Changes in membrane phosphoinositides and local Ca2+ elevations at sites of particle capture coordinate the dynamic remodeling of the actin cytoskeleton during phagocytosis. Here, we show that the phosphatidylinositol (PI) transfer proteins PITPNM1 (Nir2) and PITPNM2 (Nir3) maintain phosphatidylinositol 4,5-bisphosphate [PI(4,5)P2] homeostasis at phagocytic cups, thereby promoting actin contractility and the sealing of phagosomes. Nir3 and to a lesser extent Nir2 accumulated on endoplasmic reticulum (ER) cisternae juxtaposed to phagocytic cups when expressed in phagocytic COS-7 cells. CRISPR-Cas9 editing of Nir2 and Nir3 genes decreased plasma membrane PI(4,5)P2 levels, store-operated Ca2+ entry (SOCE) and receptor-mediated phagocytosis, stalling particle capture at the cup stage. Re-expression of either Nir2 or Nir3 restored phagocytosis, but not SOCE, proportionally to the PM PI(4,5)P2 levels. Phagosomes forming in Nir2 and Nir3 (Nir2/3) double-knockout cells had decreased overall PI(4,5)P2 levels but normal periphagosomal Ca2+ signals. Nir2/3 depletion reduced the density of contractile actin rings at sites of particle capture, causing repetitive low-intensity contractile events indicative of abortive phagosome closure. We conclude that Nir proteins maintain phosphoinositide homeostasis at phagocytic cups, thereby sustaining the signals that initiate the remodeling of the actin cytoskeleton during phagocytosis.
Phagocytosis, the engulfment of large particles by cells, is an evolutionarily conserved cellular process required to eliminate invading pathogens and to maintain tissue homeostasis (Flannagan et al., 2012). Particle recognition is mediated by surface receptors for immunoglobulins or complement fragments coating invading pathogens and by receptors for pathogen-associated sugars (Uribe-Querol and Rosales, 2020). The engagement of phagocytic receptors initiates a dramatic remodeling of the plasma membrane accompanied by acute changes in phosphoinositide (PI) composition at sites of particle capture. Sequential fluctuations in the local concentration of phosphatidylinositol mono, bis and tris phosphate regulate distinct trafficking and signaling events during the formation of phagocytic vacuoles and their subsequent fusion with endolysosomes (Botelho et al., 2000; Levin-Konigsberg et al., 2019; Montaño-Rendón et al., 2022, reviewed in Bohdanowicz and Grinstein, 2013; Flannagan et al., 2012). A spatially restricted transient PI(4,5)P2 elevation occurs at phagocytic cups coinciding with the transient recruitment of the 5-kinase that converts phosphatidylinositol 4-phosphate [PI(4)P] into phosphatidylinositol 4,5-bisphosphate [PI(4,5)P2] (Botelho et al., 2000). The PI(4,5)P2 elevation promotes actin polymerization at phagocytic cups by activating actin nucleators while inhibiting actin severing proteins (Bohdanowicz and Grinstein, 2013; Yeung et al., 2006) As the phagosome seals, PI(4,5)P2 is phosphorylated by the phosphatidylinositol 3-kinase (PI3K) into phosphatidylinositol (3,4,5)-trisphosphate [PI(3,4,5)P3], which in turn becomes transiently enriched at phagocytic cups (Marshall et al., 2001) before being converted into PI(3,4)P2 by phosphoinositide 5-phosphatases (Montaño-Rendón et al., 2022).
Following receptor stimulation, PI(4,5)P2 is also converted by phospholipase Cγ (PLCγ) into diacylglycerol (DAG) and inositol trisphosphate (InsP3). DAG is further converted into phosphatidic acid (PA) by phosphorylation at the plasma membrane (PM) (Balla, 2013; Cockcroft and Garner, 2011), whereas InsP3 promotes the release of Ca2+ from endoplasmic reticulum (ER) stores. Store depletion activates the Ca2+-sensing protein STIM1 that traps and gate the ORAI family of Ca2+ channels at ER–PM and ER–phagosomes membrane contact sites (MCS) further stabilized by junctate (Guido et al., 2015; Nunes et al., 2012; Westman et al., 2019). The combined activity of PLCγ and PI3K depletes PI(4,5)P2, promoting actin disassembly at the base of phagocytic cups (Scott et al., 2005) while periphagosomal Ca2+ elevations drive the activity of Ca2+-dependent actin-severing proteins around phagocytic vacuoles (Nunes et al., 2012). This signaling cascade is required for successful target internalization, and decreasing PI(4,5)P2 levels or preventing periphagosomal Ca2+ elevations impair phagocytosis (Botelho et al., 2000; Coppolino et al., 2002; Scott et al., 2005; Nunes et al., 2012). The signals initiated at cups implies that sufficient levels of lipid precursors are delivered at sites of forming phagosomes to fuel the sequential changes in signaling lipids, but the mechanism(s) ensuring the supply of PIs at sites of particle capture are unknown.
Nir2 (PITPNM1) and Nir3 (PITPNM2) are mammalian homologues of Drosophila retinal degeneration protein B (rdgB), and harbor an N-terminal PI transfer domain (PITP) that drives the non-vesicular exchange of ER-bound PI for phosphatidic acid (PA) between the ER and target membranes (reviewed in Balla, 2018). Nir2 and Nir3 are recruited to ER–PM contact sites upon receptor stimulation via interactions with the FFAT motif of the ER-resident vesicle-associated membrane-associated proteins VAP-A and VAP-B (Amarilio et al., 2005; Selitrennik and Lev, 2016). Nir2 maintains PI(4,5)P2 levels at the PM during signaling of Gq-coupled receptors by exchanging PI from the ER for PA on the PM (Chang et al., 2013; Chang and Liou, 2015; Kim et al., 2013, 2015), thereby preserving Ca2+ signaling competence when PIs are rapidly consumed by PLC. This homeostatic function is clinically relevant as increased Nir2 expression correlates with enhanced epithelial–mesenchymal transition and poor patient prognosis (Keinan et al., 2014). During stimulation of PLC-coupled receptors, Nir2 is dynamically recruited to ER–PM contact sites by the PA generated from DAG in the PM (Kim et al., 2013, 2015). The concomitant Ca2+-dependent recruitment of extended synaptotagmin-1 (E-Syt1) stabilizes ring-shaped ER–PM contact sites at a reduced gap distance (Chang and Liou, 2016; Kang et al., 2019). Whether Nir proteins mediate lipid transfer at phagosomes is not known, but another lipid transfer protein, ORP1L (also known as OSBPL1A), interacts with VAP proteins at ER–phagosome contact sites and contributes to phagolysosome resolution by transferring the PI(4)P accumulating in late phagosomes to the ER (Levin-Konigsberg et al., 2019).
Given the importance of PI signals as a crucial regulator of phagocytosis, we investigated the contribution of Nir-mediated non-vesicular lipid transfer in phagosome formation and maturation. Using correlative light-electron microscopy we show that Nir3 is recruited to membrane contact sites at phagocytic cups. Using GFP fused to the plekstrin homology (PH) domains of PLC (Várnai et al., 1999) or LifeAct–mCherry, we quantify PI(4,5)P2 levels and actin dynamics during FcR-mediated uptake of solid particles by cells lacking or re-expressing Nir2 and Nir3 (Nir2/3) proteins. Depletion of Nir2/3 decreased periphagosomal PI(4,5)P2 and F-actin accumulation around forming phagosomes, stalling phagocytosis at cup stage, without perturbing the subsequent PI(4,5)P2 and Ca2+ elevations around successfully internalized particles. This suggests that lipid transfer occurs at cups and maintains PI(4,5)P2 levels required for the formation of contractile actin rings during particle capture.
Nir2 and Nir3 localize to phagocytic cups enriched in PA
The two proteins Nir2 and Nir3 transfer PI from the ER to the PM to replenish PI(4,5)P2 levels following receptor-induced hydrolysis (Chang and Liou, 2015), but whether lipid transport by these proteins contributes to the phagocytic process is unknown. To assess whether Nir proteins are recruited to phagocytic vacuoles, we expressed EGFP-tagged Nir2 or Nir3 in COS7 cells rendered phagocytic by expression of the immunoglobulin receptor FcγRIIA–c-Myc (Guido et al., 2015) and assessed the location of the fluorescent proteins by confocal imaging. EGFP–Nir2 and to a larger extent EGFP–Nir3 accumulated at sites of particle capture, forming ring structures surrounding internalized particles, indicating that the lipid transfer proteins are recruited to phagocytic cups (Fig. 1A). Given that Nir proteins interact with ER-bound FFAT-binding proteins, we tested wither the latter are present at cups by co-expressing EGFP–Nir3 with mCherry–VAP-B, which colocalizes with ORP1L on late phagosomes (Levin-Konigsberg et al., 2019). EGFP–Nir3 colocalized extensively with co-expressed mCherry–VAP-B at cups (Fig. 1B), suggesting that the two proteins are co-recruited to membrane contact sites at cup stage. Nir proteins are recruited to membranes enriched in PA. We thus assessed whether PA was enriched at phagosomes using GFP fused to the membrane-binding motif of the yeast SNARE protein Spo20. Spo20–GFP accumulated around phagosomes and colocalized extensively with RFP–Nir3 (Fig. S1A), the fluorescence of the two reporters increased with similar kinetics following particle capture (Fig. S1B), indicating that Nir3 accumulates around phagosomes enriched in PA. To establish that Nir3 populates ER–phagosome contact sites, we then performed correlative light-electron microscopy (CLEM). Phagocytic COS7 cells were transfected with EGFP–Nir3 and allowed to ingest opsonized beads for 30 min. Cells were then fixed and imaged by confocal microscopy to locate Nir3-decorated phagosomes (Fig. 1C, left panel). EM tomograms of the labeled phagosomes were then acquired at nanometric resolution. Alignment of the confocal and EM images revealed that EGFP–Nir3 signals coincided with ER membranes around phagocytic cups (Fig. 1C, right). These data establish that Nir3 is recruited together with ER-bound proteins to membrane contact sites at phagocytic cups.
Nir2/3 depletion decreases plasma membrane PI(4,5)P2 levels and SOCE
To assess the role of PI transfer proteins during phagocytosis, we used CRISPR-Cas9 editing to disrupt the Pitpnm1 and Pitpnm2 genes coding for Nir2 and Nir3 in mouse embryonic fibroblasts (MEFs). Proper editing was validated by the failure of primers targeting the modified genomic regions to amplify a PCR product (Fig. S2A). Real-time quantitative (q)PCR with primers targeting non-edited regions indicated that Nir2 and Nir3 mRNA levels were decreased by 80% and 60%, respectively (in figures, Nir2CR, Nir3CR and Nir2/3CR refer to single edit for Nir2, single edit for Nir3, and double edit for Nir2 and Nir3, respectively) (Fig. S2B,C) with similar melting points, indicating a decreased stability of the modified mRNAs. We then measured basal PM PI(4,5)P2 levels with PH-PLCδ1–GFP (Balla and Várnai, 2002) in control and edited cells to verify the functional impact of these genetic manipulations. An intense PH-PLCδ1–GFP signal was detected at the edge of control cells, whereas a weaker and discontinuous signal delineated cells bearing each of the singly or doubly edited gene (Fig. 2A). Quantification of the GFP signal at the cell edge and cytosol revealed that the PM-to-cytosol intensity ratio was significantly reduced in Nir2- and Nir2/3-edited cell lines (Fig. 2B), indicating that depletion of Nir isoforms decreases basal PI(4,5)P2 levels. We then tested whether SOCE, which is mediated by the recruitment of STIM proteins to PI(4,5)P2-rich domains maintained by Nir2 at ER–PM contact sites (Chang and Liou, 2015), was impacted by Nir2/3 knockdown. As expected, the amplitude of the Ca2+ elevations evoked by the readmission of Ca2+ to cells treated with the SERCA inhibitor thapsigargin was severely impacted in Nir2/3-edited cells (Fig. 2C). Stable expression of RFP-tagged Nir2 or Nir3 restored basal PM PI(4,5)P2 levels in Nir2/3-edited cells (Fig. 2D,E), suggesting that the two lipid transport proteins complement each other for the maintenance of basal PM PI levels. Interestingly, SOCE was not restored to wild-type (WT) levels by exogenous overexpression of the tagged Nir proteins (Fig. S2D). These data indicate that depletion of the PI transfer proteins Nir2 and Nir3 decreases basal PM PI(4,5)P2 levels and SOCE in mouse fibroblasts and validate Nir2/3-edited cells as a useful tool to study the role of these transfer proteins.
Nir2/3 depletion decreases phagocytosis
PIs are central regulators of the trafficking and signaling events driving phagocytosis. To test whether the depletion of Nir proteins impacts the phagocytic process, we quantified the uptake of opsonized polystyrene beads by MEF cells rendered phagocytic by transient FcγRIIA–c-Myc expression (Guido et al., 2015). Phagocytic uptake was quantified as the percentage of cells exhibiting one or more associated fluorescent particles on flow cytometry scatter plots (Fig. 3A). Particle uptake was significantly reduced in Nir2/3-edited cells incubated for 30 and 60 min with opsonized particles (Fig. 3A,B), as was the phagocytic index measured by fluorescence imaging (Fig. 3C). Stable re-expression of RFP-tagged Nir2 or Nir3 partially restored phagocytic uptake in Nir2/3-edited cells (Fig. 3D). Of note, the PM PI(4,5)P2 levels measured with PH-PLCδ1–GFP in the different cell lines correlated well with the proportion of phagocytosing cells (Fig. 3E). These data indicate that Nir2/3 depletion reduces the efficiency of phagocytosis and that this defect is proportional to the reduction in PI(4,5)P2 levels caused by the loss of these lipid transport proteins.
Nir2/3 depletion reduces phagosomal PI(4)P and PI(4,5)P2 but not periphagosomal Ca2+ signals
Biphasic changes in PI(4,5)P2 levels occur during phagocytosis, with a rapid elevation at sites of particle engagement followed by a decrease as phagosomes seal (Botelho et al., 2000). To establish the role of Nir proteins in phagosomal PI(4,5)P2 dynamics, we imaged local PI(4,5)P2 levels around forming and internalized phagosomes using PH-PLCδ1–GFP. A transient PI(4,5)P2 enrichment was observed around nascent phagosomes 30–90 s after particle capture (Fig. 4A), both in control and Nir2/3-edited cells. However, quantification of the phagosomal PI(4,5)P2 revealed that whereas the transient elevation persisted in Nir2/3 edited cells, the local levels of the PI were dampened throughout the phagocytic process (Fig. 4B). To test whether Nir-mediated PI delivery also affects PI(4)P, we measured the accumulation of PI(4)P around phagosomes using GFP fused to the P4M domain of SidM (GFP–P4M). The PI(4)P probe transiently accumulated around phagosomes after particle capture as observed with the PI(4,5)P2 probe (Fig. S3A). In this case, however, Nir2/3 depletion reduced the transient increase in GFP–P4M around forming phagosomes but not the steady-state levels of the lipid probe on internalized phagosomes (Fig. S3B). These data indicate that Nir2/3 proteins help sustain high levels of PIs on the membrane of phagosomes and differentially impacts PI(4)P and PI(4,5)P2 dynamics, facilitating acute changes in PI(4)P and sustaining high PI(4,5)P2 levels throughout phagocytosis.
STIM1 is recruited to phagosomes, generating local Ca2+ elevations that enhance the efficiency of phacocytosis (Nunes et al., 2012; Nunes-Hasler et al., 2020; Westman et al., 2019). STIM1 binds PI(4,5)P2 via its C-terminal tail and colocalizes with Nir2 at ER-PM contact sites (Kim et al., 2015), prompting us to measure local Ca2+ elevations around phagosomes by confocal imaging using Fluo-8 and a low concentration of the Ca2+ chelator BAPTA-AM to reduce lateral Ca2+ diffusion. Unexpectedly, Ca2+ hotspots were detected around 34-40% of (successfully ingested) phagosomes independently of Nir2/3 depletion (Fig. S4), indicating that Nir proteins are dispensable for periphagosomal Ca2+ signals.
Nir2/3 depletion stalls the phagocytic process at the cup stage
We next assessed whether early steps of particle capture were impacted by the loss of Nir2/3 proteins. Cells were allowed to phagocytose opsonized fluorescent beads and the amount of total and surface-bound particles assessed using the bead-associated (green) and solvent-accessible anti-IgG (red) fluorescence signals, respectively (Fig. 4C,D; Fig. S5). The proportion of exposed particles, reflecting beads stuck at cup stage, was significantly increased in Nir2/3-edited cells (Fig. 4E, left). Re-expression of RFP–Nir2 or RFP–Nir3 reduced the proportion of particles stuck at cup stage, linking the defect to reduced Nir protein expression (Fig. 4E, right). Electron microscopy confirmed that most particles were intracellular in control cells and present at cups in Nir2/3-edited cells (Fig. 4F). These results indicate that Nir proteins facilitate the closure of phagocytic cups.
Nir2/3 depletion impairs the formation of contractile actin rings during phagocytosis
Given that PI composition controls actin dynamics during pseudopod formation and phagosome sealing, we next quantified the changes in the density of the cortical actin cytoskeleton forming around captured particles with LifeAct–mCherry. As previously reported, a transient actin enrichment was observed at sites of particle capture immediately after contact, the F-actin probe accumulating into ring structures that rapidly subsided as the particles were internalized (Fig. 5A). Remarkably, multiple rings of actin were observed arising repeatedly for up to 15 min at sites of particle capture in Nir2/3-edited cells (Fig. 5A,B; Movie 1). Repetitive ring formation was observed in 32% versus 4% of the phagocytic events analyzed in Nir2/3-edited and control cells, respectively, a highly significant difference (Fig. 5B, bottom). Moreover, the thickness of the circumferential LifeAct–mCherry signal was reduced in Nir2/3-edited cells (Fig. 5C,D). These data indicate that Nir proteins facilitates the formation of contractile actin rings during phagocytic uptake.
PI metabolism regulates the efficiency of the phagocytic process initiated at particle capture and terminating at phagolysosome resolution (Botelho and Grinstein, 2011). The formation of phagocytic cups is associated with a transient local increase in PI(4,5)P2 that is quickly replaced by an equally transient elevation in PI(3,4,5)P3 at the site of contact, a biphasic change required for the remodeling of the actin cytoskeleton driving the extension of cups and the sealing of phagosomes. Both lipids promptly disappear from sealed phagosomes through the action of phosphatases to be replaced by PI(3,4)P2 and PI(3)P as phagosomes mature (Dewitt et al., 2006; Montaño-Rendón et al., 2022). ORP1L-mediated transfer of PI(4)P to the ER contributes to the resolution of phagosomes by promoting their tubulation (Levin-Konigsberg et al., 2019), but whether a lipid transfer protein is required for the supply of PIs at phagocytic cups is unclear as PI or PI(4)P can be delivered to cups by recycling endosomes (Levin-Konigsberg and Grinstein, 2020). Here, we show that the PI transfer proteins Nir2 and Nir3 participate in the lipid signaling events triggered by the engagement of phagocytic receptors by maintaining PI(4,5)P2 levels at sites of particle capture that enable the contractile activity of the actin cytoskeleton during phagosome sealing.
Using gene editing, we show that combined depletion of Nir2 and Nir3 decreases basal PM PI(4,5)P2 levels and the efficiency of phagocytosis and that these two defects are corrected by re-expression of each of the two isoforms independently. This indicates that PM PI(4,5)P2 levels limit phagocytosis efficiency, consistent with earlier studies showing that the sequential phosphorylation of PI(4)P to PI(4,5)P2 and PI(3,4,5)P3 is required for phagosome formation (Montaño-Rendón et al., 2022). Interestingly, Nir2/3 editing reduced but did not completely eliminate dynamic elevations of PI(4)P and PI(4,5)P2 during particle engulfment, minimizing acute increases in PI(4)P around forming phagosomes and decreasing the overall PI(4,5)P2 levels throughout phagocytosis. This suggests that Nir proteins do not initiate but instead sustain the rapid amplification of the PI signal generated by the activity of kinases. Therefore, once phagosomes are sealed, dynamic changes in PIs can occur within phagocytic membranes without Nir-mediated lipid exchange. In Nir2/3-edited cells, phagocytosis was halted at cup stage, indicating a reliance on Nir proteins for the sealing of phagosomes. Interestingly, Nir3 preferentially accumulated near phagocytic cups enriched in PA together with the ER-bound protein VAP-B, decorating ER cisternae vicinal to sealing phagosomes on the electron microscope. This suggests that Nir3 mediates the exchange of PA for PI mainly at phagocytic cups, a differential accumulation that might reflect the increased affinity of Nir3 for negatively charged lipids (Chang and Liou, 2015). However, each of the two lipid transfer proteins could restore phagocytic uptake when re-expressed, indicating that focal PI supply is not a strict requirement for cup closure, at least not in the context of copious global PM supply. We propose that both isoforms regulate phagocytic activity by maintaining PI signaling at sites of phagocytosis, with Nir2 acting predominantly at the PM and Nir3 at cups.
We previously showed that localized Ca2+ elevations can persist for tens of minutes around internalized phagosomes in primary mouse neutrophils and dendritic cells as well as in MEFs expressing FcgRIIa (Nunes et al., 2012). The long-lasting local Ca2+ elevations enhance the efficiency of phagocytosis and are fueled by Ca2+ release from InsP3 receptor (InsP3R) on juxtaphagosomal ER cisternae and by Ca2+ influx across STIM-gated Orai1 channels on phagosomes (Guido et al., 2015; Nunes et al., 2012). The persistent Ca2+ activity implies a constant local production of InsP3 around phagosomes to maintain the juxtaposed ER stores in a depleted condition enabling STIM1 to trap and gate phagosomal Orai1 channels. We therefore expected that Nir2/3 depletion would abort the local Ca2+ elevations by preventing the resupply of the PI consumed by signaling at ER–phagosome MCS. Contrary to our expectations, the Ca2+ signals persisted in cells depleted of Nir proteins. This indicates that Nir-mediated exchange of PA for PI is dispensable for the local generation of InsP3 and for the anchoring of STIM1 molecules to phagosomes once the particle has been engulfed. STIM1 binds to a range of phosphatidylinositol phosphates (PIPs) via its polybasic domain and might remain bound to negatively charged lipids present on phagosomes until the resolution stage. Alternatively, STIM1 might be stabilized by junctate (Guido et al., 2015) or by binding to Orai1 channels on phagosomes. The constant generation of IP3 is more difficult to explain as the activity of PLC would stop without a constant supply of PI to regenerate its substrate PI(4,5)P2 in the membrane of phagocytic vacuoles. Other lipid transport proteins might substitute for Nir at the ER-phagosomes interface. Alternatively, the residual levels of PI(4,5)P2 might be sufficient for a small but sustained local production of IP3 enabling the opening of InsP3R clusters with a high sensitivity to inositol trisphosphate (IP3) immobilized at ER–phagosome MCS (Babu Thillaiappan et al., 2017). A pool of immobile IP3Rs licensed to respond to physiological stimuli localizes near STIM-ORAI interaction sites at ER–PM junctions (Taylor and Machaca, 2019). Whether this mechanism sustains prophagocytic Ca2+ elevations at ER–phagosome MCS remains to be established.
A striking phenotype of Nir2/3-edited cells was the abnormal pattern of actin condensation occurring around forming phagosomes. The thickness of the dense F-actin rings surrounding sites of particle capture was significantly below the levels observed in control cells, and multiple occurrences of ring formation were observed in apparent unsuccessful attempts of phagosome closure. This phenomenon was not reported previously and likely reflects the deficiency in local non-vesicular lipid transport leading to reduced PI(4,5)P2 levels at cups. A substantial amount of evidence indicates that PI(4,5)P2 accumulation at cups recruits and activates actin-modulating proteins to promote F-actin accumulation and contractile activity at cups (Raucher et al., 2000; Yeung and Grinstein, 2007). Reduced F-actin amounts at phagosomal cups has been reported in macrophages expressing a kinase-dead mutant of phosphatidylinositol-4-phosphate 5-kinase (PI4P5K), the enzyme that generates PI(4,5)P2 from PI(4)P at cups (Coppolino et al., 2002). PI4P5K-edited macrophages exhibited normal particle binding and receptor clustering but reduced accumulation of PI(4,5)P2 at cups and phagocytosis was blocked at cup stage, as we report here for Nir2/3 deficiency. In contrast, manipulations that enhance PI(4,5)P2 levels at cups cause persistent actin accumulation (Scott et al., 2005). The loss of Nir proteins therefore likely halts phagocytosis at the cup stage because the decreased local PI(4,5)P2 levels prevent the accumulation of cortical actin at cups, decreasing the contractile activity required for particle internalization.
In conclusion, we show that phosphatidylinositol transfer mediated by Nir2/3 proteins is required for the engulfment of phagocytic targets. Genetic disruption in the genes encoding for these proteins reduces PI(4,5)P2 levels in phagosomal membranes and prevents the formation of contractile actin rings around captured particles, decreasing PM and phagosomal PI(4,5)P2 levels and aborting phagocytosis at cup stage without impacting Ca2+ elevations during phagosome maturation. Nir2 preferentially mediates lipid exchange at the plasma membrane and Nir3 at phagocytic cups but either isoform can independently complement the phagocytic defect caused by the combined depletion of both proteins. We conclude that PI(4,5)P2 levels at cup stage are limiting for the recruitment of the dense cortical actin cytoskeleton driving phagosome sealing.
MATERIALS AND METHODS
Antibodies and reagents
Mouse embryonic fibroblasts (MEFs) were kindly provided by Luca Scorrano (Padova, Italy) and regularly tested for mycoplasma contamination. Myc-tagged human FcγRIIA was a gift from Sergio Grinstein (Toronto, Canada), EGFP-Nir2 and EGFP-Nir3 a gift from Tamas Balla (NIH, USA) and LifeAct–mCherry a gift from Florence Niedengang (Institut Cochin, Paris). PH-PLCδ1–GFP (#51407) and mCherry–VAP-B (#108126) were purchased from Addgene, pTagRFP-C (#FP141) from evrogen, and fura- 2-AM, BAPTA-AM and Lipofectamine 2000 from Life Technologies. TagRFP-Nir2 and TagRFP-Nir3 were subcloned from EGFP–Nir2 and EGFP–Nir3 into pTagRFP-C (evrogen, FP141), respectively. Alexa Fluor 633-coupled goat anti-human IgG (#A-21091) was purchased from Thermo Fisher Scientific, rat anti-mouse CD16/32 (Mouse BD Fc Block, #553142) from BD Biosciences and anti-human CD32–APC antibody from Miltenyi Biotec. Ca2+ recordings and live imaging experiments were conducted in physiological buffer containing 140 mM NaCl, 5 mM KCl, 1 nM MgCl2, 2 mM CaCl2, 20 mM HEPES, 10 mM glucose, made to pH 7.4 with NaOH.
Cell lines, cell culture, transfection and transduction
MEF cells were cultured in Dulbecco's modified Eagle's medium (DMEM; high glucose, catalog number 31966, Life Technologies) containing 10% fetal calf serum (FCS) and 0.5% penicillin-streptomycin (pen-strep, catalog number 15140, Life Technologies) at 37°C under 5% CO2 and were passaged twice a week. CRISPR MEF cells were generated using Double Nickase plasmid (Santa Cruz Biotechnology) against mouse Nir2 and Nir3 genes. Cells were transfected using Lipofectamine 2000 at 40–50% confluence in high glucose (4.5 g/l) DMEM without serum, selected with puromycin and sorted on GFP fluorescence. Sorted single-cell clones were screened for Nir2 and Nir3 depletion by genomic PCR. Absence of the product amplified using the primer binding to the expected cut site suggested a modification of the sequence which was validated by sequencing of the region including the cut site. qPCR was also performed to validate the decrease in mRNA levels. TagRFP-C1, TagRFP-Nir2 and TagRFP-Nir3 were subcloned into pLenti-CMVie-IRES-BlastR (Addgene #119863) and constructs were co-transfected with pMD2G and psPAX2 (Addgene #12259 and #12260) into HEK-293T cells to produce viral particles as described previously (Carreras-Sureda et al., 2021). After transduction in the presence of polybrene at 10 µg/ml, stable cells were obtained by antibiotic selection (blasticidin, 5 µg/ml) based on TagRFP fluorescence by FACS sorting.
Phagocytic target preparation
Carboxyl polystyrene microspheres (3.0 µm, Spherotech, CP-30-10) were opsonized by covalent coupling with hIgG. Following three washes in sterile PBS at 10,000 g at 4°C for 3 min, polystyrene beads were activated by 50 mM 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (Carl Roth, 2156.1) in PBS for 15 min at room temperature by shaking. Next, beads were washed three times in 0.1 M Na2B4O7 buffer (pH 8.0) at 4°C and 6 mg of hIgG were added to beads and incubated at 4°C overnight on a shaker. To prepare fluorescently labeled hIgG polystyrene beads, Alpha Fluor 488 amine (20 μg/ml, AAT Bioquest, cat no. 1705) was added at room temperature for 30 min with agitation before overnight incubation with hIgG. Beads were washed twice with 250 mM glycine in PBS and then with PBS alone, resuspended in sterile PBS and 0.5% pen-strep and counted with a cell counter.
Phagocytic activity assessment
Alpha Fluor 488 hIgG beads were added at target-to-cell ratio of 10:1 and mildly centrifuged (300 g for 1 min) onto cells. Following incubation at 37°C under 5% CO2 for indicated times in serum-containing medium, cells were washed and blocked with cold FACS buffer (1% BSA, 5 mM EDTA in PBS) for 15 min at 4°C. Next, cells were incubated with APC–CD32 antibody to determine the Fc positive population. Flow cytometry measurements were performed on a BDLSR Fortessa (Becton Dickinson) and analyzed using FlowJo software.
Periphagosomal Ca2+ microdomain measurement imaging was conducted as indicated (Guido et al., 2015; Nunes et al., 2012). Fluo8-AM (4 µM, AAT Bioquest) was loaded in physiological buffer with sulfinpyrazone, for 30 min first at 37°C, then 20 min at room temperature (RT) and with BAPTA-AM (2,5 µM) for 10 more minutes. Fluorescently labeled IgG-opsonized targets were mildly centrifuged (300 g for 1 min) onto cells and incubated for 30 min. Images were acquired on a Nipkow Okagawa Nikon spinning disk confocal microscopy with a temperature controller, motorized stage and Plan Apo 40×/1.3 Oil DICIII objective, using Visiview software (Visitron Systems). At least five snapshots, averaged over 6 s, per coverslip were quantified by custom ImageJ macros as described previously (Guido et al., 2015; Nunes et al., 2012). Briefly, local Ca+2 elevations were defined as microdomains if the area is ≥500 nm2 within a distance of ∼750 nm from the phagosome membrane with a fluorescence value at least 2× s.d. higher than the average cytoplasmic Fluo8 intensity.
All the live imaging experiments were performed on a Nikon spinning disk confocal microscopy with a temperature controller, motorized stage and Plan Apo 40×/1.3 and Plan Apo 63×/1.4 Oil DICIII objective using Visiview software (Visitron Systems). Cells were seeded on 25 mm coverslips the day before transfection and mounted on AttoFluor chambers for microscopy with 2 mM Ca2+ medium before the experiment. For phagosomal lipid assessment, IgG-opsonized targets were added 5–10 min before starting recording. At least five stage positions were selected with cells already started phagocytosing. Z-stack of 9-12 slices spaced at 0.5 µm were imaged every 30 s for each stage position during 20–30 min. Sum projections of background-subtracted images and individual phagosome tracking were done manually with ImageJ. Phagosomal PI(4,5)P2 and actin enrichment was determined by the ratio of the average signal from the phagosome membrane to local average cytosolic signal for the corresponding time point. For PM PI(4,5)P2 analysis, Z-stack of 6–8 slices spaced at 0.5 µm were acquired on 7–10 stage positions per coverslip. Background-subtracted, average projection images of three central slices were created by ImageJ. The ratio of PM fluorescence over cytosol fluorescence was used to determine the enrichment at the PM.
Cells were fixed in 4% paraformaldehyde (PFA) in PBS for 20 min at room temperature, permeabilized in PBS with 0.5% BSA plus 0.1% NP-40 for 10 min and blocked in PBS with 0.5% BSA plus 5% FBS for 1 h at room temperature. For labeling of external beads, after fixation cells were blocked in Fc block in PBS with 1% BSA (1:200) and incubated with anti-human IgG antibody in PBS with 0.5% BSA (1:500) for 1 h at room temperature. Then, cells were incubated with primary antibodies (Myc-Tag, 9B11, Cell Signaling) overnight in a humid chamber at 4°C. Next day, cells were incubated with secondary antibody (1:5000) with Hoechst 33342 (1:10,000; Molecular Probes, H-3570) for 1 h at room temperature. Coverslips were mounted in Fluoromount-GT Slide Mounting Medium (Electron Microscopy Sciences). Beads at cup stage were quantified as the number of beads labeled with human IgG and in contact with cells expressing Fc receptor divided by the total number of beads (internalized plus cup stage). Phagocytic index was determined as the number of ingested beads divided by total number of Fc receptor expressing cells.
Focused-ion beam scanning electron microscopy (FIB-SEM) was performed on a Helios NanoLabG3 microscope (FEI, The Netherlands) as described previously (Nunes-Hasler et al., 2017). Briefly, transfected cells were seeded on 35-mm Ibidi polymer dishes with gridded bottom (catalog number 81166). Following incubation with the target and fixation with 4% PFA for 20 min at room temperature, brightfield and high-resolution confocal images were captured. Next, samples were fixed for EM with 2.5% glutaraldehyde and 2% PFA in Ca-Caco buffer (2 mM CaCl2, 0.15 M sodium cacodylate, pH 7.4) for 3 h on ice and washed five times in ice-cold Ca-Caco buffer. Following dehydration, samples were embedded on Epon and prepared for FIB-SEM imaging as previously described. Samples were sputter-coated with gold for 30 s by a Q150T ES coater (Quorum Technologies, UK). Cellular footprints obtained from FIB-SEM and fluorescence and brightfield images were compared to locate the cell of interest. Images were acquired at the highest resolution setting (5×5×10 nm) using the Autoslice and View software (FEI). Drift correction and alignment of FIB-SEM images were done using Amira Software and overlay with the fluorescence image using ImageJ.
Image analysis and statistics
All images were analyzed with ImageJ software. All statistical analyses were performed with GraphPad Prism 9. Unless otherwise indicated all statistical tests conducted are two-tailed.
We thank Cyril Castelbou for the technical assistance and the bioimaging and electron microscopy core facilities of the Faculty of medicine of the University of Geneva.
Conceptualization: M.K., P.N.-H., N.D.; Formal analysis: M.K., A.C.-S., P.N.-H., N.D.; Investigation: M.K., A.C.-S., P.N.-H., N.D.; Data curation: M.K., A.C.-S.; Writing - original draft: M.K., N.D.; Writing - review & editing: A.C.-S., P.N.-H.; Visualization: N.D.; Supervision: P.N.-H., N.D.; Project administration: N.D.; Funding acquisition: N.D.
This work was funded by the Swiss National Foundation (Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung) [grant numbers 310030, 189042 (to N.D.), 310030, 189094 (to P.N.-H.)], the Sir Jules Thorn Charitable Trust Foundation (to A.C.-S.) and the Novartis Foundation (to A.C.-S.). P.N.-H. is recipient of a career award from the Prof. Dr. Max Cloëtta Foundation. Open Access funding provided by University of Geneva. Deposited in PMC for immediate release.
All relevant data can be found within the article and its supplementary information. Primary data are available from the corresponding author upon reasonable request.
Peer review history
The peer review history is available online at https://journals.biologists.com/jcs/lookup/doi/10.1242/jcs.260902.reviewer-comments.pdf
The authors declare no competing or financial interests.