Enhanced and selective translation expands the lysosome size and promotes antigen presentation during phagocyte activation

The mechanisms that govern organelle remodeling remain poorly defined. Lysosomes degrade cargo from various routes including endocytosis, phagocytosis and autophagy. For phagocytes, lysosomes are a kingpin organelle since they are essential to kill pathogens and process and present antigens. During phagocyte activation, lysosomes undergo a striking reorganization, changing from dozens of globular structures to a tubular network, in a process that requires the phosphatidylinositol-3-kinase-AKT-mTOR signalling pathway. Here, we show that lysosomes undergo a remarkably rapid expansion in volume and holding capacity during phagocyte activation. Lysosome expansion was paralleled by an increase in lysosomal protein levels, but this was unexpectedly independent of TFEB and TFE3 transcription factors, known to scale up lysosome biogenesis. Instead, we demonstrate a hitherto unappreciated mechanism of acute organelle expansion via mTORC1-dependent increase in translation of mRNAs encoding key lysosomal proteins including LAMP1 and V-ATPase subunits, which is necessary for efficient and rapid antigen presentation by dendritic cells. Collectively, we demonstrate a previously unknown and functionally relevant mechanism of rapid lysosome expansion, whereby mTORC1 senses stimuli to remodel lysosomes via modulating translation of mRNAs encoding key lysosomal proteins.


Introduction
Eukaryotic cells compartmentalize a wide-range of biochemical functions within membranebound organelles such as the endoplasmic reticulum, peroxisomes and lysosomes. These organelles can exist in disparate morphologies ranging from individual vesicular organelles, stacks of flattened membrane sacs, to a continuous membrane reticulum. Despite this complexity, cells must control organelle number, size, shape and activity to meet the needs of their differentiation state. In addition, cells must adapt these organellar properties in response to intrinsic and extrinsic stimuli that alter the metabolic and functional needs of cells [1][2][3][4][5] . Yet, how cells determine organellar properties in response to differentiation state and/or changes in their environment remains one of the most outstanding questions in cell biology.
Immune cells like macrophages and dendritic cells are highly plastic inasmuch as they can adopt "resting", highly inflammatory, and anti-inflammatory states that differ in their gene expression profile, metabolic programming, secretory pathway activity and endolysosomal membrane system [6][7][8][9][10] . With respect to the endolysosomal system, mature dendritic cells abate the degradative capacity of their lysosomal system, which is commonly referred to as the MHC-II compartment, to help preserve antigenic peptides for presentation to adaptive immune cells 11 . On the other hand, macrophages enhance their lysosomal degradative power after phagocytosis to enhance bacterial killing 12 . Perhaps, one of the most striking examples of endolysosomal remodelling occurs during lipopolysaccharide (LPS)-activation of macrophages and dendritic cells, which transform lysosomes from a collection of dozens of individual globular organelles into a striking tubular network 10,13,14 . This reorganization requires downstream TLR4 signals including the phosphatidylinositol 3-kinase-AKT-mTOR axis, which may interface with the lysosomal Rab7 and Arl8b GTPases to control lysosome association with microtubule-motor proteins 14,15 . These motors then help distort and tubulate lysosomes on microtubule tracks 13,16,17 .
While lysosome tubulation is associated with retention of pinocytic cargo, exchange of phagosomal cargo, and possibly antigen presentation [18][19][20][21][22] , it is not presently known how lysosome tubulation helps phagocytes perform their function in response to LPS and other stimulants.
Lysosomes themselves serve as signaling platforms to sense the metabolic and nutrient state of the cell [23][24][25] . For instance, a protein network involving the V-ATPase, Ragulator and Rag GTPases sense high levels of amino acids within lysosomes to activate mTORC1 on the lysosome surface [26][27][28][29][30][31] . Active mTORC1 then phosphorylates various downstream targets to stimulate occupied a larger volume than punctate lysosomes in resting cells (Fig. 1a). To test this, we quantified the total lysosome volume in LPS-activated and resting cells by employing image volumetric analysis 46,47 . We first pre-labelled lysosomes (see materials and methods) with a fluorescent fluid-phase marker and then exposed cells to LPS or vehicle-alone for 2 h to induce lysosome remodelling (Fig. 1a). Pre-labelling cells prior to stimulation ensured that lysosomes are equally loaded with the dye in both resting and activated cells. We then employed live-cell spinning disc confocal microscopy to acquire z-stacks and undertake volumetric analysis. Using this methodology, we observed a significant increase in volume occupied by the fluorescent probe in LPS-stimulated RAW macrophages, bone marrow-derived macrophages (BMDM) and bone marrow-derived dendritic cells (BMDCs) relative to their resting counterparts (Fig. 1b). This suggests that activated phagocytes have an expanded total lysosome volume relative to resting cells.
We previously demonstrated that in RAW macrophages, lysosome tubules were more mobile than punctate lysosomes 13 . Thus, to exclude the possibility that the increase in lysosome volume was due to a trailblazing effect during Z-stack image acquisition, we sought to estimate lysosome volume in fixed cells. However, typical fixation protocols with 4% PFA causes tubular lysosomes to disintegrate (Fig. S1a, b). To circumvent this issue, we developed a fixation procedure that preserves lysosome tubules in macrophages (Fig. S1a, b). Re-applying volumetric analysis to fixed RAW cells, we still observed a significant increase in lysosome volume in activated cells relative to resting phagocytes (Fig. 1c). Finally, to exclude that the increase in lysosome volume is an artifact of the limit of resolution of spinning disc confocal microscopy, we employed structured illumination microscopy (SIM) which enables imaging in super-resolution 48 .
Due to limitations of the available SIM system, we sampled three x-y planes centred at the midpoint of cells and quantified the area occupied by the fluid-phase marker (Fig. S1c). This approach also demonstrated a significant increase in label volume in activated RAW, primary macrophages, and BMDCs relative to their resting counterparts (Fig. 1d). These data demonstrate that the lysosome volume expands in response to macrophage and dendritic cell stimulation, concurrent with tubulation.

Phagocyte activation increases lysosomal holding capacity
An expanded lysosome volume may as a corollary lead to a boost in the storage capacity of lysosomes. Hence, we assessed whether activated phagocytes have a higher lysosomal holding capacity relative to resting cells by allowing cells to internalize fluorescent pinocytic tracers to saturation. Indeed, both primary and RAW macrophages pre-activated with LPS exhibited a large increase in fluid-phase accumulation relative to their resting counterparts at each time point examined (Fig. 2a, b; Fig. S2a). We also observed that pre-activated primary macrophages displayed faster rates of pinocytic uptake relative to resting macrophages (Fig. 2c). In fact, the rate of pinocytic uptake was augmented within 15 min of LPS exposure as indicated by macrophages concurrently undergoing pinocytosis and stimulation (Fig. 2c). In comparison, we showed that resting and activated primary macrophages did not differ significantly in the rate of depletion of the pinocytic tracer (Fig. 2d), suggesting that exocytosis rates were similar. RAW macrophages exhibited slightly different dynamics in that the rates of uptake and retention were similar between resting and LPS-stimulated cells (Fig. S2b, S2c). Collectively, these data indicate that activated macrophages have a higher lysosome holding capacity relative to resting macrophages. Lastly, we questioned whether dendritic cells would benefit from an increase in lysosome volume since they are reported to arrest endocytosis after maturation 49,50 . Of note, most reports examine dendritic cell function over 16 h post-stimulation, while recent work shows that mature cells can still endocytose extracellular cargo 51-53 . Importantly, we show here that dendritic cells retained their pinocytic capacity up to 8 h post-activation, which fits the timeline of lysosome reorganization and expansion (Fig. S2d). This suggests that rapidly expanding the lysosome volume may help dendritic cells accumulate more pinocytic content including antigenic material.
Overall, our observations are consistent with previous reports suggesting that tubulation in activated macrophages may aid in retaining fluid phase and that mature dendritic cells continue to engulf extracellular material 18,51,52,54 .

Activated macrophages express higher levels of lysosomal proteins
Thus far, the data presented here suggest that activated phagocytes rapidly expand their lysosome volume and retention capacity. Though other mechanisms like increased endosomal membrane influx may contribute to this, we postulated that lysosomal biosynthesis may be a significant driver of lysosome expansion during phagocyte activation. To address this hypothesis, we determined the levels of select lysosomal proteins, namely LAMP1, V-ATPase subunits and cathepsin D by Western blotting in resting and activated primary macrophages. Specifically, we compared resting macrophages to those continuously exposed to LPS for 2 h or 6 h or for 2 h with LPS followed by a 4 h chase with no LPS. In all cases, LPS induced approximately 2-fold induction of LAMP1 and V-ATPase subunits H and D protein as compared to resting macrophages (Fig. 3a, b). In contrast, cathepsin D levels were unchanged (Fig. 3a, b), suggesting that LPS exerts differential effect on lysosomal proteins. The increase in LAMP1, ATP6V1H and ATP6V1D was blunted by the translation elongation inhibitor, cycloheximide, indicating that de novo protein synthesis, rather than lower protein turnover, augments the levels of lysosomal proteins in LPS-treated phagocytes ( Fig. 3a, b). Importantly, cycloheximide blunted lysosome tubulation and expansion in macrophages in response to LPS (Fig. 3c, d). Overall, our data support a role for de novo protein synthesis in remodelling and expanding the lysosome network during phagocyte activation.

Rapid lysosome expansion is not dependent on TFEB and TFE3
Our results suggest that biosynthesis plays a major role in LPS-induced lysosome expansion in macrophages. Activation of TFEB and TFE3 transcription factors drives expression of lysosomal genes thereby stimulating lysosome function under various stresses including starvation, phagocytosis, protein aggregation and macrophage activation 12,42,[55][56][57][58][59][60][61] . Thus, we next investigated whether the observed rapid lysosome expansion was driven by TFEB-and TFE3-mediated transcriptional upregulation of lysosome genes.
To assess activation of TFEB and TFE3, we quantified nuclear translocation by determining the nucleocytoplasmic ratio of endogenously expressed proteins by immunofluorescence 12,62 . As expected, resting cells exhibited mostly cytoplasmic TFEB and TFE3, whereas inhibition of mTOR for 1 h with torin1 caused both proteins to translocate into the nucleus (Fig. 4a, b). Strikingly, while 2 h incubation with LPS rapidly induced lysosome remodelling and expansion, this did not trigger nuclear translocation of TFEB or TFE3 (Fig. 4a,   b). In comparison, a prolonged 6 h incubation with LPS initiated nuclear entry of these proteins, especially for TFE3 that was comparable to mTOR suppression (Fig. 4a, b). These results are consistent with observations by Pastore et al., who also observed delayed nuclear entry of these proteins in response to LPS-induced macrophage activation 59 . Strikingly, mRNA levels of TFEB and TFE3 target genes (i.e. LAMP1, TRPML1 and two V-ATPase subunits) were not increased even after 6 h of LPS exposure (Fig. 4c), whereas there was massive upregulation of interleukin-6 mRNA (Fig. 4d).
To further exclude the role of TFEB in lysosome expansion during macrophage activation, we measured tubulation and lysosome volume in RAW macrophages deleted for the genes encoding TFEB and/or TFE3 using CRISPR-based technology 59 . Deletion of TFEB and TFE3 did not significantly affect LAMP1 protein levels under resting conditions (Fig. S3a, b) nor fluid-phase marker trafficking, as quantified by Mander's coefficient for dextran-containing LAMP1 signal ( Fig. S3c, d). Moreover, both resting wild-type and deletion strains of RAW macrophages accumulated similar levels of the dextran probe after 1 h of uptake and 1 h chase (Fig. S3e).
Finally, TFEB and TFE3 status in the cell did not exert major influence on retention of the fluidphase probe after 2 h of LPS exposure (Fig. S3f). Collectively, these data suggest that TFEB and/or TFE3 have minimal impact on basal lysosome biogenesis, basal pinocytosis and trafficking to lysosomes.
We next examined remodelling of lysosomes by treating wild-type and TFEB and/or TFE3-deleted RAW cells with LPS for up to 2 h. Importantly, all three mutant cell lines exhibited a significant increase in lysosome tubulation after 2 h of LPS treatment relative to resting condition. This increase in lysosome tubulation in cells depleted of TFEB and/or TFE3 was indistinguishable from that observed in control, TFEB and TFE3 proficient cells (Fig. 4e).
Moreover, LPS-induced expansion of the total lysosome volume was comparable between control and TFEB and/or TFE3-deleted cells (Fig. 4f). Surprisingly, these results suggest that TFEB and/or TFE3-dependent transcription-based program is not a major driver for the rapid lysosome expansion during acute macrophage activation.

Rapid lysosome expansion depends on AKT and mTOR activity
Given that the levels of lysosomal proteins, but not corresponding mRNAs were induced by LPS treatment, we next studied the role of mRNA translation in lysosome expansion. Activated macrophages exhibit extensive metabolic reorganization, enhanced protein synthesis, selective translation of mRNAs encoding inflammatory proteins, and activation of unfolded protein response 6,8,9,63,64 . Consistently, LPS activates mTORC1 in macrophages, which not only stimulates mRNA translation, but is also necessary for lysosome tubulation 15,33,34 . Thus, we tested whether mTOR activity is necessary for enhanced lysosome volume and holding capacity. Indeed, both primary and RAW macrophages exhibited increased phosphorylation of mTORC1 substrates S6K and 4E-BP1 after exposure to LPS, which is blunted by torin1, which acts as active-site mTOR inhibitor (Fig. 5a, b). Moreover, consistent with our previous observations 15 , lysosome tubulation was suppressed upon inhibition of mTOR or AKT by torin1 and Akti, respectively (Fig.   5c). Importantly, suppression of AKT and mTOR activity abrogated the LPS-induced expansion of the lysosome volume (Fig. 5d). Finally, the increase in the holding capacity for pinocytic fluid enticed by LPS treatment was blunted by torin1 (Fig. 5e). Collectively, these findings demonstrate that mTOR is stimulated by LPS, whereby mTOR activation is required for an increase in lysosome holding capacity, volume expansion and tubulation.

In response to LPS, mTOR selectively increases translation of mRNAs encoding lysosomal proteins
Given that mTOR is hyperactivated in LPS-exposed phagocytes and its activity is necessary for lysosome expansion, we next tested whether LPS stimulates global protein synthesis in primary macrophages by employing the puromycylation assay. LPS enhanced puromycin incorporation as compared to control in a time-dependent manner, which is indicative of elevated protein synthesis ( Fig. 6a, b). Torin1 abrogated the LPS-induced increase in puromycylation (Fig. 6a, b). As a positive control, we demonstrated that translation elongation inhibitor cycloheximide abrogated puromycylation ( Fig. 6a, b). Altogether, these results demonstrate that LPS bolsters global protein synthesis in primary macrophages.
In addition to regulating global protein synthesis rates, changes in mTOR activity cause selective alterations in translation of specific mRNA subsets 65 . Considering that LPS increased lysosomal protein levels without altering their stability or corresponding mRNA abundance ( Fig.   3a and 4c), we next postulated that mTOR stimulates lysosome expansion by inducing translation of mRNAs encoding lysosomal proteins. To test this hypothesis, we employed polysome profiling wherein mRNAs are separated according to the number of bound ribosomes by sedimentation through the 5-50% sucrose gradient 66 . Distribution of mRNAs encoding lysosomal proteins across the gradient was measured by RT-qPCR. Due to technical limitations related to the amount of the material required for polysome profiling studies, these experiments were carried out using RAW macrophages. Relative to the control, LPS treatment shifted the distribution of mRNAs encoding LAMP1, and the V-ATPase subunits H and D towards the heavy polysome fractions, which is indicative of increased translational efficiency (Fig. 7a-c and Sup. Fig. S5). Importantly, although torin1 exerted minimal effect on the distribution of LAMP1 and V-ATPase subunits H and D mRNAs in control cells (Sup. Fig. S4), it dramatically reduced loading of these transcripts on heavy polysomes in LPS treated cells (Fig. 7a-c, Sup. Fig. S5). These findings indicate that LPS induces translation of mRNAs encoding LAMP1, and V-ATPase subunits H and D via mTOR. Of note, translational regulation of LAMP1 and V-ATPase subunits H and D is further buttressed by the results obtained in primary macrophages wherein LPS treatment induced LAMP1 and V-ATPase subunits protein levels, without affecting their mRNA levels or protein stability ( Fig. 3a and 4c). In striking comparison, distribution of mRNAs encoding the two heavy polysomal housekeeping proteins b-actin and PPIA remained unchanged upon addition of LPS and/or torin1 we next tested whether these translation regulatory factors play a direct role in LPS-induced lysosome expansion. To disrupt S6 kinase function in primary macrophages, we used LY258470, a potent pharmacological inhibitor of S6 kinase (Fig. S6A). We showed that LY258470 treatment of primary macrophages abrogated LPS-induced lysosome volume increase (Fig. 8a, c), but not lysosomal tubulation (Fig. 8b), which suggest a previously unrecognized uncoupling lysosome tubulation and expansion. Similarly, inhibition of S6 kinases thwarted the LPS-mediated increase in the levels of LAMP1, ATP6V1H and ATP6V1D (Fig. 8d, e). LY258470 had no effect on lysosomal transcript levels in resting cells or those co-exposed with LPS (Fig. S6b). Together, these data suggest that the mTORC1-S6 kinase axis promotes lysosomal protein expression to rapidly expand the lysosome size during phagocyte activation. We next investigated the role of 4E-BPs in regulating lysosome expansion following LPS.
For this, we generated RAW macrophages that stably expressed 4E-BP1 4Ala , a phosphorylationdeficient mutant of 4E-BP1 carrying alanine substitutions at four phosphorylation sites (Thr37, Thr46, Ser65, and Thr70), rendering it inaccessible to mTORC1 regulation 69 . This form of 4E-BP1 constitutively binds to a cap-binding protein eIF4E which prevents the assembly of the eIF4F complex thereby hindering recruitment of the mRNA to the ribosome 69 . First, relative to resting RAW counterparts, we showed that LPS augmented the lysosomal volume and lysosome tubulation in RAW cells expressing an empty pBabe retroviral vector (Fig 9a-c). In sharp contrast, LPS failed to boost the lysosome volume in RAW cells that stably expressed 4E-BP1 4Ala , though lysosome tubulation still occurred (Fig. 9a-c). Second, changes in lysosome volume were accompanied by corresponding alterations in lysosomal protein levels. Indeed, LPS exposure increased lysosomal protein levels (LAMP1, ATP6V1H and ATP6V1D) in RAW cells expressing the empty retroviral vector, compared to resting counterparts (Fig 9d, e). In contrast, LPS failed to boost the levels of these proteins in RAW cells stably expressing 4E-BP1 4Ala (Fig. 9d, e).
Collectively, these data reveal that the effects of mTORC1 on lysosome remodeling are mediated via orchestrated activity of S6Ks and 4E-BPs.

Antigen presentation is promoted by LPS through mTOR and S6K activity
Our results thus far suggest that LPS induces a rapid lysosomal expansion in primary phagocytes via mTOR-dependent boost in translation of lysosomal transcripts. Next, we sought to investigate the functional implication of LPS-mediated escalation in translation and lysosomal volume. We postulated that LPS-mediated lysosome expansion and increase in holding capacity may enhance antigen presentation by BMDCs. To test this, we used BMDCs from C57Bl/6 and C3H/He mice that respectively carry MHC-II haplotypes I-A b and I-A k . I-A b and I-A k expressing BMDCs were then respectively fed antigenic peptides, Ea 52-68 and Hen Egg Lysozyme (HEL), for 4 and 6 hours in the presence or absence of LPS. We used Aw3.18 monoclonal staining of fixed but unpermeabilized BMDCs to detect surface delivery of I-A k :: HEL 48-62 complexes 70 -this antibody could not detect MHC-II:peptide complex after permeabilization (not shown). On the other hand, we could detect total I-A b ::Ea 52-68 71 (internal and surface level) complex formation by staining fixed and saponin-permeabilized BMDCs with the monoclonal Y-Ae antibody 71 . Importantly, treatment with LPS stimulated formation and/or delivery of the MHC-II::peptide complexes even at 4 h and more potently at 6 h (Fig. 10a, c and Supp. Fig. 6c). When cells were not given antigens, signal was reduced to background, demonstrating that the fluorescence signal was dependent on MHC-II::antigen complex formation (Fig. 10c, Sup. Fig. S6c). We then inquired if antigen presentation was dependent on mTOR and S6K activities by co-treating cells with torin1 and LY258470, respectively. Remarkably, these inhibitors reduced antigen presentation of both Ea 52-68 and HEL 48-62 in unstimulated and LPS-treated cells (Fig. 10b, d and Fig. S6c). These data indicate that enhanced translation controlled by the mTORC1-S6K axis is necessary for efficient antigen presentation by BMDCs. Collectively with all the other observations, we propose that this is in part due to rapid lysosomal expansion and remodelling.

Discussion
Macrophages and dendritic cells are highly plastic inasmuch as they can dramatically alter their metabolic and gene expression profiles to adopt a range of alternative states, which exert both inflammatory or anti-inflammatory functions. Significant attention has been given to how macrophages and dendritic cells alter their metabolism and expression of cytokines, chemokines and other microbicidal agents [6][7][8][9] . Remodelling the expression level of these factors occurs both at the transcription and translation level 63,[72][73][74] . However, remarkably less is understood regarding the mechanisms that underpin changes to the endomembrane system during activation of macrophages and dendritic cells. Most notably, changes to the endomembrane system include reduced degradative capacity of endolysosomes to help conserve antigens in dendritic cells, and a dramatic morphological reorganization of the endolysosome system in both cell types, shifting from a large collection of vesicular organelles into a highly tubular network of lysosomes 11,19,20 . Tubular lysosomes are associated with increased pinocytic retention, exchange of phagosomal content within the endolysosomal system, and delivery of MHC-II-peptide for presentation, though how this occurs remains unknown 15,18,19,22,75 . Herein, we show that activated phagocytes also rapidly expand the lysosome volume and holding capacity. We provide evidence that this expansion relies on enhanced and selective translation of lysosome encoding transcripts via the mTORC1-S6K-4E-BP axes. In turn, this seems to enhance antigen retention, leading to efficient antigen presentation.

Re-organization and function of the lysosome system in activated phagocytes
Here, we disclose that activated macrophages and dendritic cells remodel their lysosome system into an expanded tubular network with augmented holding capacity. This conclusion is supported by several observations. First, imaging volumetric analysis revealed that dyes preloaded into lysosomes occupy a greater volume post-LPS activation using both live-and fixed-cell imaging, as well as super-resolution microscopy. Second, there was an increase in the expression level of several lysosomal proteins that was blunted by cycloheximide treatment or when translation was abated. Third, activated macrophages could hold a larger amount of fluid-phase relative to resting counterparts. Thus, overall, activated phagocytes not only undertake morphological reorganization of lysosomes but also expand this organelle. The increase in lysosome volume and holding capacity is consistent with work by Swanson et al. done in the 1980s showing that phorbol esteractivated macrophages retain fluid-phase more effectively than resting macrophages 18,54 .
Functionally, an expanded lysosome volume may help phagocytes engulf more material and/or process extracellular particulates and soluble cargo more efficiently. This expansion of the endolysosome system should benefit both macrophages and dendritic cells upon activation. While mature dendritic cells are reported to have reduced endocytosis 49,50 , we show here that dendritic cells exhibit extensive pinocytosis for at least 8 h post-activation, providing an avenue to internalize and accumulate antigenic material. This is also consistent with recent reports revealing that dendritic cells are still able to internalize significant amounts of extracellular content 51,52 .
Consistent with this concept, we showed here that LPS stimulation of BMDCs significantly increased presentation of two distinct antigenic peptides as early as 4-6 h of antigen uptake. Importantly, antigen presentation was highly dependent on mTOR and S6K activities, suggesting that up-regulated translation coupled to lysosome expansion helps drive antigen presentation efficiency. Of course, additional mechanisms may also assist antigen presentation in BMDCs including LPS-mediated alteration of lysosome luminal properties like pH and exocytosis of lysosome tubules, which also depends on mTOR 15,19,20 . The extent of individual contributions made by lysosome tubulation, expansion, secretion and luminal remodeling to antigen presentation will need to be assessed in future studies.

Mechanistic insight into lysosome volume expansion
Phagocyte activation leads to a rapid expansion of lysosome volume within two hours of LPS exposure. This expansion is driven by de novo protein synthesis as indicated by cycloheximide-mediated block in lysosome expansion. Unexpectedly, our data also suggest that rapid expansion of lysosome volume and holding capacity in response to LPS stimulation is not dependent on transcription changes to lysosome protein-encoding genes since we did not observe induction of corresponding mRNA levels even after 6 h post-LPS activation. Consistent with this, kinetics of TFEB and TFE3 translocation into the nucleus did not parallel initial lysosome enlargement.
Lastly, deletion of these transcription factors did not impair tubulation or rapid lysosome expansion. In contrast to transcriptional programs, we observed that mTOR-dependent translational mechanisms play a key role in LPS-mediated lysosome expansion. First, LPS activates mTORC1, as indicated by increased phosphorylation of S6K and 4E-BP1, and enhanced protein synthesis, whereas its inhibition abrogated lysosome tubulation and expansion. Second, while the translation machinery is governed by a plethora of mechanisms, including mTORindependent pathways, our data reveal that S6Ks and 4E-BPs play a major role in governing lysosome size in response to stress. Potentially, mTORC1-mediated inhibition of 4E-BPs releases the translation brake imposed on the translation initiation factor eIF4E 36,76 . On the other hand, the role of S6Ks in this process is potentially more complex given its numerous targets that modulate translation, including the ribosomal protein rpS6, the translation initiator factor eIF4B, PDCD4 which governs the function of eIF4A, eEF2K which governs elongation rates and SKAR that may promote mRNA splicing and maturation 36,77,78 . We propose that mTORC1-regulated mRNA translation is a mechanism that broadly serves to rapidly scale the activity and size of other organelles in response to metabolic alterations and stress. Consistent with this idea, inhibition of TSC1/2, an inhibitor of mTORC1, or overexpression of S6Ks increased the size and length of cilia in Chlamydomonas reinhardtii and zebrafish 79 .
We also observed that the effects of LPS on translation of mRNAs encoding lysosomal proteins are selective, whereby translation of a subset, but not all lysosomal transcripts is upregulated. It is well accepted that in addition to its effects on global protein synthesis, mTOR selectively increases translation of specific mRNA subsets. Of these, the best characterized mRNAs are those carrying a 5' terminal oligopyrimidine tract (5'TOP) which renders transcript translation mTOR-sensitive [80][81][82][83] . However, the lysosomal mRNAs encoding LAMP1 and V-ATPase subunits that were probed here do not contain the classical 5'TOP sequences. This is consistent with recent studies suggesting that a significant number of mTOR-sensitive mRNAs lack the 5'-TOP motif [84][85][86] . We anticipate that transcripts encoding lysosomal proteins may be a useful functional model to identify novel mRNA sequence or structural motifs that engender mTOR-dependent translation.
Overall, we demonstrate that activated phagocytes dramatically reorganize their lysosomal system by expanding and forming a tubular lysosome network. This amplifies the lysosome holding capacity of phagocytes, augmenting their ability to internalize and retain more extracellular cargo, likely contributing to enhanced antigen presentation. We demonstrate that this process is rapid and occurs via enhanced and selective translation of mRNAs encoding lysosomal proteins, governed by mTOR, S6K and 4E-BPs. Collectively, we propose that mTORC1 and the regulated translation machinery is an important mechanism employed by cells to scale and adapt the size and volume of organelles in response to stress signals.

Cell lines and primary cells
Murine RAW macrophage cell lines carrying CRISPR-mediated deletion of TFEB, TFE3 or both were a kind donation from Dr. Rosa Puertollano, NIH, and were previously described 59

RAW 4EBP 4Ala stable cell line production
We generated RAW cells stably expressing the HA-4E-BP1 (4Ala) phosphorylation mutant or the corresponding empty pBABE vector as previously described 69

Lysosome labelling and tubulation
For lysosome labeling, cells were pulsed with 50-100 µg/ml Alexa 546 -conjugated dextran (ThermoFisher) for 0.5-1 h, followed by 3x wash with PBS and incubated with fresh medium for at least 1 h. To induce lysosome remodeling, BMDMs and BMDCs were exposed to 100 ng/mL LPS from Salmonella enterica serotype minnesota Re 595 (Sigma-Aldrich, Oakville, ON), while RAW macrophages were incubated with 500 ng/mL for 2 hours (unless otherwise stated).

Immunofluorescence and Fluorescence Microscopy
To fix and preserve lysosome tubules in RAW cells, cells were incubated with 0.45% (v/v) glutaraldehyde and 0.5% PFA (v/v) in PBS for 15 minutes at room temperature. Cells were then washed with PBS 4x, followed by incubation with 1 mg/mL ice-cold sodium borohydride (Sigma-Aldrich) for 5 min 3x to abate unreacted glutaraldehyde and quench its autofluorescence.
To visualize endogenous TFEB and TFE3, cells were fixed using 4% PFA for 15 min following treatment conditions. Cells were then treated with treated with 100 mM glycine in PBS to quench PFA, then in permeabilization buffer (0.2% Triton-X, 2% BSA in PBS) for 10 min and Live-cell imaging was done at 5% CO 2 and 37 °C using environmental control chambers.
Live-cell and fixed-cell imaging was done with a Quorum Diskovery spinning disc confocal microscope system equipped with a Leica DMi8 microscope connected to an Andor Zyla 4.2 Megapixel sCMOS or an iXON 897 EMCCD camera, and controlled by Quorum Wave FX powered by MetaMorph software (Quorum Technologies, Guelph, ON). We also used an Olympus IX81 inverted microscope equipped with a Hamamatsu C9100-13 EMCCD camera and controlled with Volocity 6.3.0 (PerkinElmer). For super-resolution imaging, we employed the Zeiss Elyra PS1 imaging system equipped with an Axio Observer Z1 microscope fitted with the Andor iXon3 885 detector for structure illumination microscopy (SIM) and powered by Zeiss Zen 2012 software (Zeiss Microscopy, Jena, Germany). Super-resolution image acquisition was acquired by grating for 3 rotations and 5 phases. All SIM reconstructed imaging was done using default settings for image reconstruction; to avoid artifact formation, only images with peak/mean ratios above 20 and noise filter less then -4 were accepted. After reconstruction, Volocity 6.3.0 (PerkinElmer) image analysis software was used. All microscopes were equipped with standard filters appropriate to fluorophores employed in this study, optics and stage automation.

Image analysis and volumetrics
The nuclear-to-cytosolic ratio of TFEB and TFE3 was estimated as the ratio of the mean fluorescence intensity in the nucleus over the mean intensity in the cytosol after background correction using ImageJ (v. µm. For antigen presentation analysis, we acquired confocal slices over 0.3 µm z-intervals and used Volocity to determine the total fluorescence intensity of antigen-MHCII complexes for 50-100 cells per experiment. To control for background, we established a threshold fluorescence intensity measure using a no-antigen control group for during each experiment. Image manipulation was done with ImageJ or Adobe Photoshop (Adobe Systems, San Jose, CA), without altering the relative signals within images or how data may be interpreted. All figures were assembled using Adobe Illustrator (Adobe Systems).

Puromycylation and Western blotting
For puromycylation assays, cells were treated with 10 µg/mL of puromycin (Sigma-Aldrich), or an equivalent water volume for the non-puromycin group, for the last 15 min of each treatment. (Mm00522550_m1) and IL-6 (Mm00446190_m1) were done in triplicate. Target gene expression was determined by relative quantification (ΔΔCt method) to Abt1 and the vehicle-treated control sample.

Polysome profiling
Polysome profiling was performed as detailed in Gandin et al. 66      Relative lysosome volume between resting primary macrophages and those exposed to specified conditions. Shown is the mean ± standard error of the mean from 30-40 cells for each condition and experiment, across at least 3 independent experiments. Statistical analysis was done with ANOVA and unpaired Student's t-test. The asterisk * indicates a significant difference (p<0.05). Relative mRNA levels of select lysosomal genes (c) or interleukin-6 (d) in activated primary macrophages relative to Abt1 housekeeping gene and normalized against resting cells.
Quantification was done with qRT-PCR by measuring the DDCt as described in methods. Shown is the mean ± standard error of the mean from four independent experiments. e. Lysosomes in wild-type, tfeb -/-, tfe3 -/and tfeb -/-tfe3 -/-RAW strains before and after 2 h of LPS stimulation.
Images were acquired by live-cell spinning disc confocal microscopy. Yellow arrowheads illustrate tubular lysosomes. f. Relative lysosome volume between LPS-treated and resting counterpart RAW strains acquired by live-cell spinning disc confocal imaging. Shown is the mean ± standard error of the mean from 30-40 cells per condition per experiment across three independent experiments. All statistical analysis was done with ANOVA and unpaired Student's t-test. The asterisk * indicates a significant difference relative to resting condition (p<0.05). For a and e, scale bar = 5 µm.    pS6K and total S6K blots are shown to support effectiveness of LY258470 treatment. Shown is the mean ± standard deviation of the mean from five independent blots. For b, c and e, data was statistically analysed with ANOVA and unpaired Student's t-test (*p<0.05).  RAW cells were stimulated and then allowed to internalize LY over time. b. Pinocytosis rate by quantifying uptake of Lucifer yellow in RAW macrophages treated as indicated. c. Retention of Lucifer yellow chased in probe-free medium in RAW cells previously treated as indicated and prelabelled with Lucifer yellow for 1 h. In all cases, fluorescence measurements were done by flow cytometry. d. Pinocytosis in increasingly maturing DCs exposed to LPS. Microscopy was used to measure the uptake of fluorescent dextran for 30 min by DCs exposed to LPS over indicated time points. Shown is the mean ± standard error of the mean from at least three experiments. For statistical analysis, ANOVA or Analysis of Covariance was used, whereby an asterisk indicates a significant difference in fluorescent probe levels compared to resting (*p<0.05).

Supplemental Figure 3: Basal lysosome properties and trafficking is indistinguishable in
wild-type RAWs and strains deleted for TFEB and/or TFE3. a. LAMP1 levels in whole cell lysates from wild-type and deletion mutants of TFEB and/or TFE3. b. Quantification of LAMP1 levels in knock-out cells. LAMP1 levels were normalized to β-actin to control for loading.
Statistical analysis using ANOVA determined that LAMP1 levels did not vary across strains. c.