The ability of HIV-1 to replicate and to establish long-term reservoirs is strongly influenced by T cell activation. Through the use of membrane-tethered, genetically encoded calcium (Ca2+) indicators, we were able to detect for the first time, to our knowledge, the formation of Ca2+ territories and determine their role in coordinating the functional signaling nanostructure of the synaptic membrane. Consequently, we report a previously unknown immune subversion mechanism involving HIV-1 exploitation, through its Nef accessory protein, of the interconnectivity among three evolutionarily conserved cellular processes: vesicle traffic, signaling compartmentalization, and the second messenger Ca2+. We found that HIV-1 Nef specifically associates with the traffic regulators MAL and Rab11b compelling the vesicular accumulation of Lck. Through its association with MAL and Rab11b, Nef co-opts Lck switchlike function driving the formation Ca2+ membrane territories, which, in turn, control the fusion of LAT-transporting Rab27 and Rab37 vesicles and the formation of LAT nanoclusters at the immunological synapse. Consequently, HIV-1 Nef disengages TCR triggering from the generation of p-LAT and p-SLP nanoclusters driving TCR signal amplification and diversification. Altogether our results indicate that HIV-1 exploits the interconnectivity among vesicle traffic, Ca2+ membrane territories, and signaling nanoclusters to modulate T cell signaling and function.
This article is featured in In This Issue, p.3749
The ability of HIV-1 to replicate in T cells is dependent on the T cell activation state (1, 2). The accessory HIV-1 protein Nef enhances viral replication by modulating multiple signaling pathways through a plethora of interactions with cellular proteins (3). According to the literature, Nef may interact with as many as 60 cellular factors and affect the function of >180 proteins (3). Notwithstanding, there is a pressing need to progress from this encyclopedic listing of Nef interactions into a more conceptual approach and determine how Nef intersects the regulatory mechanisms controlling the activation and/or differentiation state of the infected cell.
T cell activation is coordinated at a specialized interface that forms upon contact with an APC, known as immunological synapse. TCR triggering induces the recruitment and activation of the kinases Lck and ZAP70, which, in turn, phosphorylates LAT. p-LAT recruitment of SLP76 allows for the formation of a signaling supramolecular scaffold, nucleating various signaling complexes involved in remodeling the T cell cytoskeleton (4), T cell development (5), and activation (6). Thus, the assembly of LAT-SLP76 signaling complexes, also called the LAT signalosome, is critical for T cell activation. The vesicular traffic has emerged as a central player in the assembly of the signaling machinery at the immunological synapse (7).
The TCR as well as the membrane-associated Lck and LAT signaling molecules exploit the vesicular traffic to concentrate at the immunological synapse (7). Recent works have started to address the molecular mechanisms that regulate the exocytic targeting of different vesicular compartments at the immunological synapse (8–10).
Along these lines, we have determined that the regulated release of LAT vesicles to the immunological synapse determines the functional nanostructure of the synaptic membrane and the coordination of the immune response (7, 8). We showed that Lck acts as the signal switch and Ca2+ acts as the mediator of a vesicle fusion feedback loop that builds a functional immunological synapse capable of driving T cell activation and cytokine production (8). Thus, deciphering the spatial organization of signaling proteins is not only key to understanding the mechanisms that underlie immune cell activation, but also to identifying common design principals that are likely to be co-opted by HIV-1.
In this work, we aimed to decipher molecular mechanisms underpinning the control of T cell activation to identify the common design principals that are likely to be co-opted by HIV infection. We uncovered a previously unknown immune subversion mechanism involving HIV-1 exploitation, through its Nef accessory protein, of the interconnectivity among three evolutionarily-conserved cellular processes: vesicle traffic, signaling compartmentalization, and the second messenger Ca2+. We found that Nef selectively interacts with Lck vesicle traffic regulators Rab11b and MAL restraining Lck from the immunological synapse. We pioneered the use of membrane-tethered, genetically encoded Ca2+ indicators (11) to detect and characterize the formation of Ca2+ territories at the TCR-stimulated membrane. Interestingly, we determined that Nef impairs the formation of Ca2+ membrane territories at the immunological synapse even though it has no effect on cytosolic Ca2+ fluxes (12). Finally, by restraining Lck away from the immunological synapse, Nef co-opts the switch regulatory function of synaptic Lck in promoting the generation of Ca2+ membrane territories controlling LAT vesicle fusion and the signaling nanoarchitecture of the immunological synapse.
Materials and Methods
Vectors and transfection
Lck membrane anchoring (MA) domain (LckMA)-LAT-GFP and LAT MA domain (LATTM)-Lck-GFP were constructed by PCR-fusion of the Lck kinase aa 1–18 and LAT aa 28–262 and of the LAT aa 1–27 and Lck aa 19–501, respectively. pN1-Fyn-GCaMP3 was obtained by replacing the membrane targeting sequence of Lck with the one of Fyn, through Gibson assembly, on pN1-Lck-GCaMP3 construct obtained from Adgene through Gibson Assembly. The construct of Nef allele NL4.3 and its mutants fused to GFP or mCherry were gifts from F. Niedergang (Institut Cochin, Paris, France). The Gts-Nef (revGlc-NefΔSH4) and Fyn MA domain (FynMA)-Lck-GFP (FynN18-Lck.GFP) constructs were provided by O. T. Fackler (University Hospital Heidelberg, Heidelberg, Germany); the latter was used to clone FynMA-Lck-mCherry. The MAL-GFP construct was a gift from M. A. Alonso (Universidad Autonoma de Madrid, Madrid, Spain); the Rab11b-GFP and Rab27-GFP constructs were a gift from D. Barral (Chronic Diseases Research Center, Lisbon, Portugal), and the Rab37-GFP construct was a gift from B. Goud (Institut Couchin, Paris, France). DNA constructs were inserted into Jurkat and primary CD4 T cells using the Invitrogen Neon Transfection system. Transiently transfected cells were analyzed 24 h after transfection. Primary CD4 T cells were transfected once with small interfering RNA oligonucleotide pools (Invitrogen) targeting Lck. Silenced cells were analyzed 72 h after transfection.
Virus production and infection
To generate viral stocks, we transfected 293T cells with the HIV-1 NL4.3 and HIV-1 NL4.3 ΔNef constructs. Jurkat cells were infected with HIV-1 NL4.3 and HIV-1 NL4.3 ΔNef virus for 3 d before imaging as described before (13).
Cells, reagents, and immunofluorescence
Jurkat cell clones E6.1 and Lck−/− Jurkat JCAM1.6 cells were grown in complete RPMI 1640 medium containing 10% (v/v) FBS, nonessential amino acids, and l-glutamine. Human peripheral blood cells were stimulated with PHA (1 mg/ml) and grown for 7 d in RPMI 1640 medium with 10% FCS and 20 IU/ml IL-2. Glass coverslips coated with poly-lysine overnight at 4°C and, when stated, further incubated with stimulatory (anti-CD3 MEM-92 [EXBIO] and CD28 [Genentech]) or nonstimulatory (anti-CD45 mAb GAP 8.3; American Type Culture Collection [ATCC]) Abs at a concentration of 10 μg/ml overnight at 4°C, as previously described (8). In some cases, thapsigargin (5 μM; TPS) was added during the stimulation period. For imaging of Fyn-CaGMP3, transfected cells were imaged on RPMI 1640 medium at 37°C. For all other conditions, cells were then fixed with 4% paraformaldehyde for 30 min at room temperature, incubated with blocking buffer (PBS BSA 1%) with or without 0.05% saponin, and stained with primary Abs against Lck (3A5; Santa Cruz), LAT and p-LAT (Cell Signaling), p-SLP (J141; BD Biosciences), CD45 (GAP 8.3; ATCC), hemagglutinin (HA; ab9110 and ab18181; Abcam), and Gag (National Institutes of Health AIDS reagent program) for 1 h, washed, and incubated for 45 min with Alexa 568– and Alexa 647–conjugated secondary Abs (Life Technologies).
Imaging, image processing, and quantification
Confocal images were obtained using a Zeiss LSM 710 confocal microscope (Carl Zeiss) over a 63× objective. 3D image deconvolution was performed using Huygens Essential (version 3.0; Scientific Volume Imaging), and 2D images were generated from a maximum intensity projection over a 3D volume cut of 1-μm depth centered either on the intracellular compartment when visible or on the cell center. For the quantification of Lck and LAT subcellular distribution, the plasma membrane of CD4 T cells was surface labeled with anti-CD45 mAbs. Plasma membrane segmentation was implemented through the generation of binary mask delimitating the internal and external outlines of the plasma membrane using the Fiji image analysis software.
Fluorescence resonance energy transfer imaging
A first set of fluorescence resonance energy transfer (FRET)-donor (GFP-tagged proteins) and FRET-acceptor (mCherry-tagged proteins) confocal images were acquired at low excitation intensity. Acceptor photobleaching with a 561-nm laser was then carried out through strong illumination of a region of interest corresponding to the intracellular compartment of each cell. mCherry showed 90% decrease in signal within the defined region of interest. A second set of images was acquired at low excitation intensity after acceptor photobleaching. Both prebleaching and postbleaching datasets were corrected for drift and analyzed with the AccPbFRET plugin for ImageJ software (GraphPad).
Glass coverslips were washed two to three times in optical grade acetone, soaked overnight, and sonicated in 0.1 M KOH for 20 min. Coverslips were then thoroughly rinsed in deionized water and dried. The glass coverslips were coated overnight at room temperature with 0.001% poly-l-lysine (Sigma) diluted in PBS. Dried coverslips were subsequently incubated with stimulatory (anti-CD3 MEM-92 [EXBIO] and CD28 [Genentech]) or nonstimulatory (anti-CD45 mAb GAP 8.3 [ATCC]) Abs at a concentration of 10 μg/ml overnight at 4°C. Cells were resuspended in imaging buffer, and 500,000 cells were dropped onto the coverslips and incubated for 5 min at 37°C. In some cases, TPS (5 μM) was added during the incubation period. Cells were then fixed with 4% paraformaldehyde for 30 min at room temperature. To maximize the number of detected molecules, care was taken to minimize photobleaching: cells were embedded in oxygen scavenger buffer, imaged on the same day of labeling, and kept in the dark until imaged. dSTORM images were then acquired with a Nikon Ti microscope coupled with a Hamamatsu Flash Orca 4.0 sCMOS camera and a 100× Plan Apo λ 1.45NA oil immersion objective at a 50-Hz rate. Wavelength lasers lines of 561 nm (Coherent Genesis MX 561-500 STM) and 638 nm (Vortran Stradus 642-110) were used to excite Alexa Fluor 568 and Alexa Fluor 647, respectively, using emission filters 595/50 and 655LP from Chroma Technology to further reduce and limit cross-talk. For a better signal-to-noise ratio, a pseudo–total internal reflection (TIRF) illumination (oblique illumination) was used. Acquisition was done through μManager freeware (14), and acquired data were then processed and reconstructed through Fiji (15) using the superresolution plugin ThunderSTORM (16). Once events were detected, a data table was exported with x–y coordinates of each molecule, photon count, and spatial precision, which could later be converted into an image. These datasets were then processed to check for possible cluster formation.
For spatial point pattern of molecular detection, Ripley’s K-function was then calculated through a Fiji macro of commands as follows:
where A is the area of the analyzed region (which depends on the radius), r is the radius of the analyzed area, n is the number of detected particles, and δ is the distance between points i and j. This counts the number of molecules inside the circle area around each detected particle and normalizes it to the average molecular density of all areas measured to access possible cluster formation.
Jurkat cells were stained with FITC-labeled mAbs against CD4 (RPA-T4; BD Biosciences) and with unconjugated mAbs against HA (ab9110; Abcam) followed by incubation with Alexa 647–labeled anti-rabbit Ab (Life Technologies). Cell-associated fluorescence was collected on a FACSCalibur and analyzed using the FlowJo software (BD Biosciences). For Ca2+ flow measurements, Jurkat cells (5.106/ml) resuspended in phenol red–free RPMI 1640 were loaded with 1 mM of Indo-1 (Molecular Probes) at 37°C in the dark for 45 min. Cells were washed and resuspended in the same medium. Ca2+ measurements were performed on a MoFlo flow cytometer (Beckman Coulter). Cells were acquired for 90 s to establish the baseline followed by ionomycin addition (10 μg/ml). A ratiometric analysis was performed using the FlowJo software (BD Biosciences).
The N terminus of HIV-1 Nef promotes the selective accumulation of Lck, but not LAT, in the vesicular compartment
In recent years, the vesicle traffic of signaling molecules has emerged as a central regulator of T cell activation (7–10). In particular, we have determined that the formation of nanoterritories where signaling is coordinated to promote T cell activation is regulated by a vesicular traffic amplification loop involving the sequential vesicle delivery of Lck and LAT at the immunological synapse (8). Even though the main determinant of HIV-1 pathogenesis is the modulation of TCR signaling by the viral protein Nef, it remains poorly understood how Nef exploits the interconnectivity between vesicular traffic and synaptic signaling underpinning T cell activation.
To address this open question, we started by determining the effect of Nef on Lck and LAT subcellular distribution (7). As previously reported, Nef retains Lck in an enlarged perinuclear vesicular compartment (7, 13, 17) (Fig. 1A, 1B), both in Jurkat and in primary CD4 T cells (Fig. 1A, 1B). Unexpectedly, Nef does not alter the subcellular distribution of LAT, despite their adjacent localization in the tightly filled T cell cytoplasm (Fig. 1A, 1B).
To gain mechanistic insights on how Nef specifically alters Lck subcellular distribution, we tailored the expression of a large set of GFP-tagged Lck and LAT chimeric proteins (Fig. 1C) and assessed their subcellular distribution (Fig. 1D, 1E). Previous works have reported that the biological activity of Nef is controlled by the specificity of its membrane targeting (18), but also that the LckMA, which encodes Lck traffic properties, is necessary and sufficient to undergo Nef-mediated retention (18). First, we confirmed that rerouting Nef to the Golgi compartment prevents Lck vesicular accumulation (Fig. 1A–C, Gts-Nef), underscoring the relevance of specific membrane targeting to Nef biological activity. Next, to clarify whether the LckMA is necessary and sufficient to undergo Nef-mediated retention, we used an Lck construct encoding solely the MA domain consisting of its first 18 aa and devoid of additional protein interaction surfaces fused to GFP (LckMA), and expressed it in Lck-deficient Lck−/− JCAM1.6 cells. Cotransfection of Nef-mCherry and LckMA-GFP in Lck−/− JCAM1.6 cells caused the vesicular accumulation of the LckMA-GFP to the similar extent to the one verified with endogenous Lck (Fig. 1A, 1B, 1D, 1E), confirming that the LckMA is sufficient to undergo Nef-mediated vesicular accumulation (19).
Lck and the closely related Fyn kinase display distinct MA domains and consequently are segregated to distinct vesicular compartments (20). Because Nef does not alter the vesicular distribution of the closely related Fyn kinase (17), we then expressed in Jurkat T cells a chimeric protein with all the Lck functional motifs except for the MA domain, which has been replaced by the one of Fyn (Fig. 1C, FynMA-Lck). Despite the presence of all the Lck interaction motifs in the FynMA-Lck chimeric protein, Nef did not cause the vesicular accumulation of FynMA-Lck (Fig. 1D, 1E).
We next interchanged Lck and LAT trafficking properties by constructing chimeric proteins in which the LckMA was replaced by the LATTM (Fig. 1C, LATTM-Lck) and the LATTM was replaced by the LckMA (Fig. 1C, LckMA-LAT). Conferring LAT traffic properties to Lck abrogated its retention by Nef and reciprocally, endowing LAT with Lck traffic properties led to LAT vesicular accumulation in Nef-expressing Jurkat cells (Fig. 1D, 1E).
These results indicate that the N-terminal domain of Lck is required for its selective vesicular accumulation mediated by Nef.
HIV-1 Nef selectively associates with Lck traffic regulators
The specification of the vesicular identity and traffic route rely on the Rab family GTPases (21). Lck and LAT reside and traffic in distinct vesicular compartments, with little to no overlap in their Rab specification (8). Whereas Lck localizes to Rab11b+ and MAL+ vesicles (8, 22, 23), LAT resides in vesicles marked with Rab27 and Rab37 (8). We transiently coexpressed each Rab-GFP protein with Nef-mCherry, in Jurkat and in primary T cells, and assessed their colocalization in the vesicular compartment. To exclude the possibility that Nef colocalization with MAL and Rab11b could be caused by Nef interaction with Lck (24, 25), we first investigated whether Nef colocalized preferentially with Lck traffic regulators, in Lck−/− JCAM1.6 and primary T cells (Fig. 2A). In both Jurkat and primary T cells, Nef highly colocalized with Lck traffic regulators Rab11b and MAL (∼80%; Fig. 2A, 2B) and only partially colocalized with Rab27 and Rab37 (∼20%; Fig. 2A, 2B), even when Lck (LATTM-Lck) was targeted to LAT transporting Rab27 and Rab37 vesicles (∼8%; Fig. 2A, 2B). Conversely, rerouting Nef to the Golgi (Gts-Nef) abolished its colocalization with Rab11b and MAL (∼5%; Fig. 2A, 2B). Altogether, these results indicate that Lck vesicular accumulation requires Nef association with its traffic regulators Rab11b and MAL, and that this association is established independently of Lck.
To analyze whether Nef does in fact associate with Lck traffic regulators, we performed FRET analysis through acceptor photobleaching, which detects molecular proximities <10 nm (26).
Nef associated with both Rab11b and MAL even in the absence of Lck (∼20% FRET efficiency; Fig. 3A, 3B, Lck−/− JCAM1.6). However, Nef did not associate with LAT traffic regulators Rab27 and Rab37 (∼3% FRET efficiency; Fig. 3A, 3B, Jurkat cells). Consistent with our previous findings, rerouting Nef to the Golgi abolished Nef association with Lck traffic regulators Rab11b and MAL (∼2% FRET efficiency; Fig. 3A, 3B, Jurkat cells Gts-Nef).
These results show that HIV-1 Nef associates in situ Rab11b and MAL, and that this association is supported by Nef traffic properties.
Nef co-opts the switchlike regulatory function of synaptic Lck in promoting the generation of LAT signaling nanoclusters
T cell activation relies on the generation of an amplification signalosome formed by direct association of p-LAT– and p-SLP76–Gads complexes (27). Given that Lck synaptic clustering acts as the signal switch required for the formation of LAT signaling territories (8), we investigated whether the vesicular retention would impair the functional architecture of LAT signaling in Nef-expressing cells. We triggered T cell activation by incubating the T cells for 5 or 12 min on coverslips coated with Abs to CD3 and CD28 followed by fixation. We visualized Nef effects on Lck and LAT vesicle release at the immunological synapse by 3D confocal microscopy (Fig. 4, top panels), and we determined Nef alterations to the functional architecture of the synaptic membrane using dual-color dSTORM combined with TIRF (Fig. 4, bottom panels) (28).
As expected, Nef expression or HIV-1 infection prevented the Lck vesicular compartment delivery at the immunological synapse (Fig. 4A, 4B, 4D, top panels). However, regardless of the fact that Nef does not interact with LAT vesicular traffic (Figs. 1, 2), LAT vesicle delivery is also impaired in Nef-expressing cells (Fig. 4A, 4B, 4D, top panels). Regarding the signaling architecture of activated T cells, we observed that both Nef and HIV-1 infection reduced the number and size of signaling active p-LAT nanoclusters at the immunological synapse (Fig. 4B, 4D, 4G, 4H), as well as their capacity to recruit and form signaling complexes with p-SLP76 (Fig. 4B, 4D, 4I, 4J). All of these defects were rescued when T cells were infected with a Nef-deleted virus (Fig. 4E–J, HIV-ΔNef), but not by stimulating HIV-1–infected cells for longer (Fig. 4F–J, 12′ HIV). This indicates that the signaling deficit in HIV-1–infected cells is unlikely to be a manifestation of delayed kinetics.
These data suggest that, by retaining Lck in a vesicular compartment, Nef impairs the vesicle fusion positive feedback loop that drives the formation of p-LAT nanoclusters at the immunological synapse. To formally test this hypothesis, we recovered Lck clustering at the immunological synapse of Nef-expressing cells by coexpressing the chimeric FynMA-Lck protein whose traffic is not affected by Nef (Fig. 1C, 1D). FynMA-Lck expression completely recovered LAT vesicle delivery (Fig. 4C, top panels), the number and size of p-LAT nanoclusters, as well as their ability to recruit p-SLP76 (Fig. 4C, 4G–J).
Our results collectively demonstrate that, by retaining Lck in a vesicular compartment, Nef impairs the vesicle fusion positive feedback loop that drives the formation of p-LAT nanoclusters and the subsequent recruitment of p-SLP76.
HIV-1 Nef impairs the formation of the Ca2+ membrane territories
Given that Lck clustering might participate in organizing TCR signaling (8, 29) possibly via local Ca2+ fluxes (8, 30), we investigated the impact of HIV-1 Nef on the formation of Ca2+ domains after TCR engagement. We reasoned that if membrane-localized Ca2+ are indeed determinant for the signaling architecture of the immunological synapse, the formation of Ca2+ territories should be detectable in the TCR-stimulated membrane. To pursue our hypothesis, we constructed a membrane-tethered, genetically encoded Ca2+-sensor Fyn-GCaMP3, whose membrane localization could not be altered by HIV-1 Nef (11). Our experimental setup enabled us to directly detect the formation of Ca2+ territories at the synaptic membrane, by measuring the increase in Fyn-GCaMP3 fluorescence intensity. TCR engagement, by incubating the T cells on coverslips coated with Abs to CD3 and CD28, promoted the formation of Ca2+ territories at the membrane of control and HIV-ΔNef–infected cells T cells, but not of Nef-expressing or HIV-infected cells (Fig. 5A, 5B). Interestingly, coexpressing FynMA-Lck and Nef reverted the formation of Ca2+ territories at the stimulatory membrane, indicating that Lck signaling at the immunological synapse is required for the formation of Ca2+ territories. Finally, Nef-expressing cells exhibited no intrinsic functional defect in the Ca2+ channels, because treatment with Ca2+ flux–inducing drugs TPS and ionomycin restored the formation of Ca2+ territories at the synaptic membrane of Nef-expressing cells and also induced similar cytosolic levels of Ca2+ influx, respectively (Fig. 5A–C).
These data indicate that HIV-1 impairs the formation of Ca2+ territories at the synaptic membrane by removing Lck from the immunological synapse.
HIV-1 Nef impairs the formation of calcium territories orchestrating the signaling nanoarchitecture at the synaptic membrane
Finally, to assess the direct impact of Nef on the orchestration of TCR signaling nanoarchitecture, we determined the effect of impaired Ca2+ territories on LAT vesicle docking and/or fusion at the synaptic membrane. LAT vesicles are decorated by the exocytic Rab27 and Rab37, by the v-SNARE Ti-VAMP, and by the Ca2+ sensor synaptotagmin-7, which mediate docking and/or fusion of vesicles with target membranes (8, 9, 21, 31). To distinguish between vesicle docking and vesicle fusion, we used an engineered LAT molecule with an extracellular HA tag that allows for surface labeling of LAT (LAT-HA) (9). Population comparison of HA labeling of control and Nef cotransfected T cells by flow cytometry and by confocal microscopy showed that both control and Nef cotransfected T cells expressed equivalent levels of LAT-HA (Fig. 6A) with similar subcellular distribution at the plasma membrane and within the vesicular compartment (Fig. 6B, 6C). Activation on coverslips coated with anti-CD3 and anti-CD28 Abs of control and HIV-ΔNef–infected T cells induced the fusion of the vesicular pool of LAT-HA at the synaptic membrane and the formation of LAT-HA nanoclusters measured by surface labeling (Fig. 6D, 6H), which were indistinguishable in number, size, and density from endogenous p-LAT nanoclusters (Figs. 4, 6D, 6H–J). These results indicate that LAT vesicle fusion is required for the orchestration of a LAT signaling nanoarchitecture. In Nef-expressing or in HIV-infected cells, there was a stark decrease in the number, size, and density of surface LAT-HA, reflecting Nef impairment of LAT vesicle fusion at the synaptic membrane (Fig. 6E, 6G, 6I, 6J). Importantly, restoration of Ca2+ territories by TPS treatment in anti-CD3– and anti-CD28–stimulated, Nef-expressing cells completely recovered LAT-HA nanoclustering in number, size, and density (Fig. 6F, 6I, 6J). These data collectively indicate that HIV-1 exploits the exquisite interconnectivity between Lck signaling and the formation of Ca2+ territories to impair the orchestration of a functional immunological synapse.
Cross-talk among vesicle traffic, Ca2+ gradients, and TCR signaling coordinates T cell activation (7, 32). In this article, we report a previously unknown immune subversion mechanism involving HIV-1 Nef exploitation of the interconnectivity among vesicle traffic, Ca2+ membrane territories, and TCR signaling. This inhibitory mechanism required Nef to co-opt Lck switchlike function driving the formation of Ca2+ membrane territories controlling the vesicular fusion amplification loop impelling LAT nanoclustering at the immunological synapse.
Nef has been characterized as a potent, multifunctional modulator of T cell activation (3). Recent works have shown that the vesicular traffic has a central role in regulating the assembly of signaling complexes at the immunological synapse for T cell activation (7, 32, 33). We propose that Nef has evolved to co-opt the traffic checkpoints regulating the functional nanoarchitecture of the immunological synapse rather than by inhibiting a series of signaling molecules involved in sequential steps of TCR signaling cascade (34). We have found that Nef interacts with Lck traffic regulators Rab11b and MAL, but not with LAT traffic regulators Rab27 and Rab37 (8, 22, 23). These interactions resulted in the selective retention of Lck in an enlarged vesicular compartment, in Nef-expressing cells. Interchanging Lck and LAT trafficking properties impaired Lck retention and reciprocally led to LAT vesicular accumulation in Nef-expressing cells. We used several different experimental approaches to determine whether Nef interacts directly with MAL; however, the results of these experiments were inconclusive. One possible explanation for these inconsistences is that TM proteins containing multiple hydrophobic domains, as is the case of MAL, often possess different tertiary structures and binding affinities when in solution relative to those occurring within a lipid bilayer. Such changes might confound solution-based analysis such as coimmunoprecipitation and far-Western detection, or even render physical interaction between Nef and MAL impossible in these conditions. In contrast, FRET analysis performed under physiologic conditions with MAL in its native conformation within a lipid bilayer yielded that Nef and MAL are found at interacting distances. In fact, we detected Nef association with both Rab11b and MAL in Lck-deficient cells, precluding the possibility that Nef-Rab11b/MAL in situ interaction could be the result of a tripartite association between Lck-MAL/Rab11b-Nef. We found that rerouting of Nef trafficking from the Rab11+ MAL+ recycling compartment to the close by Golgi vesicles completely abolished Nef activity. Our results suggest that Nef-mediated Lck vesicular retention is the result of the specific Nef-membrane association to the Rab11b and MAL trafficking compartment. Consistent with our results, a recent study has found that Nef biological activity, including Lck vesicular accumulation, is supported by the specificity of its MA domain (18). Curiously, an early work on Nef pathogenicity had found that the MA domain of Nef was required for Nef-mediated retention of Lck through the formation of a protein complex (25). In addition, we and others showed that Lck vesicular accumulation does not depend on its SH3-mediated interaction with Nef, requiring only the LckMA (19). It is thus possible that by promoting Lck vesicular accumulation, Nef association with Rab11b and MAL facilitates Nef modulation of Lck kinase activity through interaction with its SH3 domain.
Even though Nef motifs responsible for promoting the internalization and vesicular accumulation of the transferrin receptor (35), CD80/CD86, and MHC class I (36) differ, all of these molecules share the striking commonality of trafficking through a Rab11b compartment. It is tempting to propose that once internalized, the transferrin receptor, CD80/CD86, and MHC class I are prevented to recycle back to the plasma membrane as the consequence of Nef interaction with their common traffic regulator, Rab11b. Thus, Nef interaction with Rab11b reported in this study might allow for a unified view for the molecular mechanisms underpinning the vesicular retention of distinct molecules across distinct cell types (lymphocytes, macrophages, and dendritic cells) and might provide valuable insight into the role of Rab11b in HIV-1 budding (37).
The spatial organization of signaling proteins determines the T cell activation outcome (7–10, 32, 38, 39). Thus, to understand Nef modulation of T cell activation, it is imperative to move from simply identifying the individual TCR signaling molecules targeted by Nef to determining the molecular mechanism underpinning Nef co-opt of the spatial organization of signaling components. Superresolution microscopy has made important contributions toward revealing the molecular mechanisms controlling the signaling nanoarchitecture of the immunological synapse (8–10, 38, 39). By combining dual-color dSTORM with TIRF microscopy, we found that by interfering with the MAL-mediated route of Lck transport, Nef disengages TCR triggering from LAT vesicle fusion and alters LAT signaling nanoarchitecture at the synaptic membrane. Accordingly, Nef expression decreased the number, density, and size of p-LAT nanoclusters. The formation of elongated TCR signaling nanoclusters has been positively correlated with the efficiency of signal transmission (8, 38, 40). Interestingly, we noted a shift in the synaptic signaling pattern of Nef-expressing cells from the elongated nanoclusters, observed in control cells, toward circular p-LAT, which correlated with a decreased ability of p-LAT to recruit p-SLP76 and the consequent impaired assembly of the LAT signalosome. Finally, bypassing Lck vesicular retention was sufficient to rescue p-LAT nanoclustering at the immunological synapse of Nef-expressing cells. This further reinforced that Nef impairment of LAT synaptic delivery is not the consequence of a direct interaction, but rather the consequence of Nef co-opting the switchlike regulatory function of synaptic Lck in promoting the vesicle fusion and the consequent formation of p-LAT nanoclusters.
Lck clustering has been implicated in the organization of TCR signaling (8, 29) possibly via local Ca2+ fluxes (8, 30). Thus, by retaining Lck, Nef must impair the reorganization of the synaptic membrane that allows for the fusion of LAT vesicles. To map the impact of Nef on the formation of Ca2+ domains at the synaptic membrane, we pioneered the use of genetically encoded, membrane-tethered Ca2+ indicators (11), which detects highly localized Ca2+ signals that soluble probes fail to detect, to map the impact of Nef on the formation of Ca2+ domains at the synaptic membrane. Nef impairs the formation of Ca2+ membrane territories at the synaptic membrane even though it has no effect on cytosolic Ca2+ fluxes (12). Once again, bypassing Lck vesicular retention was sufficient to rescue the formation of Ca2+ membrane territories at the TCR-stimulated membrane, in Nef-expressing cells. These results illustrate the exquisite control that Nef exerts on the subcellular distribution of Ca2+ levels. Even though intracellular Ca2+ controls a remarkable breadth of cellular processes, including secretion, motility, and differentiation, local Ca2+ domains are essential to regulate specific signaling events (41–43). Thus, by impairing the formation of Ca2+ territories at the stimulated membrane, Nef can block the feedback loop amplifying TCR signaling and lower T cell sensitivity to Ag without affecting other Ca2+-dependent functions that might be required for HIV-1 propagation. Impairment of Ca2+ membrane territories is possibly involved in Nef inhibition of other cellular processes, such as migration (44, 45) and phagocytosis (46), which might need to be reinterpreted accordingly.
In conclusion, this work shows that Nef has evolved to co-opt the traffic checkpoints regulating the functional nanoarchitecture of the immunological synapse. Obviously, alterations to the regulatory mechanisms controlling T cell signaling thresholds and signal amplification could underpin the establishment of long-term reservoirs. Thus, the provision of such a mechanism is necessary to identify molecular targets for the development of more efficient antiviral therapies.
We thank Nuno Moreno for contributing to setting up the superresolution microscope, Cláudia Campos and Vasco Barreto for help in cloning, and Cláudia Bispo for help in intracellular calcium assay. We thank the National Institutes of Health AIDS Reagent Program for providing the HIV-1 NL4.3 and HIV-1 NL4.3 ΔNef constructs, the Ab against Gag, and the human IL-2.
This work was supported by the Fundação para a Ciência e Tecnologia under the FCT Investigator Programme (H.S.), iNOVA4Health (Grant UID/Multi/04462/2013 to H.S.), the Biotechnology and Biological Sciences Research Council (Grant BB/M022374/1 to R.H.), and the Medical Research Council (Grant MR/K015826/1 to R.H.).
The authors have no financial conflicts of interest.