Rab6 regulates recycling and retrograde trafficking of MR1 molecules

Mucosal-associated invariant T (MAIT) cells are an innate-like T cell subset important in the early response to bacterial and viral lung pathogens. MAIT cells recognize bacterial small molecule metabolites presented on the Class I-like molecule MR1. As with other Class I and Class II molecules, MR1 can likely sample ligands in the intracellular environment through multiple cellular pathways. Rab6, a small GTPase that regulates a number of endosomal trafficking pathways including retrograde transport to the trans-Golgi network (TGN), is involved in the presentation of ligands from Mycobacterium tuberculosis (Mtb) to MAIT cells. The Rab6-mediated trafficking pathway contains endosomal compartments that share features with the Mtb intracellular compartment. Using inducible expression of MR1, this study demonstrates that Rab6 regulates the recycling of MR1 molecules from the cell surface through endosomal trafficking compartments to the TGN. This Rab6-dependent pool of recycled MR1, which is available for reloading with ligands from bacterial pathogens like Mtb, may be important for early recognition of infected cells by MAIT cells in the lung.


Scientific Reports
| (2020) 10:20778 | https://doi.org/10.1038/s41598-020-77563-4 www.nature.com/scientificreports/ However, there are additional pathways governing the loading of bacterially derived ligands that are distinct from loading of exogenously added synthetic ligands and make use of endosomal MR1 21,24 . In a subsequent report, we identified 6-FP-loaded MR1 as permissive for ligand exchange, providing a stabilized pool of MR1 for loading of exogenous ligands in airway epithelial cells 25 . We previously showed that endosome-mediated pathways are utilized in the presentation of ligands from Mtb on MR1 21 , suggesting a role of endocytosis and recycling of MR1 in the presentation of ligands from intracellular microbes. This current study highlights the role of the small GTPase Rab6 in distinct pathways of MR1-dependent antigen presentation. We generated a bronchial epithelial cell line with inducible expression of MR1 to explore the trafficking of newly synthesized or preexisting MR1 proteins. This control of MR1 synthesis enabled deeper understanding of the differential pathways for ER-residing or endosomal compartment-residing MR1, respectively, to translocate to the cell surface for antigen presentation. We demonstrate here that Rab6 regulates the recycling and retrograde trafficking of MR1 through endosomal pathways, but does not impact the translocation of MR1 to the cell surface. Our data is also consistent with a model where MR1 residing in non-ER compartments is available for loading with new ligand. The intracellular localization of Mtb in the endosomal trafficking pathway suggests Mtb ligands can access MR1 through this Rab6-mediated pathway.

Results
MR1 expression kinetics using an inducible promoter. Silencing Rab6 in the bronchial epithelial cell line BEAS-2B resulted in reduced MAIT cell responses to Mtb infection 21 . We previously demonstrated an increase in the number of MR1 + endosomal vesicles in Rab6-silenced cells; however, the translocation of MR1 to the cell surface with 6-FP treatment was not affected 21 . This study used cells constitutively over-expressing MR1, which made it difficult to ascertain how Rab6 silencing specifically impacted MR1 expression and function in different cellular compartments. Additionally, although MAIT cell responses in ELISPOT assays clearly demonstrate that MR1 molecules loaded with Mtb ligands reach the cell surface in BEAS-2B cells, there are no measurable changes in the cellular localization and surface translocation of MR1 in the context of Mtb infection using flow cytometry or fluorescence microscopy, consistent with the findings of others 22 . In contrast to the mechanisms that have been identified for loading of exogenously added ligands such as 6-FP and 5-OP-RU 22 , this has created a challenge in identifying mechanisms for the loading of bacterial ligands generated during intracellular infection. To address these concerns, we generated a cell line with inducible MR1 expression. Using a previously described construct 25 , BEAS-2B cells were stably transduced to express MR1 fused to GFP behind a doxycycline (doxy)-inducible promoter. Consistent with transient transfection and induction of the doxyMR1-GFP construct 25 , addition of doxy to the culture media resulted in MR1-GFP expression in the ER and endosomal compartments and translocation of MR1 to the cell surface upon addition of 6-FP (Fig. 1A). Additionally, MAIT cell response to doxy-treated BEAS-2B:doxyMR1-GFP cells stimulated with Mycobacterium smegmatis supernatant was increased compared to BEAS-2B:doxyMR1-GFP cells not treated with doxy (expressing wildtype levels of MR1) (Fig. 1B), which confirms that the doxy-inducible MR1-GFP construct is capable of antigen presentation and activation of MAIT cells. Together these results suggest the MR1 in BEAS-2B:doxMR1-GFP cells traffics and functions similarly to endogenous and constitutively over-expressed MR1.
RT-PCR analysis of MR1 gene expression following addition and subsequent removal of the doxy was performed to determine the kinetics of MR1 overexpression in these cells. Increased MR1 expression peaked at 16-24 h following doxy addition, and remained at these levels for at least an additional 24 h ( Fig. 2A, left). Removing doxy by washing cells and replacing the media resulted in a decrease in MR1 gene expression, returning to near pre-doxy levels by 16-24 h post wash ( Fig. 2A, right). Analysis of these cells by flow cytometry and fluorescence microscopy revealed the kinetics of MR1-GFP protein expression. Flow cytometry demonstrated a substantial decrease in total cellular MR1-GFP protein expression 24 h after washing doxy from the media (Fig. 2B), mirroring the RT-PCR results.
BEAS-2B:doxyMR1-GFP cells treated with doxy were imaged in parallel to the RNA expression experiments in Fig. 2A to determine the intracellular localization of MR1 (Fig. 2C, top) and in some cases were treated with 6-FP to determine the surface translocation of MR1 (Fig. 2C, bottom). Images were analyzed using Imaris to quantify GFP endosomal compartments. Early after doxy addition (8 h), MR1 localized predominantly in the ER. By 24 h following doxy treatment, MR1 was observed in post-ER endosomal compartments (p = 0.001) and remained in the ER to a lesser extent (Fig. 2C,D, top, 24 h). Similar to the results from Fig. 2B, MR1-GFP signal was dimmer after washing doxy from the cells for 12 and 24 h (36 and 48 h, top, p < 0.001) and there were fewer MR1 + endosomal compartments per cell (36 and 48 h, p < 0.001 and p = 0.006). In cells treated with 6-FP, MR1 was observed on the cell surface at all three timepoints with little change in the overall number of MR1 + endosomal compartments per cell (Fig. 2C bottom, D). Previous evidence suggests that 6-FP protects MR1 from degradation 25 , which may explain the increase in MR1 fluorescence in 6-FP-treated cells 12 h after washing doxy (Fig. 2C,D, 36 h, p < 0.001). In contrast, the number and brightness of MR1 + endosomal compartments peaks at 24 h after addition of doxy in cells not treated with 6-FP. There are multiple possibilities for this pattern, but the observation confirmed that in our cells, MR1 translocates from the ER to endosome-like cellular compartments following initial synthesis without the addition of exogenous ligand. After 6-FP-driven surface translocation and subsequent removal of ligand, MR1 is capable of recycling into multiple cellular regions including the ER and endosomal compartments.
Taken together, these data validate that the BEAS-2B:doxyMR1-GFP cells can be used to analyze the impact of Rab6 on MR1 surface translocation in the context of "newly synthesized" versus "preexisting" MR1 ( Fig. 2E). Specifically, by manipulating Rab6 expression or function prior to the addition of doxy to the media, we can analyze the impact of Rab6 on newly synthesized, predominantly ER-localized MR1 (Fig. 2E,  www.nature.com/scientificreports/ we could analyze the impact of Rab6 on preexisting MR1 residing in endosomal compartments and on the cell surface ( Fig. 2E, right). The temporal control afforded by the BEAS-2B:doxyMR1-GFP cells is key to exploring if Rab6 functions in the translocation of antigen-bound MR1 from the ER to the cell surface (newly synthesized timing) or in the cycling of antigen-bound MR1 between endosomal compartments and the cell surface (preexisting timing).
Rab6 regulates total and surface expression of preexisting MR1. We examined the expression of total cellular or surface MR1 in Rab6-silenced BEAS-2B:doxyMR1-GFP cells by flow cytometry using the timings from Fig. 2E. In cells where Rab6 was silenced prior to inducing MR1 expression (i.e. "newly synthesized" MR1) with no addition of exogenous ligand, there were no differences in total MR1 as measured by GFP signal, or surface MR1 as measured by surface staining with the α-MR1 26.5 antibody (Fig. 3A). In cells expressing preexisting MR1, there was an increase in total MR1 in Rab6-silenced cells at steady state (Fig. 3B). Although the increase was not statistically significant, it was repeatable and not observed in cells expressing newly synthesized MR1. Additionally, it was consistent with our previous results using a system with constitutive overexpression of MR1 21 . There was no corresponding surface MR1 increase in Rab6-silenced cells expressing preexisting MR1 (Fig. 3A,B). Since Rab6 silencing did not impact expression of or translocation of newly synthesized MR1 to the cell surface, we hypothesized that the observed changes resulting from Rab6 silencing impacted preexisting MR1. We next examined whether Rab6 silencing affected total and surface MR1 expression in the context of the exogenously added ligand, 6-FP. First, we silenced Rab6 before doxy induction of MR1 and incubation with 6-FP to look at the impact of silencing Rab6 on newly synthesized MR1. Similar to our previous observations in Fig. 2C,D and in 25 , cells that were incubated with 6-FP showed an increase in total MR1 which suggests that 6-FP stabilizes MR1 molecules (Fig. 3A). Addition of 6-FP to the culture equally increased total expression and surface translocation of MR1 in Rab6-silenced cells compared to the missense control (Fig. 3A). These data demonstrate that Rab6 does not regulate loading or egress of newly synthesized MR1 from the ER in the context of 6-FP. Next, we silenced Rab6 in cells containing preexisting MR1, then added 6-FP. Similar to the steady state preexisting MR1 condition, there was a small but consistent increase in total MR1 and no increase in surface stabilized MR1 in Rab6-silenced cells compared to missense control when cells were incubated with 6-FP (Fig. 3B). Taken together, these data suggest that Rab6 could play a role in MR1 trafficking, but not at the www.nature.com/scientificreports/ level of translocation to the cell surface. Furthermore, these data provide supporting evidence that extracellular ligands like 6-FP can be loaded onto MR1 that has not been newly synthesized 25 .
Rab6 regulates MR1 recycling from the cell surface. Because we observed small but statistically insignificant changes to total cellular MR1 or surface MR1 by flow cytometry, we next used fluorescence microscopy to observe whether there were any changes to the intracellular localization of MR1 in Rab6-silenced cells.
In cells expressing newly synthesized or preexisting MR1-GFP, we quantified the number of endosome-like compartments containing MR1 and the relative amount of MR1 in these structures. As we previously showed in cells constitutively overexpressing MR1-GFP 21 , we observed an increase in the number of MR1 + endosomes in Rab6-silenced cells expressing newly synthesized MR1 (Fig. 4A). Interestingly, there was significantly less MR1-GFP in each endosome (p = 0.003) and the endosomes were smaller (p < 0.001) in this condition. In the context of 6-FP, the same differences between control and Rab6-silenced cells was observed, but to a lesser extent (Fig. 4B). This is more apparent in cells incubated without 6-FP, since translocation from the cell surface was not impacted by the presence of an exogenous antigen. Together, these results support the hypothesis that Rab6 impacts endosomal MR1 trafficking rather than surface translocation of MR1 from the ER. In cells expressing preexisting MR1, the number of MR1 + endosomal compartments per cell in control and Rab6-silenced cells was not significantly different at steady state or after incubation with 6-FP (Fig. 4B). Therefore, as above, we examined the relative amount of MR1-GFP fluorescence signal in these compartments. At both steady state and in the context of 6FP treatment, we observed a different distribution in the MR1-GFP signal among endosomal compartments (Fig. 4B), as evidenced by the multimodal distribution of MR1-GFP signal in control compared to Rab6-silenced cells. Analysis of the distribution of the data showed there was a higher mean MR1-GFP signal in endosomal compartments in Rab6-silenced cells (p < 0.001) and the distribution of the MR1-GFP signal was distinct between conditions. In support of this, there was also a significant decrease in the mean volume of MR1 + endosomal compartments in Rab6-silenced cells, with or without the addition of 6-FP (Fig. 4C, p < 0.001). This again suggested an impairment in recycling of endocytosed MR1 or in endocytosis of MR1 following surface stabilization in Rab6-silenced cells, such that MR1 follows different endosomal trafficking pathways resulting in distinct intracellular distribution.
Rab6 regulates transport of MR1 molecules to the TGN. To further understand how Rab6 mediates the recycling of MR1, we generated wild-type (WT) and a mutant (Q72L) Rab6 protein fused to RFP in BEAS-2B cells. The Rab6-Q72L mutant is GTP-locked and constitutively active 26 . As expected, Rab6-WT localized mainly to the Golgi, in areas where it colocalized with the Golgi markers golgin-97, p230, and GM130 (Fig. 5, left). Rab6-Q72L also localized to an area near the Golgi, but in a more dense and tight structure that was largely distinct from golgin-97, p230, and GM130 (Fig. 5, right).The structural change to the localization of Rab6-WT versus Rab6-Q72L was reflected in a significantly different volume of cellular Rab6 (Fig. 5, top), as well as the need to acquire images at different settings to achieve similar emission values to avoid saturation for deconvolution and analysis purposes (Rab6-WT: 100% transmission, 0.05-0.10 s; Rab6-Q72L: 32% transmission, 0.025-0.15 s).
These Rab6 constructs were then expressed in the BEAS-2B:doxyMR1-GFP cells expressing newly synthesized or preexisting MR1. When compared with the missense control in Fig. 4A,B, expression of Rab6-WT did not impact the localization of newly synthesized or preexisting MR1 (Fig. 6a,b). In general, the distance between MR1 and the center of both Rab6-RFP masses has a bimodal distribution, which may illustrate MR1 residing in distinct endosomal compartments. In cells treated with 6-FP, more newly synthesized MR1 is located farther from the center of the Rab6-Q72L mass than of the Rab6-WT mass (Fig. 6a, p < 0.001). Although not significant, a similar pattern exists at steady state, where the Rab6-Q72L cells have a larger population of distant MR1 than Rab6-WT cells. Put together, these data suggest that constitutive Rab6 action alters the transport of newly synthesized MR1 residing in post-ER endosomal compartments. Preexisting MR1 is more distant from Rab6-Q72L than Rab6-WT in cells at steady state (p < 0.001); however, there is no significant difference seen in cells treated with exogenous ligand. Together, these data suggest that Rab6 functions in recycling and retrograde transport of MR1 from the cell surface through endosomal pathways to the trans-Golgi network.    25 . Interestingly, ligands from live or fixed E. coli are also presented via distinct pathways 24 . Together, these data support the concept that MR1 can use multiple pathways to sample ligands deriving from extracellular and intracellular environments. Here, we sought to better understand the role of Rab6 in defining the MR1 antigen presentation pathway for Mycobacterium ligands, using inducible MR1 expression and 6-FP as a surrogate ligand. We previously demonstrated that silencing of Rab6 resulted in decreased presentation of Mtb ligands by WT BEAS-2B cells to MAIT cells but did not affect translocation of MR1 to the cell surface after addition of 6-FP in cells overexpressing MR1, suggesting that Mtb ligands can be loaded and presented via a different pathway than 6-FP 21 . To assess the role of Rab6 in trafficking of newly synthesized versus preexisting MR1 molecules, we generated a cell line that stably expresses MR1 behind a doxycycline-inducible promoter. By using inducible expression of MR1, we could specifically observe the impact of Rab6 silencing on newly synthesized MR1 molecules versus those already made. While consistent after removal of doxy, the decline in the level of MR1 transcripts was not as rapid as we had anticipated. The limited time between the disappearance of MR1 transcript and the loss of detectable MR1-GFP protein suggests that MR1 rapidly transits to the surface, through endosomal recycling pathways, and ultimately to degradation. This relationship between gene and protein expression resulted in challenges in analyzing the effect of Rab6 knockdown on MR1 trafficking. We had to balance the time required to achieve Rab6 silencing with the time it took to increase the probability that the observed MR1 was preexisting and not newly formed, while also maintaining the ability to detect remaining MR1 protein before degradation. Thus, there was a reduced ability to detect significant differences between the Rab6-silenced and control cells in some cases. Nonetheless, because we were able to compare the effects of Rab6 silencing on preexisting MR1 molecules to those that we knew were newly synthesized, resulting in confidence that the data show the biologically relevant impacts of Rab6 silencing.
Our data best support a function for Rab6 in retrograde transport of MR1 from the plasma membrane to the TGN. The phenotype observed in Rab6-silenced cells could be explained by several possible Rab6 functions within this pathway. First, Rab6 could be directly involved in endocytosis from the cell surface, thus when Rab6  www.nature.com/scientificreports/ surface translocation (upon treatment with 6-FP) erases this effect. For cells with newly synthesized MR1, this population is only visible after treatment with 6-FP translocates MR1 from the ER to the surface. Together, these findings suggest that Rab6 functions in retrograde trafficking of surface or endosomal MR1. In BEAS-2B cells overexpressing MR1, much of the intracellular MR1 is observed in compartments expressing late endosomal and lysosomal markers such as Rab7 and Lamp1 21 . In contrast, in the C1R lymphoblast cell line overexpressing MR1, a population of MR1 is observed with early endosomal markers like EEA1 in addition to late endosomal markers 22 , so the point in the endosomal pathway where ligands are loaded may be cell type dependent. In our model, Rab6 activity would direct MR1 into a pathway where ligand exchange and recycling to the cell surface can occur. When Rab6 is silenced, MR1 would be directed into compartments where it is unavailable for ligand exchange and recycling to the cell surface. More work will be necessary to determine the phenotype of these different subsets of endosomal compartments observed in our Rab6-silenced cells. This model suggests that microbially-derived MR1 ligands are loaded onto MR1 earlier in endocytosis, prior to trafficking of bacteria to lysosomes for degradation. In the case of Mtb, which inhibits fusion of the phagosome with lysosomes and continues to be metabolically active within the host cell, it might suggest a unique importance for ongoing presentation of Mtb ligands and MR1-dependent MAIT cell responses. Together, our data and that of our previous work 21, 25 are consistent with a model in which exchange of MR1 ligands, including 6-FP and those derived from bacterial infection, can occur in endosomal trafficking pathways, some of which are Rab6-mediated. This is distinct from what has been seen in C1R lymphoblasts 22 and suggests the possibility of cell-specific MR1 trafficking and loading pathways for ligands derived exogenously and during intracellular infection. Further exploration of MR1 ligand exchange in epithelial cells and professional antigen presenting cells may reveal the differential physiological role of MR1 presentation in these cell types. The significance of these different pathways in the context of mucosal tissues during bacterial infections remains unknown. Furthermore, perturbation of pathways that impact endosomal recycling of MR1 may also impact the internalization and trafficking of bacteria taken up by the host cell. The impact of silencing trafficking molecules such as Rab6 on the availability of MR1 ligands has yet to be explored. It is likely that the pathways www.nature.com/scientificreports/ for MR1 antigen presentation are not mutually exclusive. Ligands from different intracellular and exogenous sources may be preferentially, but not exclusively, presented through these pathways. Continued understanding of the mechanisms for presenting MR1 ligands will be critical to understanding how MAIT cell activation is regulated to prevent inappropriate inflammatory responses in mucosal tissues, and also to better target MAIT cells for therapeutics and vaccination. Cells and reagents. BEAS-2B bronchial epithelial cells were obtained from the American Type Culture

Methods
Collection (ATCC CRL-9609) and cultured in DMEM media (Gibco) supplemented with L-glutamine and 10% heat-inactivated fetal bovine serum. The MR1-restricted T cell clone D426G11 was generated and expanded as previously described 2,4 , then cultured in RPMI (Gibco) supplemented with L-glutamine and 10% heat-inactivated human serum. Mycobacterium smegmatis Mc 2 155 (ATCC) was cultured in 7H9 broth. Following growth to late log phase, the bacteria were pelleted and the supernatant was passed over a 0.22 micron filter for use as antigen in ELISPOT assays.

Lentivirus production and generation of stable BEAS-2B:doxMR1-GFP cell line. BEAS-2B cells
were stably transduced to express MR1 fused to GFP under a doxycycline (dox)-inducible promotor (BEAS-2B:doxMR1-GFP) as previously described 27 . Briefly, low passage HEK 293T cells (ATCC CRS-3216) were cotransfected with the packaging plasmids (psPAXs and pMD2.G) and the pDMLV2.1:doxMR1 construct using Lipofectamine 3000 (Invitrogen). pDMLV2.1:doxMR1 was generated by subcloning the doxMR1-GFP cassette from the pCI:doxMR1 construct used previously for transient transfection 25 into the pLenti6.2 plasmid. After 48 h, the medium was removed and passed through a 0.45 micron filter to clear cellular debris. The filtrate was then mixed with an equal volume of DMEM containing 8 μg/mL polybrene (Sigma) and used to transduce wildtype (WT) BEAS-2B cells. Cells were cultured for four days in the presence of 2 μg/mL doxy, then sorted using a BD Influx cell sorter on GFP expression. Following cell sorting and culture, the inducible expression of MR1-GFP was validated by flow cytometry and fluorescence microscopy.
Briefly, WT BEAS-2B or BEAS-2B:doxMR1-GFP cells (5e3 per well) were used as antigen presenting cells and incubated with a titration of M. smegmatis supernatant as the antigen in an ELISPOT assay. IFN-γ production by the MAIT cell clone D426G11 (5e3 per well) was used as a readout for MAIT cell activation. siRNA silencing. BEAS-2B:doxMR1-GFP cells were plated in 6-well tissue culture plates or 1.5 mm glassbottom chamber slides at 70% confluency and transfected with 5 nM siRNA (ThermoFisher Scienctific) using HiPerFect Transfection Reagent (Qiagen). Cells were incubated for indicated times prior to use and analysis in assay. A missense siRNA was used as a control in all experiments and gene silencing was validated by RT-PCR.

Real-time quantitative PCR (RT-PCR).
RNA isolation, cDNA synthesis and RT-PCR were performed as previously described 21,25 using TaqMan gene expression assays for Rab6 (Hs01042278_m1) and GAPDH (Hs02758991_g1) (LifeTechnologies). All reactions were run in triplicate. Expression data were normalized to GAPDH and calculation of the relative Rab6 expression level was determined using the 2 −ΔΔCT method 29 .
MR1 surface stabilization assay. BEAS-2B:doxMR1-GFP cells were plated in 6-or 12-well tissue culture plates, and treated with siRNA and doxy as indicated. Following a 16 h incubation with 100 μM 6-FP, cells were harvested and stained on ice with α-MR1-PE or -APC (26.5) for 40 min in the presence of 2% human serum, 2% goat serum, and 0.5% FBS. Cells were washed and analyzed live with a Beckman Coulter CytoflexS. All analyses were performed using FlowJo10 (TreeStar). www.nature.com/scientificreports/ and program G-016 on the Lonza Nucleofector 2b. Transfection efficiency was assessed by flow cytometry and fluorescence microscopy.
Fluorescence microscopy. BEAS-2B cells were seeded in #1.5 glass bottom chamber slides (Nunc) and treated with reagents as indicated. Live cell images were acquired at 37 °C with 5% CO 2 . For antibody staining, cells were washed, fixed with 4% paraformaldehyde, and permeabilized with 0.05% saponin for 30 min prior to antibody staining. Live and fixed cell images were acquired using a high-resolution wide-field CoreDV microscope (Applied Precision) with CoolSNAP ES2 HQ (Nikon). Images were taken in Z stacks in a 1024 × 1024 format using a 60 × Plan Apo N objective (NA 1.42). The researcher was blind to the cell treatment condition. For fixed cell imaging analysis of MR1 localization, cells were selected in an unbiased manner using DAPI staining. For live cell imaging analysis of MR1 localization, cells were selected in an unbiased manner using Rab6 expression. An iterative algorithm was used to deconvolve the images using an optical transfer function of 10 iterations (Softworx, Applied Precision).
Data analysis. All data were analyzed using Prism 8 (GraphPad). The Shapiro-Wilk test was used to assess normality, and the Mann-Whitney test was used to determine statistical significance. All images were analyzed using Imaris (Bitplane) as in Harriff et al. 21 except where noted. Briefly, local regions of high fluorescence intensity were algorithmically identified with the "Spots" module and the relative fluorescence intensity was averaged across each region. Rab6-RFP construct regions were similarly analyzed with the "Surfaces" module. Microscope and Imaris settings were kept constant to allow comparison between conditions and the condition of each image was blinded during analysis to prevent bias. Relative fluorescence intensity histograms were created in ImageJ 2.1.0 over linear regions of interest. Raincloud plots were created using R 4.0.0 30,31 as previously described 32 .