Deficiency of Rab26 enhances insulin secretion
Rab26 was firstly identified in rat pancreas 12, and shown to regulate the maturation of exocrine granules to mediate amylase release from rat parotid acinar cells 16. Rab26 is transcriptionally expressed in multiple tissues, and especially at higher levels in the brain (Figure S1A). To study the physiological role of Rab26, we generated Rab26 gene knockout mice. The whole fragment of Rab26 gene was deleted through CRISPR/Cas9 approach using specific sgRNA targeting to the upstream of exon 1 and downstream of exon 9 of Rab26 gene (Figure 1A). The deletion of fragment was verified by DNA sequencing. The genotypes of mice were assessed by PCR approach (Figure 1B). The homozygous Rab26 gene knock out mice (Rab26-/-) were validated not expressing Rab26 as assessed by Western-blot (Figure 1C).
To directly assess the effect of Rab26 deficiency on insulin secretion, we isolated the islets form Rab26-/- mice (as well as control mice), and measured insulin secretion in perfusion culture. Consistently, the fresh islets isolated from Rab26-/- mice secreted much more insulin under 16.7mM glucose stimulation as compared to control islets (Figure 1D and 1E), suggesting that absence of Rab26 promotes insulin secretion from islets to increase the circulating levels of insulin in response to glucose.
To further support our conclusion, we depleted Rab26 in mouse pancreatic β-cells through CRISPR/Cas9 technique (Figure 1F). The stable INS-1 cell line with Rab26 depleted was used to examine insulin secretion. As shown in Figure 1G, depletion of Rab26 significantly increased insulin secretion. These results support the conclusion that deficiency of Rab26 enhances insulin secretion.
Deficiency of Rab26 in mice improves glucose homeostasis
Insulin is the most important hormone for regulating glucose homeostasis, therefore, we examined the effects of depletion of Rab26 on glucose tolerance in mice. We first performed intraperitoneal Glucose Tolerance Test (IPGTT) to assess the consequence of Rab26 depletion. As shown in Fig 1H and 1I, Rab26-/- mice have higher insulin levels in the blood upon glucose stimulation than wild type (WT) mice. And the glucose level in the plasma of Rab26-/- mice is much lower than that of WT mice under normal feeding condition (Figure 1J and 1K). The body weight of Rab26-/- mice is less than that of WT mice (Figure S1B). Intraperitoneal insulin tolerance tests (IPITT) demonstrated that Rab26-/- is more sensitive to insulin (Figure 1L and 1M), since the blood glucose levels lowered down faster in Rab26-/- mice. Together with the above results, deficiency of Rab26 may improve glucose homeostasis to prevent diabetes development through enhanced insulin secretion.
Over-expression of Rab26 restricts insulin secretion in β-cells
As depletion of Rab26 enhanced insulin secretion, we did the converse experiment to show that over-expression of Rab26 inhibits insulin secretion. Pancreatic β-cells MIN6 or INS-1 were infected with the recombinant adenovirus expressing Rab26 (Ad-Rab26) (Figure S2A and S2B), and then insulin secretion was detected by ELISA assay. As shown in Figure 2A-D, insulin secretion was clearly suppressed in both MIN6 (Figure 2A) and INS-1 (Figure 2C) cells upon overexpressing Rab26 in response to glucose stimulation, corresponding AUCs of insulin release stimulated by 16.7 mM glucose was decreased in MIN6 (Figure 2B) and INS-1 (Figure 2D) cells. In addition, the biphasic secretion pattern was disrupted in Ad-Rab26 infected MIN6 or INS-1 cells. However, the transcripts of Ins1 and Ins2 genes were not altered by expression of Rab26 (Figure S2C).
To examine whether the inhibitory effects of Rab26 is dependent of its guanine nucleotide binding activity, INS-1 cells were infected with Ad-Rab26, Ad-Rab26T77N (dominant negative mutant prefers binding to GDP), Ad-Rab26Q123L (constitutive active mutant with GTPase activity inhibited) or Ad-vector (Figure S2D). Cells were stimulated by glucose, then insulin secretion was detected by ELISA assay. The results demonstrated that insulin secretion was significantly suppressed by Ad-Rab26 or Ad-Rab26Q123L, but not by Ad-Rab26T66N (Figure 2E), suggesting the inhibitory effect of Rab26 on insulin secretion is physiologically regulated by its guanine nucleotide binding activity.
GLP-1 is a glucose-dependent hormone stimulating insulin secretion 17, we investigated the effect of Rab26 on GLP-1 stimulated insulin secretion, the results indicated that GLP-1 indeed enhanced insulin secretion. Importantly, Rab26 also suppressed GLP-1 stimulated insulin secretion (Figure 2F), indicating Rab26 restricts the glucose-stimulated insulin secretion (GSIS) in pancreatic β-cells.
As mentioned above, depletion of Rab26 enhancing insulin secretion in INS-1 cells. A rescue experiment revealed that insulin secretion was reduced upon replenishing Rab26 by infection of Ad-Rab26 in Rab26-knockout INS-1 (Figure 2G). Taken together, our results suggest that Rab26 serves as a negative regulator to restrict insulin secretion in pancreatic β-cells and this regulation is dependent on its guanine nucleotide binding activity.
The pathophysiological relevance of Rab26 to diabetes mellitus
To further define the role of Rab26 in insulin secretion, freshly isolated mouse islets were effectively infected with Ad-Rab26 (Figure 3A). Insulin secretion from islets was stimulated by 16.7mM glucose in KRBH buffer, as shown in Figure 3B, over-expression of Rab26 significantly inhibited insulin secretion from islets at different detection time points compared with islets transduced with vector control, suggesting over-expression of Rab26 suppresses insulin secretion not only in β-cell lines, but also in freshly isolated islets which is close to the in vivo physiological condition.
Next, we examined whether Rab26 affected the function of islets in vivo. The mouse model of type 1 diabetes was established by intraperitoneal injection of streptozotocin to destroy the islets (Figure 3C). Immunofluorescence staining with insulin antibody showed that injection of streptozotocin destroyed the islets accompanied with high blood glucose levels, indicating that the model of type 1 diabetic mice was generated (Figure 3D). Islets infected with Ad-Rab26 or vector were transplanted back to type 1 diabetes mice beneath the renal capsule (Figure 3E). The transplanted mice were intraperitoneally injected with glucose, then the plasma insulin and blood glucose were monitored. As expected, the plasma insulin levels of type 1 diabetic mice transplanted with islets overexpressing Rab26 were much lower than that of control mice (Figure 3F). Meanwhile, the blood glucose levels of mice transplanted with islets overexpressing Rab26 decreased more slowly and kept at high level for longer period of time (continuously monitored for 6 days, Figure 3G, 3H). The mice transplanted with Ad-Rab26 islets exhibits hyperglycemia with glucose levels up to 25 mM, while the glucose levels of control mice were restored to normal (Figure 3G, 3H). The results suggest that overexpression of Rab26 not only restricted insulin secretion but also suppressed the rescuing function of isolated islets upon transplantation into type I diabetic experimental model.
Since Rab26 is involved in insulin secretion, the expression of Rab26 may be related to diabetes mellitus. Immuno-histochemistry analysis found the staining signals of Rab26 in islet of db/db mice are stronger than that of WT mice (Figure 3I), indicating high levels of Rab26 protein in islets of db/db mice. Furthermore, Rab26 protein level was elevated under high glucose or high palmitic acid conditions (Figure 3J, 3K). These results suggest that the expression of Rab26 is closely related diabetic pathophysiological conditions.
Rab26 associates with insulin granules and regulates the distribution of insulin granules
To investigate how Rab26 regulates insulin secretion, we examined the subcellular location of Rab26 and the effects of its over-expression on the distribution of insulin granules in β-cells. Rab26 was shown to be associated with the endocytic compartments 18. Immuno-fluorescence microscopy revealed GFP-Rab26 was present in puncta which co-localized with pro (insulin) and Vamp4 in MIN6 cells (Figure 4A), suggesting Rab26 associates with the secretory vesicles and insulin granules. Further examinations demonstrated that Rab26 associates with insulin-labeled granules (Figure 4B). In addition, over-expression of Rab26 resulted in insulin granules clustering in normal media, KRB buffer containing 2.8 mM glucose or 16.7 mM glucose in MIN6 cells (Figure 4B). Under normal conditions, over-expression of Rab26WT or Rab26Q123L mutant induced insulin granules clustering, while the inactivated mutant Rab26T77N was distributed in the cytosol and didn’t induce insulin granule granules clustering (Figure 4C). Consistent with the results that Rab26WT and Rab26Q123L but not Rab26T77N inhibit insulin secretion, this clustering distribution of insulin granules induced by Rab26 may inhibit glucose stimulation triggered exocytosis of insulin granules and consequently restrict insulin secretion.
Rab26 interacts with synaptotagmin-1
To uncover the molecular mechanism for Rab26 in regulating insulin secretion, we searched for its interacting proteins. Synaptotagmin-1(Syt1) along with several other proteins were identified as potential interacting partners of Rab26 from rat brain tissue lysate in large scale GST-pulldown experiments (Figure S3A). Syt1 is an important calcium sensor to regulate neurotransmitter release in neuron and insulin secretion in pancreatic β-cells 19, 20. The interaction of Rab26 with Syt1 was validated by analytic GST-pulldown assay. The interaction between Rab26 and Syt1 depends on Rab26’s nucleotide binding activity, as the interaction between GDP-bound mutant Rab26T77N and Syt1 was dramatically reduced (Figure 5A). In vitro binding assay using prokaryotic-expressed GST-Syt1 and His-Rab26 showed Rab26 can directly bind to Syt1 (Figure 5B). In addition, GST-pulldown experiments using GST-Rab26 revealed that Rab26 only interacts with Syt1-C2A domain, but not C2B domain in cell lysates derived from cells transfected with GFP-Syt1, GFP-Syt1-C2A or GFP-Syt1-C2B (Figure 5C), suggesting Rab26 binds directly Syt1 through interacting with the C2A domain in a manner dependent on its binding with GTP (Figure 5D). Besides Syt1, other members of Synaptotagmin such as Syt4 and Syt7 may regulate insulin secretion as well 21, 22. However, Rab26 did not interact with either Syt4 or Syt7 (Figure S3C), suggesting Rab26 specifically interacts with Syt1.
The interaction between Rab26 and Syt1 is essential to its inhibitory activity for insulin secretion
To investigate whether Rab26 restricts insulin secretion through interaction with Syt1, we first examined whether Rab26 associates with Syt1 at insulin granules. Immuno-fluorescence microscopy revealed that Rab26 is co-localized with Syt1 and insulin (Figure 5E). GFP-Rab26 also induced Syt1-containing insulin granules clustering in MIN6 cells under normal culture condition or glucose stimulation. These results imply Rab26 can associate with insulin granules by binding to Syt1, and consequently regulates exocytosis of insulin granules.
Over-expression of Syt1 promotes insulin secretion 23. We prepared Lentivirus-mediated over-expression of Syt1 in INS-1 cells, the cells then were infected with Ad-Rab26 (Figure 6A). Insulin secretion was monitored by ELISA assay. Indeed, over-expression of Syt1 enhanced insulin secretion in INS-1 cells (Figure 6B). However, insulin secretion enhanced by Syt1 was significantly suppressed when Ad-Rab26 was co-expressed with Syt1 (Figure 6B), suggesting Rab26 inhibits insulin secretion through interaction with Syt1.
By examining Genbank, we found that Rab26 gene encodes another shorter isoform (accession No. NM_001308053.1) referred to as Rab26b 24. Rab26b protein lacks N-terminal 66 amino acids. Interestingly, Rab26b was not able to interact with Syt1 in GST-pulldown experiments (Figure 6C), which was verified by using GST-Rab26 or GST-Rab26b to bind GFP-Syt1 (Figure S3B). In addition, Rab26b was not recruited to Syt1 containing insulin granules and was distributed primarily in the cytosol, not inducing granules clustering either (Figure 6D). As expected, over-expression of Rab26b had no significant inhibitory effects on glucose-stimulated insulin secretion in both INS-1 and MIN6 cells (Figure 6E, 6F). These results suggest N-terminal extension of Rab26 as compared to Rab26b is necessary for interaction with Syt1, clustering insulin granules or restricting insulin secretion. Taken together, these results suggest that Rab26 restricts insulin secretion via interacting directly with C2A region of Syt1 in a manner that is dependent on the N-terminal region and GTP-bound status of Rab26.
Rab26 influences the interaction between Syt1 and SNARE complex
Since Syt1 facilitates secretory vesicles docking to and fusion with plasma membrane through interaction with SNARE complex 25, we next examined whether Rab26 influences the interaction between Syt1 and SNARE complex. 293t cells were co-transfected with myc-Rab26 and GFP-SNAP25 or GFP-Syntaxin-1, the cell lysates were subjected for GST-pulldown using GST-Syt1. As shown in Figure 7A and 7B, the amount of SNAP25 protein bound to GST-Syt1 was significantly reduced in cells expressing Rab26 compared with control vector. However, expression of Rab26 had no effects on Syt1 binding to Syntaxin-1 (Figure 7C and 7D). Ca2+ stimulation promoted Syt1 binding to SNAP25, which was inhibited by Rab26 expression (Figure 7E and 7F). These results suggest that interaction of Rab26 with Syt1 inhibits Syt1 binding to SNAP25. Rab26 thus may restrict insulin secretion via sequestering Syn1 from interacting with SNAP-25.
As Syt1 can bind to phospholipids, we investigated whether Rab26 interaction with Syt1 influences Syt1 binding to phospholipids. In vitro overlay assay revealed that Rab26 did not affect the interaction of Syt1 with PI(4,5)P2 and PS (Figure S4). Therefore, Rab26 inhibits insulin secretory granules exocytosis probably by interfering Syt1 interaction with SNARE complex, and consequently inhibit insulin secretory granules fusion with the plasma membrane.
Rab26 inhibits exocytosis of newcomer insulin granules
Rab26 inhibits SNAP25 binding to Syt1, suggesting that Rab26 may influence Syt1 interaction with SNARE complex and consequently insulin secretory granules (ISGs) fusion with the plasma membrane. To examine this hypothesis, MIN6 cells were co-transfected with GFP-Rab26 and DsRed-Insulin, and then analyzed for the ISGs dynamics by employing time-lapse total internal reflection fluorescence microscopy (TIRFM).
The exocytotic events were, however, not uniform but could be categorized into three distinct modes of exocytosis. Pre-dock SGs that were visible before stimulation, no-dock newcomers SGs that fused without remaining at the plasma membrane for ≤ 200 milliseconds (interval of one frame), and short-dock newcomers SGs that appeared during stimulation and stably remained for > 200 milliseconds before fusion occurs. The number of pre-docked SGs was counted and averaged at the first 2 min prior to stimulation. At unstimulated state, the average number of pre-docked ISGs in Rab26-transfected cells is similar to that of vector-transfected cells (Figure 8A), when stimulated with 16.7mM glucose (containing GLP-1 and IBMX), the accumulated fusion events in Rab26-transfected cells were significantly less than that in vector-transfected cells (Figure 8B). Single SG fusion dynamic analysis revealed that no fusion events were observed before stimulation. Both the newcomer SGs no-dock and newcomer SGs short-dock were decreased in Rab26-transfected cells compared to that in vector-transfected cells, and not displaying two-phase secretion pattern in Rab26-transfected cells (Figure 8C), which is consistent with the results shown in Figure 2A and 2B. The fusion events summarized from newcomer no-dock SGs and newcomer SGs short-dock demonstrated Rab26 significantly inhibits the newcomer SGs fusion with the plasma membrane (Figure 8D). Taken together, Rab26 interaction with Syt1 may interfere Syt1 interaction with SNARE complex and consequently inhibit ISGs exocytosis, resulting in the inhibition of insulin secretion.