef5 deficiency affected the primary mouse macrophage chemotaxis. The migration of neither peritoneal macrophages in response to C5a was, however, not affected (Fig.  4A). We also examined chemotaxis of spleen T and B lymphocytes in response to SDF-1 and bone marrow neutrophils The RBD pulldown assays were carried out for cells expressing RhoA, RhoB, and RhoC, whereas the PBD pulldown assay was done for Rac, RhoQ, RhoD, and RhoV. The activity of RhoG was determined by a pulldown using GST-Elmo. Both precipitated and total GTPases were detected by Western analysis using an antibody specific for the HA tag carried by these small GTPases. B, interaction of dominant negative GTPases with Arhgef5. HA-tagged RhoA-N 19 (A), RhoC-N 19 (C), RhoD-N 31 (D), RhoF-N 33 (F), and LacZ were cotransfected with FLAG-tagged Arhgef5 in HEK293T cells. Immunoprecipitation (IP) was carried out with an anti-HA antibody and detected with an anti-FLAG antibody. C, knocking down Arhgef5 reduced SDF-1-induced RBD binding. J774 cells were transfected with pAS or Arhgef5 shRNA for 48 h and stimulated with or without 30 ng/ml SDF-1 for 15 s. Cells were fixed, permeabilized, stained with GST-RBD and Alexa 633-labeled secondary antibody, and analyzed by a flow cytometer. MFI of Alexa 633 in cells gated for GFP is shown. Three experiments were performed. A representative one is shown. D, SRE.L-luciferase assay. HEK293T cells were co-transfected with the SRE-luciferase reporter gene, GFP, LacZ, G␤␥, and/or Arhgef5. Cells were lysed and the GFP levels and luciferase activity were determined. The luciferase activity was normalized against the GFP level. E, activation of RhoA by G␤␥ and Ahrgef5 in a cotransfection assay. HEF293T cells were transfected as indicated. Twenty-four hours later, RBD pulldown assays were carried out. The numbers under the top panel are relative band intensity quantified by densitometry and normalized against total RhoA levels. OCTOBER 16, 2009 • VOLUME 284 • NUMBER 42 to formylmethionylleucylphenylalanine. Arhgef5 deficiency did not affect the chemotactic responses of these cells (Fig.  4B, data not shown).

Arhgef5 Regulates RhoA and DC Migration
DCs play important roles in modulating in the adapted immune system, and some of them share the same ontogeny as macrophages (38,39). Therefore, we examined chemotaxis of bone marrow-derived immature DCs in vitro and found that Arhgef5 deficiency markedly impaired immature DC chemotaxis in response to MIP-1␣ (Fig. 4C). We also examined RhoA activation in response to MIP-1␣ and found that MIP-1␣-induced RhoA activation is significantly reduced in the immature DCs from Arhgef5-null mice compared with those from wild type littermates (Fig. 4D). These results indicate that chemokines may regulate immature DC migration via Arhgef5 and RhoA. However, when we examined the effect of Arhgef5 deficiency on migration of mature DCs, we found that the lack of Arhgef5 had little effect (Fig.  4D). We also examined DC migration in vivo using the skin painting assay. There was a partial (35%), but statistically significant reduction in DC migration from the skin to lymph nodes in Arhgef5-null mice compared with the wild type controls (Fig. 4F). Because DCs start to mature upon antigen engagement, the differential effects of Arhgef5 deficiency on immature and mature DCs may explain this less robust effect of Arhgef5 deficiency on DC migration in vivo.
We postulated that the lack of effect of Arhgef5 deficiency on mature DC migration may be due to the molecular redundancy as there are several Arhgef5 close homologues whose expression may be upregulated in mature DCs to functionally compensate the lack of Arhgef5. Therefore, we examined the expression of Arhgef5 and its two homologs Ephexin-1 and Arhgef15 in immature and mature DCs. Although Arhgef5 could be detected in both mature and immature DCs, Ephexin-1 and Arhgef15 could only be detected in mature DCs by RT-PCR (Fig. 4G), suggesting that the relative contribution of Arhgef5 may diminish along with the increases in Ephexin-1 and Arhgef15 expression during DC maturation. To test whether these Arhgef5 homologs are indeed a functional redundancy in regulation of mature DC migration, we attempted knocking down of Arhgef5, Arhgef15, and both in cultured mature DCs with their specific siRNAs. As shown in Fig. 4H, transfection of mouse bone marrow-derived mature DCs with both siRNAs led a significant reduction in mature DC chemotaxis in response to CCL-19, although the transfection of either siRNA alone showed insignificant effects.
We also investigated the in vivo roles of Arhgef5 using two mouse model systems in which DCs are known to be involved: ovalbumin-induced allergic airway inflammation and vaginal infection of HSV-2. After OVA immunization and OVA challenge, we found that Arhgef5 deficiency reduced the number of eosinophils infiltrating into the respiratory tract, CD11c ϩ DC cells migrating to the bronchial lymph node, and the levels of IL-4 in the BAL (Fig. 5, A-C). However, Arhgef5 deficiency did not cause significant changes in the levels of IL-12 or IFN-␥ in vaginal washes. Furthermore, induction and differentiation of CD4 T cells as measured by IFN-␥ production from CD4 ϩ T cells stimulated by heat-inactivated HSV-2 or CD8ϩ T cells by the gB peptide in the HSV-2 infection model were not significantly affected in Arhgef5-deficient mice (Fig. 5, D-F). These results suggest that Arhgef5 deficiency may affect the T h 2 responses in an OVA-induced asthma model, but not T h 1 responses induced following viral infection with HSV-2. In addition, the results shown in Fig. 5F indicate that intrinsic T cell responses are not affected by Arhgef5 deficiency.

DISCUSSION
We screened a mini-Rho-GEF shRNA library and identified Arhgef5 as an important GEF in the regulation of cul- tured myeloid cells. By studying of Arhgef5-null mice, we demonstrated that Arhgef5 is a RhoA regulator in vivo and has an important role in the regulation of immature mouse DC chemotaxis. Together with the finding that Arhgef5 is directly activated by G␤␥, we have characterized a novel signaling pathway from chemokine receptors to RhoA activation for regulation of leukocyte migration. In this signaling pathway, chemokine receptors presumably act through the G␤␥ subunits of the G i proteins to activate Arhgef5, which in turn activates RhoA. RhoA are known to regulate the formation and contractility of the actomyosin structure that provides a pushing force for leukocyte migration.
In this report, we provide strong evidence showing that Arhgef5 is directly regulated by G␤␥. G␤␥ binds to the linker region between the DH and PH domains of Arhgef5 as well as part of the PH domain. Arhgef5 is the first RhoA GEF that has been found to be directly regulated by G␤␥. G␤␥ may regulate Arhgef5 through a mechanism similar to that by which G␣ q regulates p63Rho-GEF revealed by a recent structural study (40). It is also possible that G␤␥ may induce Arhgef5 membrane translocation as it does to a Rac-specific GEF, P-Rex1 (41). Additional studies, particularly structural ones, are needed to further determine the molecular mechanisms by which G␤␥ regulates Arhgef5 activity.
The study of Arhgef5-null mice revealed that Arhgef5 is a significant regulator of RhoA in immature DCs. This observation is consistent with the effect of Arhgef5 deficiency on immature DC migration. Because of the up-regulation of Arhgef5 homologs in mature DCs, as shown in Fig. 4G, Arhgef5 is no A-C, transwell migration assays. Chemotactic activity of peritoneal macrophages (A), spleen T and B lymphocytes (B), and bone marrow-derived immature DCs (C) from wild type (G5ϩ/ϩ) or Arhgef5-deficient (G5Ϫ/Ϫ) mice was determined using the transwell assay. C5a (10 ng/ml, A), SDF-1 (100 ng/ml, B), and MIF1␣ (300 ng/ml, C) were used. D, RhoA activity in immature DCs. MIP-1␣-induced Rho activation was determined using an ELISA kit that determines the levels of active RhoA. p ϭ 0.012. E, transwell migration assay of bone marrow-derived mature DCs in response to CCL19 (60 ng/ml). F, in vivo migration of DCs. Mice were painted with FITC and cells from inguinal lymph nodes were isolated and stained with CD11c. The percentage of DCs migrated from the skins (FITC/CD11c double positive) were determined by flow cytometry. n ϭ 14, p Ͻ 0.05. G, expression of Arhgef5 and its close homologs Arhgef15 and Ephexin-1 detected by RT-PCR in immature DCs (imDC) and mature DCs (mDC). H, knocking down of both Arhgef5 and -15 reduces migration of bone marrow-derived mature DCs in response to CCL-19. Bone marrow-derived mature DCs were transfected with no oligo (C) or synthetic siRNA oligos targeting Arhgef5 (G5), Arhgef15 (G15), or both (G5/ 15). p Ͻ 0.05, fourth bar versus others. P values for other bars are Ͼ0.05. CD4 ϩ or CD8 ϩ T cells were isolated from iliac and inguinal lymph nodes. They were co-cultured with naïve wild type (WT) splenocytes and heat-inactivated HSV-2 of the indicated plaque-forming units (for CD4 ϩ T cells) or 1 g/ml gB peptide (for CD8 ϩ T cells) for 3 days. The levels of IFN-␥ in the conditioned medium were determined by ELISA (F). n ϭ 3 (wild type) or 4 (Arhgef5Ϫ/Ϫ). longer the primary GEF for RhoA regulation, and hence its deficiency no longer has a significant impact on mature DC migration. Immature DCs, once activated by antigens, start to migrate toward lymph nodes accompanied by their maturation process. It is reasonable to speculate that up-regulation of Arhgef5 homolog expression upon DC maturation would mask some of the in vivo effects of Arhgef5 deficiency. This molecular redundancy may help to explain the differential effects of Arhgef5 deficiency on the two model studies shown in Fig. 5. DCs are known to play important roles in both of the model systems, even though these two models are primarily mediated by T h 1 and T h 2 cells, respectively. The lack of significant effects of Arhgef5 deficiency on the responses in the HSV-2 infection model may be due to early maturation of DCs or an involvement of different types of DCs in which Arhgef5 has no primary role. In fact, DCs are rapidly activated through innate recognition of HSV-2 virions through TLR9 (42), and inflammatory monocytes are recruited to the site of infection and there differentiate into DCs (43). Although significant reduction in DC infiltration into the bronchial lymph nodes observed in Arhgef5-null mice should contribute to the reduction in eosinophil infiltration and IL-4 production in the OVA-induced asthma model, our results do not exclude other possible contributions by Arhgef5 deficiency to the attenuation of allergic airway inflammatory responses. Further studies are needed to clarify these questions. In addition, the presence of Ephexin-1 and Arhgef15 in neutrophils and macrophages (data not shown) may explain why Arhgef5 deficiency failed to affect migration of these leukocytes in response to chemoattractants. Inactivation of additional GEFs in this class may be needed to reveal if these GEFs are actually involved in chemotaxis of these leukocytes and to more comprehensively investigate the significance of this signaling pathway in immune responses in vivo.