Regulator of G-Protein Signalling-14 (RGS14) Regulates the Activation of αMβ2 Integrin during Phagocytosis

Integrin-mediated phagocytosis, an important physiological activity undertaken by professional phagocytes, requires bidirectional signalling to/from αMβ2 integrin and involves Rap1 and Rho GTPases. The action of Rap1 and the cytoskeletal protein talin in activating αMβ2 integrins, in a RIAM-independent manner, has been previously shown to be critical during phagocytosis in mammalian phagocytes. However, the events downstream of Rap1 are not clearly understood. Our data demonstrate that one potential Rap1 effector, Regulator of G-Protein Signalling-14 (RGS14), is involved in activating αMβ2. Exogenous expression of RGS14 in COS-7 cells expressing αMβ2 results in increased binding of C3bi-opsonised sheep red blood cells. Consistent with this, knock-down of RGS14 in J774.A1 macrophages results in decreased association with C3bi-opsonised sheep red blood cells. Regulation of αMβ2 function occurs through the R333 residue of the RGS14 Ras/Rap binding domain (RBD) and the F754 residue of β2, residues previously shown to be involved in binding of H-Ras and talin1 head binding prior to αMβ2 activation, respectively. Surprisingly, overexpression of talin2 or RAPL had no effect on αMβ2 regulation. Our results establish for the first time a role for RGS14 in the mechanism of Rap1/talin1 activation of αMβ2 during phagocytosis.


Introduction
Phagocytosis has a critical physiological function as part of the feeding process in amoebae, or as part of the innate immune system which functions to remove microorganisms and apoptotic cells in mammals [1]. In humans the process of phagocytosis is undertaken by professional phagocytes such as neutrophils and macrophages. Phagocytosis involves receptor-mediated particle recognition, actin-driven uptake, phagosome maturation and particle clearance. Surfaceexpressed phagocytic receptors exist that can bind their target directly or indirectly through opsonins, depending on cell type and the nature of its targets [2]. Two well characterized phagocytic receptors are the Fcγ receptor (FcγR) and complement receptor 3 (aka CR3, Mac-1, αMβ2, CD11b/ CD18), that bind IgG-or C3bi-opsonised particles, respectively [1,2].
CR3 is an integrin, comprised of a single αM and a single β2 subunit. Integrin subunits are made up of a large extracellular ligand-binding domain, a single pass transmembrane domain and a short cytoplasmic tail and they are bi-directionally regulated. "Inside-out" activation of αMβ2 involves the cytoskeletal protein, talin, the small GTP-binding protein, Rap1, and calcium/calmodulin kinase II and leads to particle binding [3][4][5][6]. Association with a complement-opsonised particle leads to "outside-in" signalling which ultimately leads to the RhoAdependent uptake, and subsequent destruction, of the particle [7,8].
Previously, we established both that Rap1 action was upstream of talin and that exposure of talin1 head domain was crucial for "inside-out" activation of αMβ2 during phagocytosis in macrophages [4,5]. This signalling pathway occurs independent of RIAM, a well-documented Rap1 effector crucial for αIIbβ3 integrin function [9]. Several other potential Rap1 effectors have been identified, although most of these also bind to other Ras GTPases in vitro -e.g. Vav2, Regulator for cell Adhesion and Polarization enriched in Lymphoid tissues (RAPL), Afadin/MLLT4 (AF6) and Regulator of G-Protein Signalling-14 (RGS14) [10][11][12][13]. Others, like Phosphatidylinositol 3-kinases, RalA and Raf have been dismissed (Caron, unpublished data) [14][15][16]. Recent work has also identified other FERM-containing proteins, such as talin2 and kindlin3, the latter of which has a positive role in integrin activation [17,18]. The aim of this study was to identify the Rap1 effector involved in activation of αMβ2 and to establish the role, if any, of talin2 in this process. We show that the Rap1 effector, RGS14 regulates αMβ2, and that this regulation is dependent on R333 of RGS14 and F754 of β2. However, we find no evidence that Talin2 is involved in this process.

Cell culture and transfection
Cells from the murine macrophages J774.A1 and simian kidney fibroblast COS-7 (American Type Culture Collection numbers TIB-67 and CRL-1651, respectively) were maintained and seeded as previously described [7]. Transfection of COS-7 and J774.A1 cells with plasmid DNA or siRNA were performed using Genejuice (Merck Millipore) or RNAiMAX (Invitrogen), respectively, according to manufacturers' instructions. For gene knockdown of RGS14, J774.A1 cells were transfected with 60pmol ON-TARGET plus siRNA (individual or a pool of four, Dharmacon/Thermo Scientific) and allowed to undergo gene silencing for 5 days. For transfection of J774.A1 cells with plasmid DNA, magnetofection using the magnefect-nano II system (nanoTherics Ltd, Stoke-on-Trent, UK) was employed. Briefly, 1.5μg of plasmid DNA was complexed with 1μl of Neuromag (nanoTherics Ltd) for 15min in serum-free DMEM before adding drop-wise to 10 000 non-activated J774.A1 cells per well in 24-well plate. Plates were incubated for 30min over a magnet array moving laterally at 2Hz with 0.2mm amplitude of displacement (oscillating). After transfection, plates were removed from the magnetic arrays and placed back in the incubator. 24hr later, the same procedure was repeated (retransfection) and cells were incubated for a further 24hr before phagocytic challenge or lysed with sample buffer for western blotting. Intensities of bands were determined by densitometric analysis by using the ImageJ software (National Institutes of Health) and related to the levels of Scrambled siRNA/FAK loading control.

Immunofluorescence and scoring
Since only associated RBC were counted, J774.A1 or COS-7 cells were permeabilised with 0.1% Triton X-100 and stained for either β2 (1:100 dilution), myc-or Flag-(both at 1:1000 dilution) tagged proteins before counterstained using Alexa-Fluor-488-conjugated goat anti-mouse (GFP-tagged proteins remained unstained). To highlight the C3bi-SRBC, the Alexa-Fluor-594-conjugated goat anti-rabbit antibodies were applied, which would be against the opsonising rabbit antibodies. Coverslips were finally mounted in Mowiol (Calbiochem) containing p-phenylenediamine (Sigma) as antifading reagent and analyzed by confocal or epifluorescence microscopy (LSM710, Zeiss or DM-IL, Leica). Association index is defined as the number of RBC bound to 100 macrophages or COS-7 cells expressing appropriate plasmids. Data shown are the mean ± SEM of at least three independent experiments.

Statistical analysis
One-way analysis of variance (ANOVA) was used to compare and determine statistical significance (GraphPad InStat software). P values of less than 0.05 were considered significant.

Results and Discussion
To identify the Rap1 effector required for activation of αMβ2, plasmids expressing known Rap1 effectors were transiently transfected into COS-7 cells together with wild type (wt) αM and β2 chains. After 48hr, the transfected cells were challenged with C3bi-opsonised sheep red blood cells (C3bi-RBC), immunostained and scored for associated C3bi-RBCs relative to COS-7 expressing only wt αMβ2 (Figure 1). Consistent with our previous findings, binding of C3bi-RBC increased by twofold to cells expressing αMβ2 together with either constitutively active Rap1 (V12) or full length talin1 [5]. Co-expression with Vav2, AF-6, talin2 or RAPL resulted in no increase in C3bi-RBC binding. However, co-expression of RGS14 with αMβ2 resulted in an increase in binding (158.5±32.7, p<0.05), though not to levels seen with either V12Rap1 (199.2±38.7, p<0.05) or talin1 (198.8±28.2, p<0.05). RGS14 was first identified as a Rap1 effector in a yeast two hybrid screen, using a mouse brain cDNA library, together with wt and point mutants of Rap1 and Rap2. This was confirmed using His-tagged RGS14 to pull down endogenous Rap1 from mouse brain membranes [13]. Therefore it is likely that RGS14 is the primary Rap1 effector responsible for activating αMβ2.
To confirm this result, varying doses (20-400pmol) of short interfering RNAs (siRNA) against human RGS14 or a control scrambled sequence were transfected into differentiated THP-1 human monocytes/macrophages. Transfected cells were left for 48hr before either challenging these cells with C3bi-RBC or lysing the cells for western blot analysis to determine levels of gene knock-down, with tubulin used as a loading control. However, we were unable to efficiently knockdown RGS14 in THP-1 cells (data not shown) so we used J774.A1 mouse macrophages transfected with 60pmol of either an individual or a pool of 4 siRNAs against mouse RGS14 siRNA. Gene silencing of RGS14 using individual siRNAs vary between 0-25% knock-down (#5, #6 and #8, Figure 2B), as determined by western blot analysis (Figure 2A, B). Knockdown of RGS14 using pooled siRNA was not complete with ~55% decrease in expression of RGS14 -when compared to the corresponding scrambled siRNA controls and focal adhesion kinase (FAK) loading controls (Figure 2A,B). This is reflected in a physiological effect as demonstrated with a decrease in bound C3bi-RBC -set at 100 (60pmol scrambled siRNA) c.f. 46.5±6.1 (pooled RGS14 siRNA) ( Figure 2C). While there is a possibility of off-target effects, these findings and previous results with the overexpressed RGS14 gene ( Figure 1) suggest a role for RGS14 in activating αMβ2 prior to particle binding and uptake. The modest decrease in bound C3bi-SRBC could be due to the presence of talin, a well documented activator of integrins, including αMβ2, on its β2 cytoplasmic tail [4,24]; kindlin-3, another activator of β2 class of integrins and shown to be involved in αMβ2-mediated spreading of αMβ2-expressing K562 cells [25,26]; and/or Rap1-interacting adaptor molecule (RIAM) which was recently demonstrated to regulate complement dependent phagocytosis, an αMβ2-mediated event [27], though the role of RIAM is still controversial [5]. Nevertheless, this suggests that RGS14 acts on αMβ2, either directly or in conjunction with the other integrin regulators described above. Interestingly, depletion of RGS14 led to a decrease in nerve-and fibroblast-growth factor induced neurite outgrowth in the neuronal cell line, PC12 [22] -a process generally thought to be driven by α3β1 integrins, specifically the cytoplasmic tail of the α3 integrin subunit [28,29].
Finally, to confirm this RGS14/αMβ2 signalling mechanism, J774.A1 murine macrophages were transfected with either GFP as control, GFP-tagged talin1, talin2, or Flag-tagged RGS14. Although the transfection efficiency was low using magnetofection (<1%), cells remained viable and there were sufficient transfected cells (minimum of 25 cells) to allow analysis of phagocytosis ( Figure 3). Consistent with previous data (Figure 1), J774.A1 cells transfected with RGS14 showed an increase in C3bi-RBC binding (170.0±10.7) compared to GFP control (set at 100, p<0.05), though not the same levels of binding as seen in the talin1-expressing control cells (207.5±18.9, p<0.001). Expression of talin2 did not increase binding of C3bi-RBC (73.3±18.1), further confirming that talin2 was not involved in activation of αMβ2. Studies showed that, unlike talin1, talin2 did not affect embryonic development, focal adhesion assembly or cell spreading [30,31]. However, talin2 is able to compensate the loss of talin1 in fibroblasts derived from talin1 knockout mice [32] and is possibly involved in the formation of actin-integrin links in mature muscle cells [33]. Interestingly, mouse testis and kidney express a talin head domain (N-terminal)-deficient isoform of talin2 [17] and it would be interesting to see if such a form of talin2 exists in macrophages. While we recognise the controversy surrounding RGS14 being Rap1 effector and interacting partner [22] and the lack of clarity regarding the function of talin2 in cellular models, we believe this study is a first step in delineating the pathway that governs αMβ2 activity. Interestingly, it has been shown recently that the Ras-binding domain (RBD) of RGS14 regulates the activity of RGS domains by interacting with, and enhancing their GTPase (GAP) activity [34]. This suggests the possibility of RGS14 acting in conjunction with Rap1. It also has to be noted that R-Ras, a Rap1 homologous small GTPase, like Rap1, also shares RGS14 as an effector [22]. R-Ras was previously shown to be required for αMβ2 activity during binding and phagocytosis of C3bi opsonised SRBC [3,35]. Therefore there is still a need to distinguish the role of Rap1 on RGS14 from R-Ras and how that regulates αMβ2 function, which would hopefully open up the field of integrinand macrophage-biology. John Hepler (Emory University, GA, USA) for the kind gifts of plasmid constructs, as described in this manuscript. We dedicate this manuscript to the memory of Emmanuelle Caron who passed away on 8 July 2009 before this work was completed.