Research ArticleChemotaxis

The transmembrane adaptor protein NTAL limits mast cell chemotaxis toward prostaglandin E2

See allHide authors and affiliations

Science Signaling  13 Nov 2018:
Vol. 11, Issue 556, eaao4354
DOI: 10.1126/scisignal.aao4354

Adapting to PGE2

The transmembrane adaptor proteins non–T cell activation linker (NTAL) and linker for activation of T cells (LAT) are structurally similar but nonredundant for immune cell function. Whereas LAT promotes mast cell degranulation in response to antibody-antigen complexes, NTAL inhibits this process. Halova et al. found that NTAL, but not LAT, inhibited the chemotaxis of mouse bone marrow–derived mast cells to the signaling lipid PGE2. NTAL reduced the phosphorylation of AKT and ezrin family proteins in PGE2-stimulated mast cells. Experiments with truncation and point mutants of NTAL indicated that C-terminal phosphorylation sites were critical for its inhibitory activity. These data suggest that NTAL not only inhibits mast cell degranulation after antigen stimulation but also limits mast cell migration in response to PGE2.


Chemotaxis of mast cells is one of the crucial steps in their development and function. Non–T cell activation linker (NTAL) is a transmembrane adaptor protein that inhibits the activation of mast cells and B cells in a phosphorylation-dependent manner. Here, we studied the role of NTAL in the migration of mouse mast cells stimulated by prostaglandin E2 (PGE2). Although PGE2 does not induce the tyrosine phosphorylation of NTAL, unlike IgE immune complex antigens, we found that loss of NTAL increased the chemotaxis of mast cells toward PGE2. Stimulation of mast cells that lacked NTAL with PGE2 enhanced the phosphorylation of AKT and the production of phosphatidylinositol 3,4,5-trisphosphate. In resting NTAL-deficient mast cells, phosphorylation of an inhibitory threonine in ERM family proteins accompanied increased activation of β1-containing integrins, which are features often associated with increased invasiveness in tumors. Rescue experiments indicated that only full-length, wild-type NTAL restored the chemotaxis of NTAL-deficient cells toward PGE2. Together, these data suggest that NTAL is a key inhibitor of mast cell chemotaxis toward PGE2, which may act through the RHOA/ERM/β1-integrin and PI3K/AKT axes.


Chemotaxis plays a key role in numerous biological processes including embryonic development, tissue homeostasis, inflammation, and angiogenesis. Chemotaxis is also important in the development and function of mast cells, which are prominent effector cells of the immune system and sentinels within tissues exposed to the external environment (1). Mast cell progenitors are released from hematopoietic tissues into circulation and subsequently migrate into peripheral tissues, where they continue their development (2). Mast cells also accumulate in the proximity of solid tumors, where their mediators are involved in antitumorigenic as well as tumor-promoting and metastatic consequences depending on the type of tumor (3). Chemotaxis is tightly controlled by a plethora of chemoattractants and proper cross-talk between numerous signal transducers, which include surface receptors, kinases, phosphatases, adaptor proteins, and components of the cytoskeleton (1). Among them, transmembrane adaptor proteins (TRAPs) play a critical role, especially in hematopoietic cells.

Although structurally similar, the TRAP non–T cell activation linker (NTAL) mediates B cell activation, and the TRAP linker for activation of T cells (LAT) is required for T cell activation, respectively (46). They exhibit a complementary expression pattern in lymphocytes reflected in their names: Whereas LAT is predominantly found in T but not B lymphocytes, the reverse is true for NTAL (4). Both of these TRAPs are found in mast cells, where they act as key mediators of signals generated by activation of the high-affinity receptor for immunoglobulin E (IgE) (FcεRI) and the receptor for stem cell factor (SCF), c-KIT (710). Despite high structural and sequence homology, NTAL and LAT are found in distinct microdomains of the mast cell plasma membrane (8), and their role in mast cell signal transduction also differs. In bone marrow–derived mast cells (BMMCs), NTAL inhibits (8, 9, 11) and LAT promotes (7) cellular degranulation. However, mast cells deficient in both adaptors degranulate poorly after FcεRI activation (8, 9). NTAL also limits mast cell migration toward IgE-specific multivalent antigen (Ag) but has no effect on chemotaxis toward SCF (10). In contrast, LAT deficiency does not affect chemotaxis toward Ag. Mast cells that lack both adaptors chemotax toward Ag more than wild-type (WT) cells, but less than NTAL-deficient cells (12). NTAL and LAT have several tyrosine motifs in their cytoplasmic domains, on which the function of these adaptors depends (13). Both adaptors are phosphorylated after activation of mast cells; NTAL and LAT are extensively phosphorylated after IgE immune complex (IC)–mediated activation by LYN and SYK (spleen tyrosine kinase), whereas NTAL, but not LAT, is also phosphorylated after SCF-mediated activation by c-KIT (14). Although both Ag and SCF can act as important chemoattractants, their main roles in mast cell physiology are in activation and maturation, respectively (13). Another important chemoattractant for mast cells and some other cell types is prostaglandin E2 (PGE2). One of the major eicosanoids produced during inflammation by a variety of cell types, PGE2, acts as chemoattractant for both mature and immature mast cells through prostaglandin E receptor 3 (EP3), which is a G protein–coupled receptor (GPCR) (15). Although the PGE2-EP3 axis serves important roles in mast cell chemotaxis, its function and regulation are incompletely understood. Furthermore, there are no data on the role of TRAPs in this pathway.

In this study, we examined the role of NTAL in PGE2-mediated chemotaxis in mast cells. We found that phosphorylation of the ezrin/radixin/moesin (ERM) inhibitory threonine and activation of β1 integrin were increased in resting NTAL knockout (KO) BMMCs when compared to WT cells. Ras homolog family member A (RHOA) is more active in resting NTAL KO BMMCs than in WT cells (10). All these features are commonly observed in invading tumor cells (1621). Activation of NTAL KO BMMCs by PGE2 also increased phosphorylation of AKT and phosphatidylinositol 3,4,5-trisphosphate (PIP3) production without altering phosphorylation of NTAL. The combined data provide evidence that transmembrane adaptor NTAL inhibits mast cell chemotaxis toward PGE2.


PGE2 enhanced mast cell chemotaxis in the absence of the adaptor protein NTAL

The aim of this study was to determine the role of NTAL in mast cell chemotaxis toward PGE2. Using Transwell chemotaxis assays, we tested the ability of mast cells derived from WT mice, or mice deficient in NTAL (NTAL KO), LAT (LAT KO), or both adaptors (dKO) to migrate toward PGE2. We found that loss of NTAL significantly increased PGE2-driven chemotaxis of BMMCs (Fig. 1A). Chemotaxis of NTAL KO BMMCs was also increased when compared to LAT-deficient cells or cells that lacked both adaptors (dKO). LAT deficiency also increased chemotaxis when compared to WT cells. However, no difference in PGE2-mediated chemotaxis between WT BMMCs and those deficient in both adaptors (dKO) was observed. To determine whether BMMCs with NTAL KO also exhibit changes in their motility after exposure to PGE2, we examined the motility of BMMCs on fibronectin-coated plates. Once cells were attached to fibronectin, PGE2 was added to the chemotactic medium, and individual cells were tracked for 1 hour. We found that loss of NTAL increased BMMC motility after PGE2 stimulation. However, in response to SCF stimulation, we found that there were no significant differences in the chemotaxis of WT, NTAL KO, and LAT KO BMMCs (fig. S1). Because PGE2 acts through EP3 in BMMCs (22, 23), we measured the expression of EP3 in BMMCs of all genotypes used and found no difference in the expression of EP3 at both the mRNA (Fig. 1C) and protein level (Fig. 1, D and E). Chemotaxis toward PGE2 is likely mediated through the EP3, which is coupled to the Gαi subunit that can be selectively inhibited by pertussis toxin (PTX). When we treated BMMCs of all cell types with PTX for 4 hours and measured chemotaxis toward PGE2, we found that chemotaxis was inhibited to a similar extent in cells of all genotypes (Fig. 1A). The function of TRAPs as scaffold molecules in signal transduction is mostly dependent on their phosphorylation [reviewed by Draber et al. (13) and Rivera (24, 25)]. However, PGE2-stimulated GPCR activation does not cause NTAL and Tyr191 phosphorylation (26). To determine the global tyrosine phosphorylation of LAT and NTAL after PGE2 activation, these adaptor proteins were immunoprecipitated, and their phosphorylation was analyzed by immunoblotting with phosphotyrosine-specific monoclonal antibody (mAb). We found that neither LAT nor NTAL was tyrosine-phosphorylated after PGE2 stimulation. In contrast, the TRAPs were rapidly phosphorylated in BMMCs activated with IgE ICs, and NTAL was also partially phosphorylated after SCF exposure (Fig. 1F).

Fig. 1 NTAL controls BMMC motility and chemotaxis toward PGE2.

(A) Transwell chemotaxis assay analysis of migration in response to PGE2 by WT, NTAL KO, LAT KO, or dKO BMMCs treated (+PTX) or not (−PTX) with pertussis toxin for 4 hours. Data are means ± SEM from 6 (−PTX) to 10 (−/+PTX) experiments. (B) Microscopy analysis of BMMC motility on the fibronectin-coated plates after exposure to PGE2 for 1 hour. Data are means ± SEM of 210 cells pooled from seven independent experiments. (C and D) Western blot analysis of EP3 abundance in lysates of BMMCs. Blots (C) are representative of four independent experiments. Quantified band intensity values (D) are means ± SEM pooled from all experiments. (E) qRT-PCR (quantitative reverse transcription polymerase chain reaction) analysis of EP3 mRNA expression in BMMCs. Data are means ± SEM from two to three different mice per each genotype tested in two independent experiments. (F) Western blot analysis of pY of NTAL (upper) and LAT (lower) immunoprecipitated from lysates of WT BMMCs treated as indicated. Blots are representative of three to four independent experiments. Quantified band intensity values for NTAL (G) and for LAT (H) are means ± SEM pooled from all experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 by one-way analysis of variance (ANOVA) with Tukey’s post-test.

Loss of NTAL promoted AKT phosphorylation and translocation in PGE2-activated mast cells

Because we found that NTAL and LAT were not tyrosine-phosphorylated after PGE2 stimulation, we examined whether other proteins involved in chemotaxis were phosphorylated in the absence of these adaptors. The absence of NTAL increases the phosphorylation of extracellular signal–regulated kinase (ERK) in IgE IC–activated BMMCs (8, 9). When we examined ERK phosphorylation after PGE2 activation, we found that ERK Tyr204 phosphorylation was not affected by the absence of NTAL and/or LAT (Fig. 2, A and B). However, one of the crucial steps in chemotaxis is phosphorylation of protein kinase B (also AKT) and its translocation to the leading edge of the migrating cell (27). When we examined AKT Ser473 phosphorylation, we found that phosphorylation of AKT was significantly greater in PGE2-activated NTAL KO BMMCs than in WT cells (Fig. 2, A and C). The pleckstrin homology (PH) domain of AKT binds to the plasma membrane–bound PIP3 generated by the enzymatic activity of phosphatidylinositol 3-kinase (PI3K) (27). Because PGE2-induced phosphorylation of AKT was inhibited by NTAL, we also tested the role of NTAL in translocation of AKT to the plasma membrane. A plasmid containing the PH domain of AKT fused to enhanced green fluorescent protein (EGFP) (PH-AKT-EGFP) was transfected into NTAL WT and NTAL KO BMMCs, and the location of the PH-AKT-EGFP fusion protein was analyzed by confocal microscopy after PGE2 activation. We found the fusion protein located at the plasma membrane more in NTAL KO than in WT BMMCs (Fig. 2, D and E, and fig. S2). As a control for the PIP3 specificity of the PH-AKT-EGFP protein, we found that a mutant AKT (PH-AKTmut-EGFP) that does not bind to PIP3 was not found at the plasma membrane after PGE2 activation (fig. S2).

Fig. 2 The absence of NTAL increases membrane PIP3 abundance in BMMCs.

(A to C) Western blot analysis of the indicated proteins in lysates from WT, NTAL KO, LAT KO, or dKO BMMCs cytokine-starved overnight and activated with PGE2 for the indicated times. Blots are representative of four to six experiments. Quantified pERK (B) and pAKT (C) band intensity values are means ± SEM pooled from all experiments. (D and E) Immunofluorescence microscopy analysis of PIP3 (red) and nuclear DNA (blue) in WT and NTAL KO BMMCs previously transfected with a plasmid encoding the PH domain of AKT (PH-AKT-EGFP) that were treated with PGE2 for 1 min. Images (D) are representative of four independent experiments. Quantified relative membrane PIP3 data (E) are means ± SEM of about 120 cells from all experiments. Scale bars, 10 μm. *P < 0.05 and **P < 0.01 by one-way ANOVA, Tukey’s post-test (B and C) or by Student’s t test (D). PM, plasma membrane.

NTAL inhibited phosphorylation of ERM and F-actin polymerization in BMMCs

Important players in chemotaxis are the ERM family of proteins that act as a dynamic connection between the plasma membrane and actin cytoskeleton. Transient dephosphorylation of their C-terminal threonine is crucial for proper cell migration (2830). We therefore examined the changes in ERM Thr567/564/558 phosphorylation in PGE2-activated cells by Western blot and found that ERM proteins were extensively dephosphorylated after PGE2 activation (Fig. 3, A and B). We also found that in nonactivated NTAL KO BMMCs, phosphorylation of ERM proteins was increased when compared with WT cells (Fig. 3, A and B). To determine whether NTAL had an effect on the cytoplasmic localization of phosphorylated ERM proteins, we used confocal microscopy. We found that the absence of NTAL increased the abundance of phosphorylated ERM at the plasma membrane of nonactivated BMMCs (Fig. 3, C and D, and fig. S3). We also explored the changes in actin polymerization in WT and NTAL KO cells stimulated with PGE2 stained with labeled phalloidin by confocal microscopy. Activation increased the F-actin content of BMMCs, which peaked at 3 to 5 min after PGE2 stimulation and then gradually returned to the basal level. We found that the peak of F-actin abundance was significantly higher in BMMCs that lacked NTAL (Fig. 3E). Together, these data showed that loss of NTAL expression increased basal phosphorylation of plasma membrane–associated ERM proteins, which enhanced actin polymerization in PGE2-activated mast cells.

Fig. 3 NTAL inhibits ERM phosphorylation, F-actin polymerization, and β1-integrin activation in BMMCs.

(A and B) Western blot analysis of pERM in lysates from WT, NTAL KO, LAT KO, or dKO BMMCs cytokine-starved overnight and activated with PGE2 for the indicated times. Blots are representative of four to seven independent experiments. Quantified pERM band intensity values (B) are means ± SEM from all experiments normalized to WT phosphorylation at 5 min. (C and D) Immunofluorescence microscopy analysis of pERM (red) and nuclear DNA (blue) in cytokine-starved WT and NTAL KO BMMCs. Images (C) are representative of >120 cells from two independent experiments performed in duplicate. Average relative pERM fluorescence intensity (D) was calculated from each experiment (dashed lines). Scale bars, 10 μm. (E) Flow cytometry analysis of F-actin abundance in WT and NTAL KO BMMCs activated with PGE2 for the indicated times. Normalized mean fluorescence intensity (MFI) data are means ± SEM of six independent experiments performed in duplicate. (F) Flow cytometry analysis of activated β1-integrin abundance in WT and NTAL KO BMMCs treated with control or PGE2 for 3 min. Normalized MFI data are means ± SEM of 12 independent experiments. (G) Flow cytometry analysis of β1-integrin abundance in WT and NTAL KO BMMCs. Data are means ± SEM from at least five experiments. *P < 0.05 and **P < 0.01 by one-way ANOVA with Tukey’s post-test (B), paired t test (C), or Student’s t test (E to G).

Absence of NTAL increased basal β1-integrin activation in BMMCs

An important role in adhesion, spreading, and chemotaxis of mast cells is played by β1 integrin, which exists in two basic conformations: closed (nonactivated) and open (activated). The difference in β1-integrin activation can be detected by antibodies (Abs) that bind only to β1 integrin in open conformation (31, 32). When BMMCs are activated with Ag, the abundance of the activated form of β1 integrin on the cell surface increases (33), and this enhances BMMC binding to fibronectin (31). Using WT and NTAL KO BMMCs, we found that the activity of β1 integrin in resting cells was inhibited by NTAL (Fig. 3F). However, after exposure of the BMMCs to PGE2, although the fraction of cells with β1 integrin in open conformation was enhanced, we found no difference in the abundance of activated β1 integrin between WT and NTAL KO BMMCs. In contrast, in IgE IC–activated mast cells, NTAL promoted β1-integrin activation (fig. S4A). The differences that we observed were not due to the differences in total β1-integrin expression, which was similar in both cell types (Fig. 3G). These data suggested that NTAL controls β1-integrin activation in a stimulation-dependent manner.

When cultured in regular cell culture media, BMMCs are nonadhesive. In the resting state, they attach to fibronectin weakly and are easily released from the surface. However, after activation with IgE IC or SCF, they adhere strongly and spread on fibronectin. NTAL promotes this process in IgE-activated cells, but it has no role in SCF-activated cells (10). When we compared PGE2-induced spreading of BMMCs isolated from WT mice or mice deficient in NTAL, we found that NTAL did not affect BMMC spreading (fig. S3, B and C). For comparison, we also determined spreading on fibronectin in the same cells activated with IgE IC and found that PGE2 induced BMMC spreading less efficiently than IgE IC. These data confirmed that NTAL has distinct functions in cell spreading after exposure to either PGE2 or IgE ICs.

Cholesterol and CD9 controlled the migration of BMMCs

NTAL deficiency decreases expression of several genes involved in cholesterol biosynthesis (34). Furthermore, loss of NTAL exacerbates a cholesterol-dependent decrease in BMMC chemotaxis toward Ag that occurs when BMMCs are cultured in media containing cholesterol-depleted fetal bovine sera (FBS) (35). Ag-driven chemotaxis of NTAL KO BMMCs is also more sensitive to depletion of cholesterol from the plasma membrane by methyl-β-cyclodextrin (MβCD) (34). Therefore, we examined the effect of cholesterol reduction on PGE2-driven chemotaxis. When we cultured BMMCs from WT and NTAL KO mice in cholesterol-depleted FBS during PGE2 stimulation, we found that chemotaxis toward PGE2 by NTAL KO BMMCs, but not WT BMMCs, was reduced (Fig. 4A). In parallel experiments, we exposed BMMCs of both genotypes to increasing concentrations of MβCD (0.1 to 5 mM) and analyzed their chemotaxis toward PGE2. We found that at the lowest concentration of MβCD (0.1 mM), cholesterol depletion increased chemotaxis of both WT cells and NTAL KO cells; however, the effect was less prominent in NTAL KO than in WT BMMCs. However, when MβCD was used at higher concentrations, the chemotactic response of NTAL KO BMMCs sharply declined, whereas chemotaxis of WT cells remained unchanged. At the highest concentration of MβCD tested, chemotaxis of both cell types was almost completely inhibited (Fig. 4B). Because these data are different from those obtained when Ag is used as chemoattractant (34), they suggest that the effects of cholesterol on BMMC migration may depend on the stimulation used.

Fig. 4 Enhanced sensitivity of NTAL KO cells to cholesterol depletion and positive role of CD9 expression in PGE2-mediated chemotaxis.

(A) Transwell chemotaxis assay analysis of migration in response to PGE2 by WT and NTAL KO BMMCs previously cultured in media with (+) or without (−) cholesterol. Data are means ± SEM from three to four independent experiments performed in duplicate. (B) Transwell chemotaxis assay analysis of migration to PGE2 by WT and NTAL KO BMMCs after cholesterol depletion with MβCD at the indicated concentrations. Data are means ± SEM from four to six independent experiments. (C) Transwell chemotaxis assay analysis of migration in response to PGE2 by BMMCs transduced with lentivirus vector alone (pLKO.1) or vectors containing two independent shRNA against tetraspanin CD9. Data are means ± SEM from six independent experiments. (D) Fluorescence-activated cell sorting (FACS) was used to evaluate the amount of surface CD9 in transduced BMMC. Data are means ± SEM from three independent testing. *P < 0.05, **P < 0.01, and ***P < 0.001 by one-way ANOVA with Tukey’s post-test (B) or by using Student’s t test (A, C, and D).

Important players of membrane integrity are tetraspanins (34, 36). NTAL and tetraspanin CD9 colocalize at the plasma membrane, and CD9 Ab blockade inhibits the migration of mast cells toward several chemoattractants (12). Reducing the expression of CD9 by 80% has no effect on mast cell migration toward Ag (12). When we reduced the expression of CD9 in BMMCs by knockdown with lentiviral-encoded short hairpin RNAs (shRNAs) (KD1 and KD2), we found that BMMC chemotaxis toward PGE2 was significantly reduced (Fig. 4C). These results suggest that CD9 is differentially coupled to FcεRI and EP3.

NTAL phosphorylation inhibited chemotaxis toward PGE2

Our data appeared to suggest that PGE2 provides a very different stimulation to BMMCs than IgE IC. To test whether PGE2 stimulates similar NTAL phosphorylation, we mutated the residues that are important for the effects of NTAL in response to IgE IC. We complemented NTAL KO BMMCs with EGFP-tagged constructs encoding (i) WT human NTAL (NTALWT), (ii) human NTAL where all 10 tyrosines were mutated to phenylalanines (NTAL10Y-10F), and (iii) a human NTAL truncation mutant (NTALshort), EGFP only (CTRL) (Fig. 5A). After 40 hours, we found that BMMCs expressed all three NTAL isoforms (NTALWT, NTAL10Y-10F, and NTALshort) at the plasma membrane by confocal microscopy (Fig. 5A). In contrast, in NTAL KO BMMCs that received CTRL construct, EGFP was localized in the nucleus and cytoplasm. When we examined Ag-induced degranulation by measuring CD107a on the surface of EGFP-positive cells after Ag stimulation, we found that expression of NTALWT, but not NTAL10Y10F or NTALshort, in BMMCs decreased surface CD107a abundance when compared to the CTRL BMMCs (Fig. 5B). We also examined the role of NTAL phosphorylation in mast cell chemotaxis, by stimulating the same number of EGFP-positive BMMCs with PGE2 or Ag in Transwell chambers (Fig. 5, C and D). Although introduction of NTALWT into NTAL KO BMMCs decreased cellular chemotaxis toward PGE2, no decrease was observed in BMMCs expressing NTAL with tyrosine mutation or truncation (Fig. 5C). We found similar effects after Ag stimulation, although expression of NTAL tyrosine mutation or truncation in BMMCs increased migration over CTRL BMMCs (Fig. 5D). Further, in NTAL KO BMMCs transfected with NTALWT or CTRL constructs, we analyzed pERM phosphorylation by confocal microscopy (Fig. 5, E and F). We found that introduction of NTALWT decreased ERM phosphorylation, which was comparable with the results obtained when NTAL WT and NTAL KO BMMCs were used. These data suggest that tyrosine phosphorylation of NTAL is as important for its inhibitory function after PGE2 stimulation as IgE IC stimulation (8, 9, 14). Further, NTAL may reduce PGE2-mediated chemotaxis through the RHOA/ERM/β1-integrin and PI3K/AKT axes (Fig. 6).

Fig. 5 Importance of NTAL tyrosines in mast cell migration.

(A) Confocal microscopy analysis of NTAL localization in NTAL KO BMMCs transfected with pEGFP-N1 vector (CTRL) or three different NTAL-EGFP constructs (NTALWT, NTAL10Y-10F, and NTALshort). NTAL constructs (top) contained changes to the transmembrane domain (blue boxes) or GRB2-binding site (red), as indicated. Images of nucleus (blue) and NTAL-EGFP constructs or EGFP alone (green) are representative of X independent experiments. Scale bars, 5 μm. (B) Flow cytometry analysis of cell surface CD107a abundance on control or Ag-activated BMMC. Frequency of positive cells are means ± SEM of six independent experiments. (C and D) Transwell chemotaxis assay analysis of migration in response to PGE2 (C) or Ag (D) by NTAL KO BMMCs transfected with the indicated constructs. Data are means ± SE from at least seven independent experiments. (E and F) Microscopy analysis of pERM (red) and control or NTAL (green) non-nuclear (blue) localization in NTAL KO BMMCs transfected with the indicated constructs. Bright-field (BF) and fluorescence images (E) are representative of three independent experiments. Quantified pERM intensity (F) and means are from 30 EGFP-positive cells analyzed in three independent experiments. Scale bars, 5 μm. *P < 0.05, **P < 0.01, and ***P < 0.001 by one-way ANOVA with Tukey’s post-test (B and C) or Student’s t test (E and F).

Fig. 6 Role of NTAL in the activation of ERM, β1 integrin, and RHOA after PGE2 triggering.

When compared with resting WT cells, β1 integrin and RHOA guanosine triphosphatase (GTPase) were increased in NTAL-deficient cells, which promoted phosphorylation of ERM proteins. The N-terminal region of ERM proteins could also directly interact with RHOGDI (RHO guanine nucleotide dissociation inhibitor) (GDI) and uncouple it from the guanosine diphosphate (GDP) form of RHOA. Activation of the EP3 by PGE2 leads to the switch of RHOA from guanosine triphosphate (GTP)– to GDP-bound inactive form and subsequent transient dephosphorylation of ERM inhibitory threonine. This dephosphorylation causes transient uncoupling of the plasma membrane from the actin cytoskeleton and activates β1 integrin. PGE2 stimulation also activates PI3Kγ and subsequent PIP3 production, which serve as binding sites for the PH domain of AKT. AKT is phosphorylated on Ser473 by the mammalian target of rapamycin complex 2 (mTORC2) that is also activated by PGE2. In this model, the size of individual forms of RHOA reflects the amount of its active/inactive form under the indicated conditions. Although switching between GDP- and GTP-bound form of RHOA is also regulated by several GTPase activating proteins (GAPs) and guanine nucleotide exchange factors (GEFs), it remains unclear which individual GAPs and GEFs in PGE2-induced activation and their possible cross-talk with NTAL have not been specified (question marks). PIP2, phosphatidylinositol 4,5-bisphosphate; cAMP, adenosine 3′,5′-monophosphate.


NTAL inhibition of IgE IC–induced activation and chemotaxis of mast cells requires phosphorylation of conserved tyrosine residues to bind Src homology 2 (SH2) domain–containing proteins (13). Similarly, we found that NTAL inhibited chemotaxis of mast cells toward PGE2. Analysis indicated that NTAL inhibited phosphorylation of the inhibitory threonine at the C-terminal part of ERM proteins. Our data fit a model suggesting that NTAL inhibited the activity of β1-integrin and PI3K/AKT axes.

Phosphorylation and transient dephosphorylation of ERM proteins act as a dynamic clutch between plasma membrane phospholipids and actin cytoskeletal proteins, and its precise regulation is necessary for the proper migration in various cell types, including B cells and T cells (30). One of the kinases responsible for ezrin phosphorylation is RHO kinase (ROCK), which is activated by the small GTPase RHOA (18). ERM proteins influence the RHOA activation through direct interaction of the ERM N-terminal region with RHOGDIs and reduce the GDP/GTP exchange of RHOA (37). The amount of active RHOA in resting NTAL KO BMMCs is at least twice as high as in WT cells, and active RHOA transiently decreases after FcεRI activation in WT cells and more rapidly in NTAL KO cells (10). Thus, the increased activity of RHOA in NTAL-deficient cells may contribute to the increased phosphorylation of the ERM family proteins. How exactly NTAL inhibits RHOA activity remains unknown. The activity of RHOA as well as other GTPases is orchestrated not only by RHOGDIs but also by a group of RHO guanine nucleotide exchange factors (RHOGEFs) and RHOGTPase-activating proteins (RHOGAPs), and we speculate that NTAL may serve as an anchor for some of them. For example, VAV, a well-known RHOAGEF, is able to associate with phosphorylated LAT and NTAL (38, 39). The phosphorylation-dependent activity of VAV is influenced by phosphatases like Src homology region 2 domain-containing phosphatase-1 (SHP-1), which also associates with NTAL (40). However, we have not found any differences in VAV phosphorylation and p190GAP phosphorylation between nonactivated WT and NTAL KO cells (10).

The RHOA signaling axis is necessary for the movement of other cell types in which ezrin is localized in the direction of cell movement (41). Increased expression and activation of moesin through the RHOA/ROCK-2 pathway promotes cervical cancer metastasis (42). In NTAL-deficient BMMCs, we found that the basal β1-integrin activation was increased. These results suggested that NTAL may inhibit the RHOA/ERM/β1-integrin axis, which is necessary for mast cell chemotaxis. This conclusion is supported by data indicating that ezrin phosphorylation at Thr567 increases the activity of β1 integrin and the invasiveness of cancer cells (16). NTAL expression is decreased in patients with acute myeloid leukemia, where the Lat2 gene is one of the targets of t(8;21) chromosomal translocation (43, 44), and reduced NTAL expression in T cell acute lymphoblastic leukemia inhibits cellular responses to prednisone therapy (45, 46). In addition, in mast cells, ERM is dephosphorylated by phosphatases, including protein phosphatase 2A (30, 47). This phosphatase interacts with p21-activated kinase 1, and together, they are responsible for F-actin rearrangements (48). However, whether NTAL is involved in the regulation of phosphatases remains unclear.

PGE2 activation of NTAL KO cells enhanced the phosphorylation of AKT Ser473, which is connected with activation of PI3Ks (26, 27, 49). Phosphorylation at this site is mediated by mTORC2, which was identified as central signaling locus in PGE2-mediated chemotaxis of mast cells (50). Chemotaxis is dependent not only on AKT phosphorylation but also on its translocation to the proximity of the plasma membrane (27). In response to IgE IC, PI3K activity promotes the production of PIP3, which serves as binding sites for AKT and other PH domain–containing proteins (49). NTAL constrains PI3K activity in IgE IC–activated BMMCs (8). Here, we found that NTAL also reduced PI3K activity in PGE2-activated cells. Similarly, inhibiting PI3K in mast cells with wortmannin reduces IgE IC–mediated chemotaxis (49). In contrast to chemotaxis toward Ag (10) and PGE2, NTAL has no effect on chemotaxis toward SCF (10). Studies suggests that distinct class I PI3Ks are activated by different chemoattractants, where class IA PI3Kδ is associated with SCF-induced activation and chemotaxis (51, 52) and PI3Kγ is involved in activation and chemotaxis mediated through GPCRs (49, 52, 53). Although which of these PI3Ks is involved in the activation and chemotaxis mediated by Ag is still controversial [discussed by Halova et al. (1) and Draber et al. (13)]. Kuehn et al. suggest that only PI3Kδ is involved in IgE IC activation of mast cells and outlined an autocrine axis that goes through GPCRs (53). Others consider participation of PI3Kγ in autocrine signaling in IgE IC–mediated degranulation and chemotaxis as a key event (54, 55). PI3Kγ is the main PI3K responsible for IgE IC–induced mast cell migration in vitro and in vivo (52). However, the exact mechanism by which NTAL regulates the activity of PI3K remains to be elucidated.

Although NTAL exhibits strong tyrosine phosphorylation after FcεRI- and cKit-mediated activation (810, 12), after exposure of mast cells to PGE2, we could not detect NTAL tyrosine phosphorylation by immunoblotting. NTAL limits lipopolysaccharide-mediated cytokine production in dendritic cells (DCs) independently of its tyrosine phosphorylation (56). Knock-in of NTAL with mutated five distal tyrosines was able to restore the WT phenotype DCs, suggesting the dispensability of these tyrosines for cytokine production. In contrast, mast cells having a mutant form of NTAL10Y-10F were unable to chemotax toward PGE2, indicating that these tyrosines are important for that function. In PGE2-activated mast cells, NTAL tyrosines may be only transiently phosphorylated, thus not readily detectable by immunoblotting. In this case, phosphorylation could still be sufficient for binding of other molecules responsible for signal transduction that are indispensable for chemotaxis. Additionally, NTAL Tyr residues at distinct sites could be of greater importance for chemotaxis. For example, of the five NTAL Tyr residues closer to the membrane, there is one within the growth factor receptor–bound protein 2 (GRB2)–binding motif (57) and one that was described to bind SHP-1 (40). The N-terminal part of NTAL is also important for pre–B cell differentiation (58). Alternatively, our study cannot exclude that mutation of Tyr into Phe changed the charge and/or conformation of the cytoplasmic part of NTAL in such a way that it was no longer able to fulfill its function in the chemotaxis regulation. However, in mast cells stimulated by an inhibitor of the endoplasmic reticulum Ca2+ adenosine triphosphatase, thapsigargin, the uptake of extracellular calcium directly correlated with the amount of cellular NTAL, although no enhanced tyrosine phosphorylation of NTAL was observed in thapsigargin-activated cells (11). Similarly, although NTAL inhibits PGE2-mediated chemotaxis, we found that NTAL had no effect on both PGE2-induced spreading and β1-integrin activation. Similarly, NTAL did not affect ERMT567/564/558 dephosphorylation in PGE2-activated cells. However, PGE2 induces only weak attachment and subsequent spreading of BMMCs on fibronectin.

Cleavage of a short C-terminal part of NTAL reduces the activation of human B cells and monocytes (59). In contrast, we found that transfection of NTAL KO BMMCs with a truncated version of NTAL that lacked the cytoplasmic domain (NTALshort) did not prevent IgE IC–induced degranulation and chemotaxis toward PGE2 in these cells. However, both NTALshort and NTAL10Y-10F increased chemotaxis toward Ag. These differences may suggest that the response to different stimuli has different requirements for NTAL. Alternatively, in contrast to B cells and monocytes, mast cells have the transmembrane adaptor LAT, which is structurally similar to NTAL. Although LAT promotes mast cell activation, it can also limit chemotaxis (12). Our data indicated that chemotaxis toward PGE2 was not further increased when NTALshort and NTAL10Y-10F were transfected to the cells. However, we cannot exclude the possibility that chemotaxis toward PGE2 in the absence of NTAL is at its maximum and could not be further increased.

Depletion of cholesterol differentially affected chemotaxis of WT and NTAL-KO cells toward PGE2 and Ag. Organization of the plasma membrane microdomains is dependent on the cholesterol content, which is critical for proper cross-talk between FcεRI and TRAPs (13, 60). Pretreatment of BMMCs with MβCD before IgE IC activation decreases phosphorylation of LAT, NTAL, SYK, and AKT, whereas phosphorylation of ERK is unaffected (61). Cholesterol directly and indirectly modulates the function of GPCRs in a receptor-dependent manner (62), and the function of p84 to p110γ subunits of PI3Kγ is sensitive to cholesterol depletion (63). Thus, our data suggest that cholesterol may differentially affect the cross-talk between NTAL and distinct mast cell receptors.

Tetraspanin CD9 augments mast cell chemotaxis toward PGE2. NTAL colocalizes with tetraspanin CD9, and NTAL KO BMMCs are more susceptible to chemotaxis inhibition by mAb against CD9 than WT BMMCs (12). Both TRAPs and tetraspanins are palmitoylated transmembrane molecules that can modulate chemotaxis through association with membrane microdomains (36). Tetraspanins CD9 and CD81 are important scaffolds for GPCRs (64) and in this way could be important for the proper PGE2-mediated chemotaxis. Although NTAL and CD9 are colocalized at the plasma membrane, they have opposite roles in PGE2 signaling. We therefore propose that the cross-talk between NTAL and CD9 contributes to PGE2-induced chemotaxis.

In conclusion, our data showed that although NTAL was not tyrosine-phosphorylated in PGE2-activated mast cells, it was a potent inhibitor of mast cell migration toward PGE2. In mast cells, NTAL reduced PGE2-mediated chemotaxis by limiting RHOA/ERM/β1-integrin activation, as well as PI3K/AKT signaling axes. The chemotaxis of mast cells is also modulated by both cellular cholesterol content and plasma membrane tetraspanin CD9. Our findings connecting TRAPs with ERM functions and integrin signaling in mast cells may also be relevant to cancer cell migration where changes in ezrin expression correlate with metastasis and aggressiveness of tumor progression (1621).


Abs and reagents

Abs against the following were used in this study: pY204 ERK (RRID:AB_668964), pS473AKT (RRID:AB_667741), AKT1 (RRID:AB_630849), hypoxanthine phosphoribosyltransferase (HPRT) 1 (F-1, catalog number sc-376938), EP3 (RRID:AB_2174758), horseradish peroxidase (HRP)–conjugated goat Ab specific for mouse IgG (RRID:AB_631736), HRP-conjugated goat Ab specific for rabbit IgG (RRID:AB_631746), and HRP-conjugated donkey Ab specific for goat IgG (RRID:AB_631730) were from Santa Cruz Biotechnology Inc.; pezrin (T565)/radixin (T564)/moesin (T558) (RRID:AB_823497) and ezrin (RRID:AB_2100309) were from Cell Signaling Technology; NTAL (RRID:AB_10737203) was from EXBIO Praha; GFP (RRID:AB_303395) was from Abcam; HRP-conjugated mouse pY-specific mAb (RRID:AB_397433), activated β1 integrin (HM β1-1, RRID:AB_395077 and 9EG7, RRID:AB_395001), and V450-conjugated rat Ab specific for mouse CD107a (RRID:AB_1727420) were from BD Biosciences; Alexa Fluor 488 (AF488)–conjugated donkey Ab specific for rabbit IgG (RRID:AB_2535792), AF488-conjugated goat Ab specific for hamster IgG (RRID:AB_2535759), AF568-conjugated goat Ab specific for rabbit IgG (RRID:AB_10563566), and AF555-conjugated goat Ab specific for rabbit IgG (RRID:AB_2535849) were from Life Technologies; and DyLight 488–conjugated goat Ab specific for rat Fcγ (RRID:AB_2338316) was from Jackson ImmunoResearch Laboratories. We also used mAb against LAT (65) and γ-tubulin complex component 2 (GCP2) (66), as well as polyclonal Abs specific for LAT and NTAL (67). Trinitrophenol (TNP)–specific IgE Ab (IGEL b4 1) (68) and TNP-BSA (bovine serum albumin) conjugate (Ag; 15 to 25 mol TNP/mol BSA) were produced as described (69). All other chemicals were obtained from Sigma-Aldrich, if not stated otherwise.

Cells and lentiviral infection

BMMCs were derived from 6- to 8-week-old littermates (WT, Ntal−/−, Lat−/−, and Ntal−/−/Lat−/− mice) obtained by crossing of Ntal+/−/Lat+/− mice (8). All work with animals was conducted in accordance with the Institute of Molecular Genetics guidelines (permit number 12135/2010-17210) and national guidelines (2048/2004-1020). Bone marrow cells were isolated and cultured to obtain BMMCs as described (70). Cells with CD9 KD and corresponding controls were prepared by lentiviral transduction of BMMCs with suitable shRNAs (TRCN0000066393 or TRCN0000066395 cloned into pLKO.1 vector or empty pLKO.1 vector) as described previously (12).

Plasmid preparation and cell transfection

Plasmids containing the PH domain of AKT or its mutated version fused to EGFP (PH-AKT-EGFP and PH-AKTmut-EGFP, respectively) were a gift of T. Balla [National Institutes of Health (NIH), Bethesda, USA]. pEGFP1-N1 plasmids containing human NTAL (NTALWT) and NTAL with all tyrosines mutated to phenylalanines (NTAL10Y-10F) were engineered using PCR in which expression vector pEFBOS containing NTALWT or NTAL10Y-10F [provided by V. Horejsi, Institute of Molecular Genetics (IMG), Prague, Czech Republic] (14) was used as matrix and subcloned into pEGFP1-N1 expression vector (Clontech Laboratories Inc.; 6085-1) using Eco RI and Sal I or Kpn I restriction sites, respectively. The following primers were used for PCR (restriction sites underlined): forward (for both constructs), 5′-ttgaattcgccgccaccatgagctcggggactgaactg-3′; reverse (NTALWT), 5′-ttgtcgacttctgtggctgccacc-3′ NTAL10Y-10F 5′-ttggtaccgcttctgtggctgccac-3′. NTALshort was prepared by deleting the intracellular part of NTAL from pEGFP-N1- NTALWT plasmid (between 263 and 749 nucleotides counting from the NTAL start codon) by Bam HI. All constructs were verified by DNA sequencing. Plasmids were introduced into BMMCs by Amaxa Nucleofector II (Lonza, Germany) nucleofection according to the manufacturer’s instructions (program Y-001). Cells were left to recover for 36 hours, then starved for further 4 hours in media without interleukin-3 (IL-3) and SCF, and used for experiments.

Activation, immunoprecipitation, and analysis of protein phosphorylation

For all experiments, cells were kept in SCF- and IL-3–free culture medium for 16 hours in the presence or absence of IgE mAb (IGEL b4 1; 1 μg/ml), unless stated otherwise. Preparation of whole-cell lysates, immunoblotting (70), and immunoprecipitation of LAT and NTAL (12) were described previously. The amount of phosphorylated proteins was normalized to the loading controls (nonphosphorylated protein).

Quantitative reverse transcription polymerase chain reaction

RNA was isolated, and quantitative PCRs were performed as previously described (70). Glyceraldehyde phosphate dehydrogenase, ubiquitin, and HPRT were used as reference genes. The expression levels of EP3 mRNA (primers: 5′ aacctggcgaccatcaaag 3′ and 5′ ccgtctccgtggtgattctg 3′) were normalized to the geometric mean of the reference genes.

Measurement of activated integrin on the cell surface, cell spreading, and F-actin content

Surface localization of integrin β1 and its activated epitope 9EG7 in BMMCs activated or not with TNP-BSA (100 ng/ml) or PGE2 (100 nM) was quantified by FACS as described (33). Cell spreading was analyzed on fibronectin-coated glass plates (12). Image processing and analysis were completed by means of CellProfiler software (Broad Institute, Boston, MA) (71). For the F-actin content measurement, cells (2 × 105 per well) were activated in a 96-well plate by 100 μM PGE2 for the indicated time periods and then fixed with 4% paraformaldehyde, washed three times in buffered salt solution with 0.1% BSA, and then stained with lysophosphatidylcholine (120 μg/ml) and AF488-phalloidin (0.2 U/ml). The stained cells were analyzed in a BD LSR II flow cytometer.

Confocal microscopy

Preparation, labeling, and examination of samples by confocal microscopy were described previously (33). After labeling, the cells were washed and mounted in 90% (w/v) glycerol in phosphate-buffered saline (pH 8.5) supplemented with antifading agents 2.5% 1,4-diazabicyclo-(2,2,2-octane), whenever EGFP-transfected cells were used, or 0.1% n-propyl gallate in all other cases. In each experiment, all images were acquired at identical microscope settings. Image analyses were performed using a pipeline generated in CellProfiler cell image analysis software (Broad Institute, Cambridge, MA) (71).

To quantify the phosphorylation of ERM proteins at the plasma membrane, NTAL WT and KO BMMCs were used (two different WT and KO BMMCs in each experiment), the average signals for WT and KO BMMCs were paired for each experiment to avoid possible differences in fluorescent intensities from independent experiments and analyzed by paired Student’s t test (average intensities in all four independent experiments were 10 to 20% higher in NTAL KO BMMCs than in WT). Cells were labeled with ERM protein phosphorylated threonine (T567/564/558) specific rabbit Ab, followed by goat anti-rabbit AF568. For each cell type, 20 to 40 cells were identified using nuclei, and fluorescent signal was calculated. In case of a rescue experiment, only cells positive for EGFP were taken into analysis. Five to 20 EGFP-positive cells were taken into each experiment, and 30 EGFP-positive cells were analyzed in total. For analysis of PIP3 production, NTAL KO and WT cells were transfected with PH-AKT-EGFP or PH-AKTmut-EGFP to monitor PIP3 production. BMMCs were attached to fibronectin-coated plates and activated (PGE2, 100 nM) or not for 1 min. Signal was amplified by staining with rabbit anti-GFP followed by anti-rabbit AF555. To examine the PH domain plasma membrane localization, 20 to 40 cells were identified using nuclei, and EGFP signals were segmented into four concentric rings centered on the nucleus. The percentage of plasma membrane–associated EGFP was determined as fraction of the EGFP signal of the outermost ring.

Chemotactic response and detection of surface CD107a

Chemotactic responses were assayed using 24-well Transwell chambers (Corning) with 8-μm porosity polycarbonate filters and 2 × 105 of different BMMC types in the upper well (12). For studies of the role of cholesterol in chemotaxis, BMMCs were treated with various concentrations of MβCD or cultivated in cholesterol-free FBS (34). Analysis of BMMC motility and chemotaxis on fibronectin-coated plates was performed as described (72). PGE2 at a final concentration of 100 nM was used as chemoattractant. Rescue experiments were carried out with NTAL KO BMMCs transfected with vectors containing three different NTAL-EGFP constructs or empty vector. The same number (3 × 105) of EGFP-positive cells was put into each Transwell chamber, and chemoattractant was added into the lower well. After 8 hours, the number of EGFP-positive cells that migrated into the lower well was counted by flow cytometer Accuri C6. For analysis of surface CD107a, transfected BMMCs were activated or not with Ag (TNP-BSA; 250 ng/ml, 15 min; 37°C). Cells positive for EGFP were analyzed by flow cytometry and determined using FlowJo software (Tree Star, Ashland, OR), and the percentage of CD107a-positive cells was determined (33).

Statistical analysis

Statistical analyses were performed using Prism software (GraphPad Software). Comparisons between groups were analyzed by Student’s t test (two groups), or one-way ANOVA followed by Tukey’s post-test (more than two groups), as indicated.


Fig. S1. Neither NTAL nor LAT reduce chemotaxis of BMMCs toward SCF.

Fig. S2. Increased translocation of AKT-PH domain to the membrane in the absence of NTAL.

Fig. S3. Enhanced phosphorylation of ERM protein inhibitory threonine in NTAL-deficient BMMCs.

Fig. S4. NTAL increases β1-integrin activation and cell spreading on fibronectin in Ag-activated cells but has no role in PGE2-induced spreading.


Acknowledgments: We would like to thank V. Horejsi and Pavel Draber (both IMG CAS, Prague, Czech Republic) and T. Balla (NIH, Bethesda, USA) for providing essential Abs and plasmids (described in Materials and Methods), J. Janacek (IPHYS CAS, Prague, Czech Republic) for the help with statistics, and S. Takacova for critical reading of the manuscript. Funding: This work was supported by projects 17-20255S (to L.D.), 17-20915S (to P.D.), and 18-18521S (to P.D.) from the Czech Science Foundation and by the Institute of Molecular Genetics of the Academy of Sciences of the Czech Republic (RVO 68378050 to P.D.). Author contributions: I.H., M.B., L.D., V.B., and P.D. designed the experiments. I.H., M.B., L.D., and V.B. performed the experiments and analyzed the results. I.H. and P.D. wrote the manuscript. Competing interests: The authors declare that they have no competing financial interests. Data and materials availability: All data needed to evaluate the conclusion in the paper are present in the paper or the Supplementary Materials.

Stay Connected to Science Signaling

Navigate This Article