Introduction

The Synaptotagmin-Like (SYTL) protein family consists of 5 members (SYTL1-5), each containing an N-terminal Synaptotagmin homology domain (SHD) and two C-terminal C2 lipid-binding domains. The SHD domains of SYTL1-5 have been found to directly interact with the small GTPase protein RAB27^1,2. The C2 domain is generally involved in phospholipid binding and interacts with cellular membranes³, either in a calcium-dependent or calcium-independent manner^3,4. Calcium was shown to be required for SYTL3 and SYTL5 phospholipid binding activity^1,5, whereas SYTL2 binding to phosphatidylserine (PS) was inhibited by calcium⁶.

The SYTL protein family is generally involved in plasma membrane transport and exocytosis^2,4, mostly through their binding to RAB27 proteins. The main functions of RAB27 are related to vesicle budding, delivery, tethering and fusion with membranes⁷ and its activity is regulated by a cyclic activation and inactivation state depending on it binding to guanosine-5’-triphosphate (GTP) or guanosine diphosphate (GDP), respectively⁷. RAB27A and RAB27B are the two RAB27 isoforms found in vertebrates⁷ and their binding to effector proteins occurs only when they are bound to GTP⁷. SYTL5 binds to the GTP-bound RAB27A form, indicating its possible role in membrane trafficking events¹.

Functional mitochondria are fundamental for normal cellular metabolism. Mitochondria are the primary source of adenosine triphosphate (ATP) obtained through oxidative phosphorylation (OXPHOS), and are important for cellular calcium homeostasis, lipid metabolism, reactive oxygen species (ROS) generation, and detoxification⁸. Mitochondria can integrate and generate signalling cues to adjust their metabolism and biogenesis to maintain cellular homeostasis^8,9. Mitochondrial function and health are tightly regulated by several quality control processes involving targeting of parts of mitochondria to lysosomes for degradation, including macromitophagy^10,11 (selective degradation of mitochondria by autophagy), piecemeal mitophagy^12,13, mitochondrial-derived vesicles (MDVs)^14,15, and vesicles derived from the inner mitochondrial membrane (VDIMs)¹⁶. It has also been found that damaged mitochondria can be directly released to the extracellular space¹⁷, either contained within extracellular vesicles¹⁸ or by mitocytosis, where damaged mitochondria are expelled in migrasomes¹⁹.

Exposure of mitochondria to various stressors can affect their function and lead to severe diseases, such as neurodegenerative disorders and cancer^20,21. Cancer cells can trigger a change in mitochondrial metabolism to promote cell proliferation and survival by a process known as the Warburg effect, characterised by a switch from OXPHOS to ATP production through glycolysis even under normoxic conditions²². Adrenocortical carcinoma (ACC) is a rare type of cancer that develops in the adrenal gland cortex, having a very poor prognosis and limited treatment options²³. Most patients are diagnosed at advanced cancer stages and present an excess in adrenocortical hormone production²³. Mitochondria are known to be essential organelles in the synthesis of steroid hormones, including cortisol, since they contain specific enzymes that catalyse steroid synthesis from cholesterol delivered to the mitochondria²⁴.

Here we demonstrate that SYTL5 localises to mitochondria in a RAB27A-dependent manner and that both proteins regulate mitophagy and cellular metabolism. Cells lacking SYTL5 undergo a shift from mitochondrial oxygen consumption to glycolysis, which may explain the correlation between low SYTL5 expression and poor survival of ACC patients.

Results

SYTL5 localises to mitochondria and endolysosomal compartments

SYTL5 was identified as a putative candidate in a screen for lipid-binding proteins involved in the regulation of mitophagy²⁵. To characterise the cellular localisation and function of SYTL5, we first generated U2OS cells with stable inducible expression of SYTL5-EGFP as there are no antibodies recognizing endogenous SYTL5. Intriguingly, live cell imaging analysis revealed that SYTL5-EGFP co-localised with filamentous structures positive for MitoTracker red (Figure 1A), which labels active mitochondria. Moreover, several SYTL5-EGFP positive vesicles were observed, including some small and highly mobile SYTL5-EGFP vesicles moving along the mitochondrial network (Supplementary video 1 and arrows in Figure 1B). SYTL5-EGFP staining at the plasma membrane was also seen, particularly in cells with higher expression levels (Figure 1C), which may suggest a role for SYTL5 in secretion to the plasma membrane.

Figure 1.

To determine the identity of the SYTL5 positive vesicles, U2OS SYTL5-EGFP cells were infected with a panel of lentiviral RAB GTPase constructs representing different cellular compartments (Supplementary Figure 1A-F) and analysed by live cell imaging. SYTL5-EGFP positive structures were observed to co-localise with vesicles positive for RAB4 and RAB11, two GTPases that mainly localise to recycling endosomes²⁶ and to some extent with RAB5, mostly localised to early endosomes²⁶ (Supplementary Figure 1A-C). Co-localisation was also observed with RAB7, a marker of late endosomes, lysosomes and autophagosomes²⁶ and partially with RAB9, which associates with late endosomes and mediates the transport between late endosomes and the trans-Golgi network²⁶ (Supplementary Figure 1D-E). SYTL5-EGFP showed little or no co-localisation with RAB6 positive structures, representing Golgi-derived vesicles (Supplementary Figure 1F). In line with an endocytic identity of SYTL5-EGFP positive vesicles, they were also positive for lysosomal markers, including LysoTracker red, which labels acidic compartments in the cell (Figure 1C), and the lysosomal marker LAMP1 (Figure 1D).

Thus, based on live cell imaging analysis we conclude that SYTL5 localises to the mitochondrial network and to mitochondria-associated vesicles that partly overlap with endolysosomal compartments.

Mitochondrial localisation of SYTL5 requires both the RAB27-binding SHD domain and the lipid-binding C2 domains

To investigate whether the observed intracellular localisation of SYTL5-EGFP depends on its binding to RAB27 and/or lipids, we generated U2OS cells with stable expression of EGFP-3xFLAG-tagged wild type SYTL5 or SYTL5 lacking either the RAB27-binding SHD domain (SYTL5(ΔSHD)) or the two lipid-binding C2 domains (SYTL5(ΔC2AB)) (Figure 1E).

To validate the SYTL5(ΔC2AB) mutant and determine the lipid-binding specificity of SYTL5, lysates from these cells were incubated with membranes containing various phosphoinositides and other lipids. Full-length SYTL5 was observed to bind to mono-phosphorylated phosphoinositides (PI(3)P, PI(4)P, PI(5)P), as well as PI(4,5)P2, PI(3,4,5)P3 and to some extent to phosphatidic acid (PA) and cardiolipin (CL) (Figure 1F). As expected, upon removal of the C2 domains, specific binding to all lipid species was lost, while SYTL5 lacking the SHD domain retained the ability to bind to lipids (Figure 1F).

The SYTL5(ΔSHD)-EGFP and SYTL5(ΔC2AB)-EGFP expressing cell lines were then incubated with MitoTracker red and analysed by live cell imaging. Indeed, the SYTL5 colocalisation to mitochondria was lost and both SYTL5(ΔSHD)-EGFP and SYTL5(ΔC2AB)-EGFP were dispersed in the cytosol (Figure 1G-H), indicating that both the RAB27-interacting SHD domain and the lipid binding C2 domains are required for mitochondrial localisation of SYTL5.

Mitochondrial localisation of SYTL5 requires RAB27A GTPase activity

Given that the SHD domain of SYTL5 has been shown to bind RAB27A¹ and our observations that the SHD domain is required for mitochondrial localisation of SYTL5, we asked whether RAB27A activity is required for SYTL5 recruitment to mitochondria. U2OS SYTL5-EGFP cells were infected with a mScarlet-RAB27A lentiviral construct to generate a stable cell line. Co-expression of SYTL5-EGFP and mScarlet-RAB27A resulted in a clear co-localisation of both proteins to filaments positive for MitoTracker DeepRed (DR) (Figure 2A). The mitochondrial localisation of SYTL5-EGFP and mScarlet-RAB27A was confirmed by correlative light and EM (CLEM) analysis using the same cell line (Figure 2B, zoom in 1). The mitochondrial localisation of SYTL5-EGFP was enhanced when co-expressed with mScarlet-RAB27A compared to cells expressing SYTL5-EGFP only (Figure 1A-B), indicating that RAB27A might facilitate mitochondrial recruitment of SYTL5. Indeed, in U2OS cells with stable expression of mScarlet-RAB27A only, RAB27A strongly co-localised with MitoTracker green (Supplementary Figure 2A), demonstrating its mitochondrial targeting. Moreover, SYTL5-EGFP and mScarlet-RAB27A were both detected in the mitochondrial fraction (containing

Figure 2.

TIM23 and COXIV) of cells expressing SYTL5-EGFP or mScarlet-RAB27A (Supplementary Figure 2B). Importantly, the mitochondrial localisation of SYTL5 and RAB27A is not specific to U2OS cells as HeLa cells with stable expression of SYTL5-EGFP or mScarlet-RAB27A show a similar expression pattern to U2OS cells (Figure 1A & C), as demonstrated by co-localisation of SYTL5-EGFP with MitoTracker Red and plasma membrane staining in cells having higher expression (Supplementary Figure 2C) and of mScarlet-RAB27A with MitoTracker Green (Supplementary Figure 2D). All further experiments were conducted in U2OS cells.

To further characterise mitochondrial recruitment of SYTL5 and RAB27A and elucidate their role at the mitochondrion, we used CRISPR/Cas9 to knock out (KO) RAB27A and SYTL5 in U2OS cells (Supplementary Figure 3A-D). This double-KO (dKO) cell line was further transduced with lentiviral constructs to constitutively express SYTL5-EGFP, mScarlet-RAB27A, or both, and their mitochondrial localisation was analysed by live cell imaging (Figure 2C-E). Intriguingly, mScarlet-RAB27A co-localised extensively with MitoTracker green and to small vesicles dispersed in the cytoplasm and located at or near mitochondria (Figure 2C), indicating that SYTL5 is not required for RAB27A mitochondrial localisation. In contrast, SYTL5-EGFP was not recruited to mitochondria when expressed alone in dKO cells, although some SYTL5-EGFP vesicles were seen near mitochondrial structures (Figure 2D). Upon co-expression of SYTL5-EGFP and mScarlet-RAB27A in dKO cells the mitochondrial localisation of SYTL5-EGFP was rescued (Figure 2E), indicating that mitochondrial recruitment of SYTL5 is dependent on RAB27A.

Since SYTL5 has previously been described as a RAB27A effector¹, we asked whether its mitochondrial recruitment depends on binding to the GTP-bound form of RAB27A. To address this, dKO cells expressing SYTL5-EGFP were rescued with the constitutively active (RAB27A-Q78L) or inactive (RAB27A-T23N) mutant forms of mScarlet-RAB27A. As expected, SYTL5-EGFP interacted specifically with RAB27A wild type (WT) and RAB27A-Q78L, as well as with endogenous RAB27A in control cells, while no interaction was detected with RAB27A-T23N, as assessed by GFP-pulldown (Figure 2F). To our surprise, live cell imaging of the same cells revealed that, in contrast to mScarlet-RAB27A WT, neither the RAB27A Q78L nor the T23N mutant localised to the mitochondrial network when expressed together with SYTL5-EGFP in dKO cells. Also, SYTL5-EGFP failed to localise to mitochondria in cells expressing RAB27A Q78L, further demonstrating that mitochondrial recruitment of SYTL5 depends on mitochondrial RAB27A localisation (Figure 2E). Taken together, our data indicate that the GTPase activity of RAB27A is required for its mitochondrial localisation, which facilitates further recruitment of SYTL5.

SYTL5 interacts with proteins involved in vesicle-mediated transport and cellular response to stress

As few molecular interactors of SYTL5 are known¹, we analysed the SYTL5 interactome by immunoprecipitating SYTL5-EGFP or an EGFP control, followed by mass spectrometry (MS) and protein-hit enrichment analysis. RAB27A was one of the most significant hits found, providing evidence that the IP was successful, together with several proteins known to participate in secretion such as PDCD6IP²⁸ (Figure 3A). In total, 163 proteins were identified as significant SYTL5-EGFP interactors compared to the EGFP control (Figure 3A, Table 1) and were subjected to Gene Ontology (GO) term enrichment analysis. Considering GO enrichment within the subcellular compartment ontology (GO-CC), we discovered the following compartments as enriched (Figure 3B) (number of hits for each category in parentheses): cytoskeleton (43), intracellular vesicle (38), extracellular vesicle (36), plasma membrane-bound cell projection (33), endosome (18) and focal adhesion (17). Furthermore, for biological processes, encoded in the GO-BP subontology, we discovered the following categories as enriched (Figure 3A and C): vesicle-mediated transport (38), cellular response to stress (34), response to oxygen-containing compound (31), intracellular protein transport (22), secretion (21), autophagy (13), stress-activated MAPK cascade (10), cellular response to oxidative stress (9), reactive oxygen species metabolic process (7), regulation of mitochondrion organization (6), endosome organization (4) and protein insertion into mitochondrial membrane (3). Taken together, the SYTL5 interactome indicates a role for SYTL5 in mitochondrial processes and cellular response to stress, in line with its localisation to mitochondria, as well as a possible role in secretion and endocytosis, in line with its localisation to endocytic compartments and the plasma membrane (Figure 1).

Figure 3.

Table 1.

SYTL5-RAB27A positive vesicles contain mitochondrial components

Several proteins involved in autophagy were identified as specific SYTL5 interactors, including SQSTM1/p62, ATG4B, TBK1, and ATP6V1A (Figure 3A), and given the mitochondrial localisation of SYTL5, we first investigated a possible role of SYTL5 in mitophagy. To induce mitophagy, dKO cells expressing mScarlet-RAB27A and SYTL5-EGFP were subjected to hypoxia or the drugs DFP (deferiprone, an iron chelator) or DMOG (Dimethyloxalylglycine, a prolyl-4-hydroxylase inhibitor) that mimics hypoxia by activation of HIF1α. We noticed an increase in the number and size of vesicles positive for SYTL5-EGFP and mScarlet-RAB27A that also contained MitoTracker Deep Red (MitoTracker DR) in all conditions, compared to control cells (Figure 4A).

Figure 4.

HIF-1α is a transcription factor that is a main driver of the Warburg effect¹⁵ and an inducer of mitophagy via upregulation of the outer mitochondrial membrane proteins BNIP3 and BNIP3L, which bind to LC3. To identify the nature of the SYTL5/RAB27A/MitoTracker DR positive vesicular structures that increase with HIF-1α stabilisation, dKO cells expressing mScarlet-RAB27A and SYTL5-EGFP were immunostained with antibodies against various autophagy markers. Intriguingly, while observing strong co-localisation of SYTL5/RAB27A/MitoTracker DR positive vesicles with LAMP1 (Figure 4B), the same structures did not co-localise with LC3 or the autophagy receptor p62 (Supplementary Figure 4A). Moreover, the mitochondrial localisation of SYTL5 and RAB27A was not affected in cells treated with the ULK1 inhibitor MRT68291 or with the class III phosphoinositide 3-kinase VPS34 inhibitor IN1 (Supplementary Figure 4B), together indicating that SYTL5 and RAB27A are recruited to mitochondria independent of active autophagy.

SYTL5 and RAB27A regulate turnover of inner mitochondrial membrane proteins

To investigate whether SYTL5 and RAB27A are required for the turnover of mitochondrial proteins in hypoxic conditions, we took advantage of two mitophagy reporter U2OS cell lines, one expressing the inner mitochondrial membrane (IMM) reporter, pSu9-Halo-mGFP²⁹ and one expressing the mitochondrial targeting signal of the matrix protein NIPSNAP1 tagged with EGFP-mCherry³⁰ (hereafter referred to as IMLS cells).

The pSu9-Halo-mGFP cells allow a measure of mitophagy flux via the production of free Halo^TMR, which is stable in lysosomes when bound to the ligand. Activation of HIF1α by treatment with DFP or DMOG resulted in the appearance of a Halo Tag band due to cleaved Halo^TMR (Figure 4C-D) consistent with an activation of mitophagy. The use of small interfering RNA (siRNA) to deplete SYTL5, RAB27A or both simultaneously in these cells, resulted in a significant reduction in the amount of free Halo^TMR (Figure 4C-D) highlighting a reduction in mitophagy. This indicates that both SYTL5 and RAB27A are required for the turnover of mitochondrial inner membrane components in response to HIF1α activation.

Two independent siRNAs were used to deplete SYTL5 (Supplementary Figure 5A) in IMLS cells treated with DFP to induce Parkin-independent mitophagy and in IMLS cells with stable expression of Parkin treated with the uncoupler carbonyl cyanide m-chlorophenyl hydrazone (CCCP) to induce Parkin-dependent mitophagy. We performed imaging-based quantification of mCherry-only puncta (which represent mitochondria delivered to lysosomes as the EGFP signal is quenched in the acidic lysosome). Addition of the lysosomal V-ATPase inhibitor Bafilomycin A1 (BafA1) was used as a control to confirm that the mCherry-only puncta corresponded to lysosomal structures (Supplementary Figure 5B-C). There was no significant difference in mCherry-only area per cell relative to the siControl in cells depleted of SYTL5, neither in cells treated with DFP (Supplementary Figure 5B) or with CCCP (Supplementary Figure 5C). Moreover, IMLS turnover was not affected by depletion of RAB27A in these cells (Supplementary Figure 5D), together indicating that macro-mitophagy proceeds independently of SYTL5 and RAB27A. In line with this, the starvation-induced autophagic flux was not affected in cells lacking SYTL5 (Supplementary Figure 5E).

Taken together, our results indicate that SYTL5 and RAB27A function to regulate the lysosomal turnover of mitochondrial inner membrane proteins in response to hypoxic conditions, in a macro-mitophagy independent manner.

SYTL5 regulates mitochondrial respiration

Our finding of mitochondrial material within SYTL5-RAB27A positive structures upon induction of mitochondrial stress through DFP and hypoxia (Figure 4A), together with reduced turnover of the Fo-ATPase subunit 9 upon knockdown of SYTL5 and RAB27A (Figure 4C-D) and the identification of several proteins involved in the cellular response to stress (Figure 3C), including mitochondrial proteins (e.g., DECR1 and TIMM13), in the SYTL5 interactome (Figure 3A) led us to investigate a potential role of SYTL5 and RAB27A in mitochondrial function and bioenergetics.

The cellular oxygen consumption rate (OCR) can be measured with the Seahorse analyser upon sequential addition of different inhibitors (oligomycin, CCCP, antimycin and rotenone) that block the function of specific electron transport chain (ETC) complexes³¹. SYTL5 KO cells had a significant decrease in both basal and ATP-linked respiration compared to WT U2OS cells (Figure 5A-B), indicating reduced mitochondrial OXPHOS activity. As for RAB27A, both basal respiration and ATP-linked OCR were significantly reduced in cells depleted of RAB27A (RAB27A KO and RAB27A/SYTL5 dKO) relative to the control (Figure 5A-B), indicating an important function of these proteins in regulation of mitochondrial respiration.

Figure 5.

To test whether the decreased respiration in cells depleted of SYTL5 and/or RAB27A is coupled to a shift to glycolysis, we performed Seahorse analysis of the extracellular acidification rate (ECAR) in cells upon the sequential addition of glucose (increases glycolytic capacity), oligomycin (inhibits mitochondrial ATP production), and 2-DG (a glycolysis inhibitor which causes an abrupt decrease in ECAR). Intriguingly, glycolysis was significantly increased in SYTL5 KO cells as compared to control (Figure 5C-D), indicating a possible metabolic shift to compensate for the decrease in OCR observed when SYTL5 was depleted (Figure 5A-B). In contrast, ECAR was reduced in cells lacking RAB27A compared to the control, which may indicate an impairment in glucose uptake (Figure 5C) and therefore in glycolysis (Figure 5D). After adding oligomycin the ECAR increased in all cell lines, but in RAB27A KO and dKO cell lines, the increase was reduced when compared to the control cell line (Figure 5C-D).

The clear reduction in mitochondrial respiration observed in cells lacking RAB27A and SYTL5 led us to investigate the abundance of various ETC proteins. Depletion of SYTL5 and/or RAB27A did not affect the abundance of ATP5A (complex V), UQCRC2 (complex III) or SDHB (complex II), but there was a significant increase in the level of MTCO1 (complex IV) in cells lacking RAB27A (Figure 5E-F). This is in line with the reduced turnover of Halo-tagged ATPase subunit Su9 observed in cells depleted of SYTL5 and/or RAB27A (Figure 4C-D).

Taken together, our data show that both RAB27A and SYTL5 are important for normal mitochondrial respiration and ATP production while SYTL5 seems to regulate a metabolic switch to glycolysis.

Low SYTL5 expression is related to reduced survival for adrenocortical carcinoma patients

Since our data indicate that SYTL5 may be a regulator of the metabolic switch from oxidative phosphorylation to glycolysis (the Warburg effect) (Figure 5A-D) that is frequently associated with cancer cells, we decided to analyse publicly available expression datasets to explore a possible link between SYTL5 and cancer. Using bulk RNA expression data from healthy/normal tissue samples in the GTEx project, we found that SYTL5 is preferentially expressed in adrenal gland and various regions of the brain (Figure 6A). Adrenocortical carcinoma (ACC) is a rare type of aggressive cancer with poor prognosis and limited treatment options that develops in the adrenal cortex²³. A comparison of gene expression levels³² in normal adrenal gland samples and ACC samples showed that SYTL5 mRNA expression is significantly reduced in tumour samples. (Figure 6B). Furthermore, patient survival analysis using data obtained from the TCGA-ACC cohort, indicates that low SYTL5 mRNA expression is associated with a significantly reduced survival of ACC patients (Figure 6C), suggesting a possible tumour suppressor function for SYTL5 in ACC cancer progression. RAB27A expression levels were also reduced in tumor tissue compared to normal samples, although to a lesser extent than SYTL5 (Figure 6D), and overall patient survival in ACC patients did not show any relationship with RAB27A expression levels (Figure 6E).

Figure 6.

The adrenal gland cortex secretes several steroid hormones, such as mineralocorticoids (e.g. aldosterone), glucocorticoids (e.g. cortisol), and adrenal androgens (DHEA and DHEA-sulphate)³³, to the bloodstream. ACC tumours are characterised by excess adrenocortical hormone production in nearly 45–70% of patients, with hypercortisolism being the most common²³. As mitochondria contain enzymes essential for steroid hormone biosynthesis from cholesterol²⁴, we asked whether SYTL5 may regulate cortisol production. SYTL5 or RAB27A were depleted by siRNA in the ACC cell line H295R, followed by quantification of cortisol secretion using a colorimetric assay. We did not observe a significant change in cortisol secretion upon SYTL5 or RAB27A knockdown (Supplementary Figure 6), indicating that the correlation between low SYTL5 expression levels and ACC patient survival likely is unrelated to effects on cortisol production.

Discussion

SYTL5 has been identified as a RAB27A effector protein with affinity for membranes through its C2 lipid-binding domains¹, but nothing is known about a potential role of SYTL5 in regulation of RAB27A-dependent membrane trafficking pathways. Using the osteosarcoma derived U2OS cell line, here we confirm that SYTL5 interacts with RAB27A, and show that it localises to mitochondria and small, highly mobile vesicles interacting with the mitochondrial network, as well as to endolysosomal and endocytic compartments, indicating a role of SYTL5 in cellular membrane trafficking events.

Most interestingly, mitochondrial recruitment of SYTL5 requires its RAB27-binding SHD domain and its lipid-binding C2 domains. Indeed, we demonstrate that SYTL5 is recruited to mitochondria in a RAB27A-dependent manner and that RAB27A itself is localised to mitochondria in a manner dependent on its GTPase activity, as both RAB27A GDP and GTP mutants (T23N and Q78L, respectively) showed a dispersed cytoplasmic localisation. This is in line with a study showing that the GTPase activity of RAB27A is required for its localisation to melanosomes³⁴, but the mechanisms involved in mitochondrial recruitment of RAB27A are not clear. It is however interesting to note that ubiquitination of the RAB27A GTPase activating protein alpha (TBC1D10A) is reduced in the brain of Parkin KO mouse compared to controls³⁵, suggesting a possible connection of RAB27A with regulatory mechanisms that are linked with mitochondrial damage and dysfunction.

Given the presence of mitochondria-associated small and highly mobile SYTL5/RAB27A-positive vesicles that contain mitochondrial material and stain positive for LAMP1, we speculate that RAB27A and SYTL5 facilitate mitophagy-mediated lysosomal delivery of mitochondria or selected mitochondrial components. Indeed, SYTL5 and RAB27A positive vesicles containing mitochondrial material were observed in cells upon hypoxia or hypoxia-mimicking (DFP and DMOG) treatments. Intriguingly, in cells treated with DFP and DMOG we observed a significant decrease in lysosomal cleavage of the mitochondria inner membrane reporter pSu9-Halo-mGFP upon depletion of SYTL5 and/or RAB27A. Additionally, when blotting for various proteins of the electron transport chain complexes, we found that the level of the cytochrome c oxidase (complex IV) subunit MTCO1/COX1 was significantly increased in cells lacking RAB27A. In contrast, lysosomal targeting of a mitochondrial matrix reporter (IMLS) was not affected by depletion of SYTL5 and/or RAB27A, neither upon induction of Parkin-independent or Parkin-dependent mitophagy. In line with this, mitochondrial recruitment of SYTL5 and RAB27A was independent of the core autophagy machinery components ULK1 and VPS34 and the mitochondria-containing SYTL5/RAB27A-positive vesicles induced upon hypoxia did not stain positive for endogenous LC3B or p62, suggesting that SYTL5 and RAB27A-dependent clearance of mitochondrial components is independent of macroautophagy. The role of SYTL5 and RAB27A in the turnover of proteins involved in mitochondrial oxidative phosphorylation and ATP production is reminiscent of the recently described VDIM (Vesicles Derived from the Inner Mitochondrial membrane) pathway¹⁶. VDIMs are formed by IMM herniation through pores in the outer mitochondrial membrane, followed by their engulfment by lysosomes in proximity to mitochondria in a process that is independent of the core autophagy machinery. In line with our findings, VDIMs lack LC3 and p62 and VDIM formation was found to increase upon oxidative stress, leading to selective degradation of IMM proteins (including COX4 and other proteins involved in OXPHOS) while sparing the remainder of the organelle.

Mitochondria are known stress sensors and hubs for cellular adaptation⁸. When analysing the interactome of SYTL5, we found that many candidates were related to the cellular response to stress and oxygen-containing compounds, as well as vesicle-mediated transport and secretion. In line with a role for SYTL5/RAB27A in the regulation of mitochondrial stress, we found that cells lacking SYTL5 and/or RAB27A displayed reduced OXPHOS activity and ATP-production. Interestingly, the depletion of SYTL5 triggered a shift to glycolysis as observed by an increase in the extracellular acidification rate (ECAR) due to cellular lactate production. A shift from OXPHOS to anaerobic glycolysis occurs in normal cells during hypoxic conditions and often in cancer cells even under normoxic conditions, commonly referred to as the Warburg effect²².

The observed switch from OXPHOS to glycolysis in U2OS cells depleted of SYTL5 prompted us to look at cancer patient databases. We observed, using data from GTEx, that the adrenal gland is among the tissues with highest levels of SYTL5 gene expression, and that its expression is reduced in adrenocortical carcinoma samples (TCGA-ACC) when compared to normal adrenal gland tissue. Moreover, low SYTL5 expression is associated with reduced ACC patient survival. One characteristic of this type of cancer is an increase in steroid hormone production and secretion. Most steroid hormones are produced in mitochondria²⁴ but we did not observe any changes in the level of cortisol in ACC cells depleted of SYTL5. The mechanisms linking SYTL5 to the Warburg effect and ACC need further investigation, but it is interesting to note that RAB27A expression levels do not correlate with ACC survival.

To summarise, we show that SYTL5 is a RAB27A effector protein being recruited to mitochondria in a RAB27A-dependent manner. Upon hypoxia, RAB27A and SYTL5 promote lysosomal clearance of OXPHOS machinery components and both proteins are needed for mitochondrial respiration. We identify SYTL5 as a negative regulator of the Warburg effect and potentially tumorigenesis.

Materials and methods

Antibodies and dyes

Cell culture, lentivirus production and stable cell line generation

Human bone osteosarcoma epithelial cells (U2OS FlpIn TRex cell line, kindly provided by Steve Blacklow, Harvard Medical School, US) were grown in complete medium of Dulbeccós Modified Eagle Medium (DMEM; Lonza #12-604F and Gibco # 31966047) with 10% v/v foetal bovine serum (FBS; Sigma-Aldrich #F7524),100 U/ml Penicillin and 100 μg/ml Streptomycin (P/S; Thermo Fischer Scientific #15140122) at 37 °C with 5% CO2.

For lentiviral particle production HeK-FT cells were co-transfected with 1.6 μg of pLVX lentiviral vector, pCMV-VSV-G and psPAX2 using X-tremeGENE™ 9 DNA Transfection Reagent (Sigma #XTG9-RO). For virus infection, cells were seeded in complete medium and supplemented with 8 μg/ml polybrene (Santa Cruz Biotechnology #sc-134220). 2 μg/ml Puromycin (Sigma Aldrich #P7255) was added after 24 h post-infection for selection.

To generate U2OS stably expressing SYTL5-EGFP, U2OS FlpIn TRex cells and U2OS FlpIn TRex_mScarlet-RAB27A were co-transfected with 0.1 μg pcDNA™5/FRT/TO_SYTL5-EGFP and 0.9 μg pOG44 Flp recombinase expression vector using Lipofectamine™ 2000 (Thermo Fisher Scientific #11668019) according to the manufactureŕs instructions. To produce stable cell lines expressing pLVX-SV40-mScarlet-RAB27A(T23N/Q78L) and pLVX-CMV-SYTL5-EGFP-3xHA, U2OS dKO cells were infected with lentiviral particles followed by fluorescence-activated cell sorting (FACS) for double mScarlet/EGFP expressing cells. To generate stable cells expressing pLVX-CMV-SYTL5-EGFP-3xFLAG, pLVX-CMV-SYTL5 (ΔC2AB)-EGFP-3xFLAG and pLVX-CMV-SYTL5 (ΔSHD)-EGFP-3xFLAG, U2OS FlpIn TRex cells were infected with lentiviral particles. To generate U2OS stably expressing pSu9-Halo-mGFP, U2OS cells were infected with retroviral particles generated using pUMVC (addgene # 8449) and pMRX-IB-pSu9-HaloTag7-mGFP (addgene # 184905). Stable cell lines were maintained in complete medium with additional 5 μg/ml blasticidin S (Thermo Fisher Scientific #R21001) or 2 μg/ml puromycin (Sigma Aldrich #P7255). For starvation experiments, cells were washed with PBS and incubated for 4 h in Earlés balanced salt solution (EBSS; Gibco #24010043).

NCI-H295R cells were obtained from ATCC (#CRL-2128) and maintained in DMEM: F12, HEPES medium (Gibco #11330032) supplemented with 2.5% Nu-Serum (Corning #355100), 1% ITS+Premix Universal culture supplement (Corning #354352) and 100 μg/ml P/S (Thermo Fischer Scientific #15140122) at 37 °C with 5% CO2.

Hypoxia and drug treatments

After cells were seeded, they were transferred to a INVIVO₂ 200 Hypoxic workstation (Ruskinn). Hypoxia treatment was carried out for 24 hours with oxygen concentration set to 1%.

Drugs were added into the cell media. Treatment durations and drug concentrations are indicated in the figure legends. Drugs used were DFP (Sigma-Aldrich #379409), DMOG (Sigma-Aldrich #D3695), VPS34IN1 (Selleckchem # S7980) and MRT68291 (Selleckchem #S7949).

Molecular cloning

SYTL5 was amplified from WT U2OS cDNA (5’-ATGTCTAAGAACTCAGAGTTCATC-3’ and 5’-TTAGAGCCTACATTTTCCCATG-3’) and cloned into Zero Blunt TOPO vector using Zero Blunt™ TOPO™ PCR Cloning Kit (Thermo Fisher Scientific #450245) according to the manufactureŕs instructions. pcDNA™5/FRT/TO_SYTL5-EGFP was generated by Gibson Assembly using the TOPO-SYTL5 vector as a template for SYTL5 and cloned into a pcDNA™5/FRT Mammalian Expression Vector (Thermo Fisher Scientific #V601020) or into a pLVX-CMV lentiviral expression vector. Truncated SYTL5 (ΔSHD) and SYTL5 (ΔC2AB) were generated by Gibson assembly using the pcDNA™5/FRT/TO_SYTL5-EGFP construct as a template and cloned into a pLVX-CMV lentiviral expression vector.

RAB27A was amplified from WT U2OS cDNA and cloned into a pLVX-SV40 lentiviral expression vector. RAB27A mutations were performed using the site-directed mutagenesis QuikChange II kit (Agilent #200524) according to the manufactureŕs instructions using the following primer pairs:

CRISPR knockout cell line generation and validation

SYTL5 and RAB27A gene knockouts were performed as previously described by Ran et al., 2013. Briefly, gRNA target sequences were designed using Wellcome Sanger Institute Genome Editing (WGE) design tool available at https://wge.stemcell.sanger.ac.uk/find_crisprs.

DNA oligos were annealed and phosphorylated before cloning into pSpCas9(BB)-2A-Puro (PX459) V2.0 (addgene # 62988). The obtained CRISPR plasmids were transfected into U2OS Flp-In TRex cells using Lipofectamine™ 2000 according to the manufactureŕs instructions and selection with 2 μg/ml puromycin (Sigma Aldrich #P7255) was applied for 24 h. Limiting dilutions were performed using a concentration of 15 cells/ml and seeded in complete medium in a 384 well plate. Single colonies were collected ∼10 days after seeding and expanded for a further 3 weeks. For genotyping, gDNA from each single clone line was isolated using QIAamp DNA Mini Kit (QIAGEN #51304) and the targeted genomic region was amplified by PCR using flanking primers (RAB27A: 5’-ATAACCTCTCCCTTGACCTTGTATG-3’ and 5’-TAGATGCCTTTGGGATTTGTACTGA-3’; SYTL5: 5’-CAGGTCCCTTTCTTCTCGCA-3’ and 5’-TCAGTCAGCTGCAAGAGTGG-3’). The PCR product was cloned into Zero Blunt TOPO vector using Zero Blunt™ TOPO™ PCR Cloning Kit according to the manufactureŕs instructions. RAB27A clonal lines were validated by western blot, and SYTL5 deletions were detected by running PCR products in 1% w/v agarose gel (Thermo Fisher Scientific #R2801).

siRNA knockdown by reverse transfection

For siRNA knockdown, cells were incubated with OptiMEM (Thermo Fisher Scientific # 31985047) and Lipofectamine RNAiMAX (Thermo Fisher Scientific #13778150) for 5 min at room temperature (RT). Followed by addition of 20 nM siRNA diluted in OptiMEM and 15 min incubation at RT. Cells were fixed/harvested after 72 h of knockdown. The siRNA used for mRNA depletion are the following: siControl: 5’-UAACGACGCGACGACGUAAtt-3’; siSYTL5 #1: 5’-CUCUUAGAAGCAAAACGUAtt-3’; siSYTL5 #2: 5’-CAACAAGCGUAAGACCAAAtt-3’; siSYTL5 #3: 5’-GGUUUGUGCUUCAACCCAAtt-3’; siRAB27A #1: 5’-GGAAGACCAGUGUACUUUAtt-3’ and siRAB27A #2: 5’-CCAGUGUACUUUACCAAUAtt-3’.

RNA isolation, cDNA synthesis and RT-PCR

RNA was isolated using TRIzol reagent (Thermo Fisher Scientific #15596026) and cDNA was synthesised using SuperScript III Reverse Transcriptase (Thermo Fisher Scientific #18080085) according to the manufactureŕs instructions. qPCR analysis was performed using KAPA SYBR FAST qPCR Kit (Sigma-Aldrich #KK4601) and relative mRNA expression levels were normalised to the expression levels of TATA-binding protein (TBP) using the comparative ΔCt method. The primers used were the following:

SYTL5: 5’-GTGACAAAATCGCGCAGCTA-3’, 5’-GGACAACATCAGTGCCGAGA-3’

RAB27A: 5’-GGAGAGGTTTCGTAGCTTAACG-3’, 5’-CCACACAGCACTATATCTGGGT-3’

TBP: 5’-CAGAAAGTTCATCCTCTGGGCT-3’, 5’-TATATTCGGCGTTTCGGGCA-3’

Immunofluorescent staining

Cells were fixed in 4% paraformaldehyde (PFA) (Sigma-Aldrich #158127) in PBS for at least 15min at 37°C, quenched for 10 min using 0.05 M NH4Cl in PBS, permeabilised in 0.05% saponin in PBS for 5 min and blocked in 1% BSA in PBS for 30 min at room temperature. Antibody staining was performed in a wet chamber for 1 h at room temperature and the antibodies were diluted in PBS containing 0.05% saponin. Nuclei staining was performed using Hoechst 33342 (Thermo Fisher Scientific #H1399) diluted in PBS at 1µg/ml.

Light microscopy

Cells expressing inducible fluorescent proteins were induced with 100 ng/ml doxycycline (Clontech #631311) for 24 h. For live cell confocal microscopy cells were incubated with indicated dyes prior to imaging. Imaging was carried out with either the Andor Dragonfly High Speed Confocal Microscope or the Nikon Crest V3 with a 60x oil immersion objective (NA 1.4). High content widefield imaging was conducted with the ImageXpress Micro Confocal (Molecular Devices) microscope with a 20x objective (NA 0.45). Fixed immunofluorescence confocal imaging was performed using the Andor Dragonfly High Speed Confocal Microscope or Nikon Crest V3 both with a 60x oil immersion objective (NA 1.4) or Zeiss LSM 800 microscope (Zen Black 524 2012 SP5 FP3, Zeiss) with a 63x oil immersion objective (NA 1.4).

Correlative light electron microscopy

U2OS cells co-expressing SYTL5-EGFP and mScarlet-RAB27A were seeded in gridded glass-bottom cell culture dishes (MatTek #P35G-1.5-14-CGRD). Cells growing in monolayer were fixed in warm (≈37°C) 3.7% PFA in 0.2 M HEPES (pH 7) and imaged using the Andor Dragonfly High Speed Confocal Microscope. After imaging, the samples were fixed using 2% v/v glutaraldehyde in 0.2 M HEPES, pH 7.4. After post-fixation in 2% v/v osmium tetraoxide and 3% v/v K4[Fe(CN)6], samples were embedded in Epon and 60 nm thickness sections were cut with a diamond knife. Sections were analysed using a JEM-1400 Plus Transmission Electron Microscope.

Western blot

Cells were harvested in RIPA buffer (50 mM TRIS pH 7.4, 150 mM NaCl (Merck #1064041000), 0,5% sodium deoxycholate (Sigma-Aldrich #30970), 1% NP-40 (Sigma-Aldrich #I3091), 1 mM EDTA (VWR #20302-293), 0,1% SDS (Sigma-Aldrich #L3771) containing 1x protease inhibitor (Merck #11697498001) and 1x phosphatase inhibitor (Merck #4906845001). 20 µg of lysate was loaded on SDS-PAGE gel (BioRad #5671095) and proteins transferred to a PVDF membrane (Merck #IPFL00010) which was blocked using 5x v/v casein blocking buffer in PBS for 1 h at room temperature. The obtained membranes were immunoblotted using primary and secondary antibodies with washing steps using PBST (1x Phosphate-Buffered Saline with 1% Tween20 (Sigma-Aldrich #P1379)). Membranes were analysed using the Odyssey CLx imaging system (Li-cor biosciences) and protein levels were quantified by densitometry with ImageStudio Lite software (Li-cor biosciences).

Macro-mitophagy analysis

siRNA knockdown was conducted as previously described in U2OS-IMLS cells seeded in complete media supplemented with 100 ng/ml doxycycline. Culture media was replaced 24 h after transfection and 1 mM DFP (Sigma-Aldrich #379409) treatment was added after 48 h. For U2OS-IMLS cells expressing PARKIN, 20 μM CCCP (Enzo Life Sciences #BML-CM124-0500) was used for 16 h and 10 μM Q-VD-OPh (Sigma-Aldrich #SML0063-1MG) pan-caspase inhibitor was included to support cell survival^11,36. Control wells with 100 nM BafA1 (AH diagnostics #BML-CM110-0100) were dosed 2 h prior to fixation. After 72 h, cells were washed with PBS pH 7 and fixed in warm (≈37°C) 3.7% PFA in 0.2 M HEPES (pH 7) for 15 min at 37°C. After washing with PBS, cells were incubated with 2 μg/ml Hoechst and imaged.

Halo assay for mitophagy flux

siRNA knockdown was conducted as previously described in U2OS cells expressing the pSu9-Halo-mGFP reporter. 48 h after transfection culture media was changed to contain 100 nM tetramethylrhodamine (TMR)-conjugated Halo ligand (Promega #G8251). After 20 min incubation, cells were washed twice with PBS and the media replaced with either media containing no treatment, 1 mM DFP or 1 mM DMOG. 72 h after transfection cells were harvested and western blot conducted (as described earlier). Membranes were incubated with the Halo Tag antibody and actin antibody. Membranes were imaged using the Chemidoc MP system (Bio-Rad) and band intensity quantified in Image Lab 6.1 (Bio-Rad). Bands were normalised to actin loading control. The percentage cleaved Halo was calculated by the formula (Free Halo / (Free Halo + Full Length)) x 100.

Oxygen consumption rate measurement by Seahorse XF analyser

U2OS cells were seeded in complete medium at the density of 4x10⁴ cells/well into XF24 cell culture microplates (Agilent #100777-004). The sensor cartridge was hydrated for 24 h using XF calibrant in a 37°C non-CO₂ incubator. Before analysis, cell medium was replaced by warm (≈37°C) DMEM without Sodium Bicarbonate (pH 7.4) and placed for 1 h in a 37°C non-CO₂ incubator. Each measurement cycle consisted of a mixing time of 3 min and data acquisition period of 3 min (three data points). OCR results refer to the average rates during each measurement cycle. The final concentration for each mitochondrial inhibitor used in this assay was 1.5 µM Oligomycin A (SelleckChem #S1478), 1 µM CCCP, 0.5 µM Rotenone (Sigma Aldrich #R8875-1G) and 0.5 µM Antimycin (Sigma Aldrich #A8674). Four baseline measurements were performed before adding each compound and three response measurements were taken after each compound was added. Data analysis was performed using Seahorse Analytics software available at https://seahorseanalytics.agilent.com/ and OCR data was normalised to protein concentration in each well.

Extracellular acidification rate (ECAR) measurement by Seahorse XF analyser

U2OS cells were seeded in complete medium at the density of 4x10⁴ cells/well into XF24 cell culture microplates. The sensor cartridge was hydrated for 24 h using XF calibrant in a 37°C non-CO₂ incubator. Before analysis, cell medium was replaced by warm (≈37°C) XF DMEM media (Seahorse Bioscience #102353-100) supplemented with 2 mM L-Glutamine (BioNordika #BE17-605E) and pH adjusted to 7.4. The final concentration for each reagent used in this assay was 11 mM Glucose, 1.3 µM Oligomycin A, 0.1 M 2-Deoxy-Glucose (2-DG; Sigma Aldrich #D8375-1G)³⁷. Each measurement cycle consisted of a mixing time of 3 min and data acquisition period of 3 min (four data points). ECAR results refers to the average rates during each measurement cycle. Five baseline measurements were performed before adding each compound and four response measurements were taken after each compound was added. Data was normalised to protein concentration using Seahorse Analytics software available at https://seahorseanalytics.agilent.com/ and the glycolysis rate was calculated by subtracting the ECAR value after 2-DG treatment from the ECAR value after addition of glucose.

Purification of crude mitochondrial fraction

To purify crude mitochondrial fraction, cells expressing EGFP (control), mScarlet-RAB27A, SYTL5-EGFP and SYTL5(ΔC2AB)-EGFP were harvested from one 150 mm (80-90% confluency) dish, washed twice with ice cold homogenisation buffer (210 mM mannitol (Sigma Aldrich #M4125), 70 mM sucrose (Sigma Aldrich #S0389), 5 mM HEPES pH 7.12 at 25°C) and harvested in 1 mL homogenisation buffer supplemented (HBS) with 1 mM EGTA (Sigma Aldrich #E3889) and 1x protease inhibitor. The obtained cell lysate was further mechanically homogenised at 4°C using a cell homogeniser (Isobiotech) equipped with a 16 µm clearance ball by repeatedly passing the cell lysate for 10x repeats with the aid of 2 disposable syringes. The homogenised solution was further spun down at 1500xg for 3 min at 4°C to collect the TCL and the obtained supernatant collected and spun down at 13,000xg for 20 min. The obtained supernatant corresponds to the cytosolic fraction and the pellet was resuspended in 2 mL HBS and spun down at 13,000xg for 20 min. The obtained pellet corresponds to the enriched mitochondrial fraction and was resuspended in 900 µL HBS. 50 µL of each cellular fraction was resuspended in 30 µL 2x SDS-sample buffer and boiled at 100°C for 5 min and further analysed using SDS-PAGE.

GFP-TRAP and FLAG immunoprecipitation

Cells expressing EGFP were harvested from a 100 mm dish (80-90% confluency), washed with ice cold PBS and scraped using ice cold isolation buffer (1% NP-40, 2 mM EDTA, 136 mM NaCl, 20 mM Tris-HCl pH 7.4, 10% glycerol, 1x protease inhibitor and 1x phosphatase inhibitor). Cell lysate was spun down at 20,000xg for 10 min and the supernatant added to 25 µL equilibrated GFP-trap agarose beads (Chromotek #gta-20). Samples were rotated end-over-end for 1 h at 4°C. After 4x washing steps, the beads were collected by centrifugation at 2700xg for 1 min, resuspended in 30 µL 2x SDS-sample buffer and boiled at 100 °C for 5 min. The obtained supernatant was analysed by SDS-PAGE.

U2OS cells expressing pLVX-CMV-SYTL5-EGFP-3xFLAG, pLVX-CMV-SYTL5(ΔC2AB)-EGFP-3xFLAG and pLVX-CMV-SYTL5(ΔSHD)-EGFP-3xFLAG were harvested from a 150 mm dish (80-90% confluency), washed 2x with cold PBS and scraped using 1 ml of cold lysis buffer (50 mM Tris HCl, pH 7.4,150 mM NaCl, 1% Triton X-100, 5 mM CaCl₂ (Merck #1023821000, and 1x protease inhibitor. Cell lysates were spun down at 20,000xg for 10 min and the resulting supernatants added to 40 µL of pre-washed anti-FLAG M2-Agarose Affinity Gel (Sigma Aldrich #FLAGIPT-1). Samples were rotated end-over-end for 2 h at 4°C. After 3x washing steps, the affinity gel was centrifuged at 5000xg for 30 seconds and the FLAG fusion proteins eluted using 3x FLAG peptide according to manufacturer’s protocol. The eluted FLAG fusion proteins were diluted in 4 ml 1x TBST supplemented with 3% BSA and 5 mM CaCl₂.

In-vitro protein-lipid binding assay

Eluted FLAG fusion proteins (SYTL5-EGFP-3xFLAG, SYTL5(ΔC2AB)-EGFP-3xFLAG and SYTL5(ΔSHD)-EGFP-3xFLAG were diluted in 1x TBST supplemented with 3% BSA and 5 mM CaCl₂. PIP strips (Echelon Biosciences #P-6001) and membrane lipid strips (Echelon Biosciences #P-6002) were blocked in 1x TBST supplemented with 3% BSA for 1 h at room temperature and incubated with the diluted FLAG fusion protein for 1 h at 4°C with gentle agitation. Lipid membranes were washed 3x with TBST and incubated with FLAG Tag primary antibody for 1 h at room temperature and detected using SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific #34076).

LC-MS/MS analysis of SYTL5-EGFP interactome

Cells expressing SYTL5-EGFP or EGFP (control) were harvested using NP-40 lysis buffer (10 mMTris Cl pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 0.5% NP-40, 1x protease inhibitor. SYTL5-EGFP was immunoprecipitated using GFP-Trap agarose beads according to the manufacturer’s protocol. The obtained beads were washed twice with 50 mM ammonium bicarbonate, reduced, alkylated and further digested by trypsin for overnight at 37 °C. Digested peptides were transferred to a new tube, acidified and the peptides were de-salted for MS analysis.

Peptide samples were dissolved in 10 µl 0.1% formic buffer and 3 µl was loaded for MS analysis. LC-MS/MS analysis of the resulting peptides was performed using an Easy nLC1000 liquid chromatography system (Thermo Electron) coupled to a QExactive HF Hybrid Quadrupole-Orbitrap mass spectrometer (Thermo Electron) with a nanoelectrospray ion source (EasySpray, Thermo Electron). The LC separation of peptides was performed using an EasySpray C18 analytical column (2 µm particle size, 100 Å, 75 μm inner diameter and 25 cm (Thermo Fisher Scientific). Peptides were separated over a 90min gradient from 2% to 30% (v/v) ACN in 0.1% (v/v) FA, after which the column was washed using 90% (v/v) ACN in 0.1% (v/v) FA for 20 min (flow rate 0.3 μL/min). All LC-MS/MS analyses were operated in data-dependent mode where the most intense peptides were automatically selected for fragmentation by high-energy collision-induced dissociation.

Raw files from LC-MS/MS analyses were submitted to MaxQuant 1.6.17.0 software³⁸ for peptide/protein identification. Parameters were set as follow: Carbamidomethyl (C) was set as a fixed modification and PTY; protein N-acetylation and methionine oxidation as variable modifications. First search error window of 20 ppm and mains search error of 6 ppm. Trypsin without proline restriction enzyme option was used, with two allowed miscleavages. Minimal unique peptides were set to one, and FDR allowed was 0.01 (1%) for peptide and protein identification. The Uniprot human database was used. Generation of reversed sequences was selected to assign FDR rates. Further analysis was performed with Perseus³⁹, limma/voom⁴⁰ and Package R⁴¹. Volcano plots were plotted with EnhancedVolcano⁴². GO analysis was performed with shinyGO⁴³.

Bulk gene expression analysis – normal and tumor tissue

Bulk gene expression distribution in healthy/normal tissue samples was collected from GTEx v8 (PMID: 26484571). We utilised a combined RNA-seq expression dataset of TCGA and GTEx samples available through the Xena platform to investigate SYTL5 and RAB27A expression in normal and tumor tissue (PMID: 32444850). Here, data from a total of n = 77 samples provided normal adrenal gland expression (GTEx), and a total of n = 128 samples provided expression in adrenocortical carcinoma samples (TCGA-ACC). Survival data from samples in the TCGA-ACC cohort (one record per patient) was obtained using the TCGABiolinks R package⁴⁴. Survival curve analysis was performed using GraphPad Prism 8.0.1 using Logrank test (Mantel-Cox).

Cortisol assay

To measure cortisol in cell culture supernates, siRNA knockdown by reverse transfection with siSYTL5 and siRAB27A was conducted as described earlier in NCI-H295R cells. Cells were harvested with TNTE Lysis Buffer. Cortisol from lysates was measured using the R&D Systems® Cortisol Immunoassay kit (# KGE008B) as described in the protocol provided from the company. Controls were performed using cells treated with 50 uM Forskolin as a positive control to stimulate cortisol secretion⁴⁵, or 10 uM mitotane to inhibit cortisol production⁴⁶.

Statistical analysis

Statistical analysis was performed with GraphPad Prism (8.0.1, 9.3.1 and 10.2.0 versions) and the tests used are indicated in the respective figure legends. All values come from distinct samples. **** = p<0.0001, *** = p<0.001, ** = p<0.01, * = p< 0.05.

Supplementary figures

Figure S1

Figure S2

Figure S3

Figure S4

Figure S5

Figure S6

Acknowledgements

We thank Jenni Lane for assisting with EM sample preparation, Sachin Singh and Tuula Nyman at the OUS Proteomics Core Facility for assisting with mass spectrometry-based proteomic experiments, Ankush Sharma for help with analysis of mass spectrometry data, Anna Lång at ALM Core Facility Gaustad for assisting with widefield microscopy experiments, Santosh Phuyal for experimental assistance and Serhiy Pankiv for providing the lentiviral mScarlet-RAB expression constructs. We also thank Patricia González-Rodríguez for providing us with the U2OS pSu9-HaloTag7-mGFP cells.

This project was funded by the Research Council of Norway through its Centres of Excellence funding scheme (project number 262652) and FRIPRO grant (project number 249753), the Norwegian Cancer Society (project number 190251), Marie Skłodowska-Curie ETN grant under the European Union’s Horizon 2020 Research and Innovation Programme (Grant Agreement No 765912 DRIVE) and by the University of Oslo Scientia Fellow program through the MSC scheme – Co-funding of Regional, National and International Programmes (COFUND). Mass spectrometry-based proteomic analyses were performed by the Proteomics Core Facility, Department of Immunology, University of Oslo/Oslo University Hospital, which is supported by the Core Facilities program of the South-Eastern Norway Regional Health Authority. This core facility is also a member of the National Network of Advanced Proteomics Infrastructure (NAPI), which is funded by the Research Council of Norway INFRASTRUKTUR-program (project number: 295910).

Additional information

Author contributions

A.L., L.S.J., L.T.M. and A.S. designed and conceived the experimental work. A.L., L.S.J., L.T.M. and M.Y.W.N. performed the experiments and data analyses. S.S. carried out MS and GO term enrichment data analysis. S.N. provided GTEx and TCGA data analysis. E.L.E. performed and analysed CLEM data. A.L., L.S.J. and A.S. wrote the manuscript.

Additional files

Supplementary video 1. Live confocal microscopy imaging tracking SYTL5 (green) vesicle movement along filaments positive for MitoTracker Red (magenta).