While molecular regulation of insulin granule exocytosis is relatively well understood, insulin granule biogenesis and maturation and its influence on glucose homeostasis are relatively unclear. Here, we identify a novel protein highly expressed in insulin‐secreting cells and name it BIG3 due to its similarity to BIG/GBF of the Arf‐GTP exchange factor (GEF) family. BIG3 is predominantly localized to insulin‐ and clathrin‐positive trans‐Golgi network (TGN) compartments. BIG3‐deficient insulin‐secreting cells display increased insulin content and granule number and elevated insulin secretion upon stimulation. Moreover, BIG3 deficiency results in faster processing of proinsulin to insulin and chromogranin A to β‐granin in β‐cells. BIG3‐knockout mice exhibit postprandial hyperinsulinemia, hyperglycemia, impaired glucose tolerance, and insulin resistance. Collectively, these results demonstrate that BIG3 negatively modulates insulin granule biogenesis and insulin secretion and participates in the regulation of systemic glucose homeostasis.
This study identifies a regulatory role for the novel protein BIG3 in systemic metabolism through its inhibitory effects on insulin granule biogenesis and insulin secretion.
BIG3 is associated with immature insulin granules in pancreatic β‐cells.
BIG3 restricts prohormone processing and insulin granule biogenesis.
BIG3 deficiency results in elevated insulin content and insulin secretion.
BIG3‐knockout mice show postprandial hyperglycemia and insulin resistance.
Insulin is the key molecule to regulate glucose homeostasis , and pancreatic β‐cells are the primary source of insulin production and release . Besides defects in insulin secretion, increased proinsulin levels are seen in patients with abnormal fasting glucose, impaired glucose tolerance, and overt diabetes and are used as a predictive marker for disease onset , , . Elevated proinsulin levels may be due to defects in proinsulin processing and disproportionate proinsulin release, both of which occur in immature secretory granules (ISGs), the major organelles that store and process proinsulin . Although insulin secretion process is well defined , the mechanisms underlying insulin granule biogenesis and maturation and its influence on systemic glucose homeostasis remain unclear.
The biogenesis of post‐Golgi secretory vesicles is under the tight control of Arf GTPases and coat proteins , , , . Precise temporal and spatial activation of Arf proteins is controlled by their corresponding Arf‐GTP exchange factors (GEFs) , , . Both the formation and maturation of ISGs require functional Arf and brefeldin A (BFA)‐sensitive Arf‐GEFs , . The GBF/BIG subfamily of proteins is the only known BFA‐sensitive Arf‐GEFs. Moreover, proinsulin exits the trans‐Golgi network (TGN) via clathrin‐coated compartments , , , which requires Arf1 and BFA‐sensitive Arf‐GEF , , , .
The insulin secretory granule contains a wide range of proteins that coordinate and regulate its behavior . Multiple proteomic studies reveal a large number of unknown or uncharacterized proteins that are associated with insulin granules , . While searching for novel regulators of membrane trafficking by database mining, we identified a large Sec7 domain‐containing protein, BIG3 as a novel insulin granule‐associated protein. Here, we report functional studies of BIG3 in insulin granule biogenesis, insulin secretion, and glucose homeostasis.
Results and Discussion
BIG3 is a novel insulin granule‐associated protein
In a search for novel regulators of the insulin secretion pathway, we found the gene kiaa1244 with a high tissue‐specific expression pattern. As the novel protein showed a significant sequence similarity to the GBF/BIG subfamily of Arf‐GEF proteins, it was named BIG3. BIG3 was highly expressed in pancreatic islet, and to a lesser extent the brain, but was undetectable in other tissues when examined using a BIG3‐specific antibody (Fig 1A and Supplementary Fig S1), consistent with previous observations . We confirmed specific co‐staining of BIG3 with insulin in the mouse endocrine pancreas, but failed to detect BIG3 in the exocrine pancreas or pancreatic ductal cells (Fig 1B and C). Furthermore, BIG3 predominantly localized to insulin granules of islet β‐cells, as revealed by immuno‐EM (Fig 1D). These results confirm BIG3 as a novel insulin granule‐associated protein and suggest that BIG3 may play a specific role in the insulin secretory pathway.
BIG3 has a negative regulatory role in insulin secretion
BIG3 was highly expressed in several insulin‐secreting β‐cell lines, including MIN6 (Supplementary Fig S2). To analyze the cellular functions of BIG3, we first generated stable BIG3 knockdown (BKD) and scrambled control MIN6 cells by using lentiviral infection and FACS. BIG3 protein level in BKD cells was efficiently suppressed, while the expression levels of selected constitutive or vesicle‐associated proteins were unaffected (Fig 1E). We next investigated the effect of BIG3 depletion on insulin secretion. BIG3 KD resulted in a marked increase in insulin secretion upon high glucose stimulation, and this effect was further exacerbated when combined with high K+ (Fig 1F and G). To examine whether increased insulin granule exocytosis contributed to the enhanced insulin secretion in BKD cells, we compared depolarization‐evoked membrane capacitance at the single cell level by using the whole‐cell patch‐clamp technique. The change in capacitance, reflecting exocytosis‐induced membrane expansion, was increased by approximately 44% in BKD cells compared to controls (Fig 1H and I). Importantly, expression of exogenous human BIG3 (hBIG3) in BKD cells completely reversed the KD phenotype, confirming that the observed insulin secretion phenotype is specific to BIG3 depletion (Fig 1J). Consistent with a negative role in insulin secretion, hBIG3 expression led to reduced insulin granule exocytosis (Fig 1J). Collectively, these results reveal that excessive BIG3 attenuates insulin granule exocytosis, while BIG3 deficiency promotes insulin granule exocytosis and insulin secretion in cultured β‐cells.
Generation of BIG3‐knockout mice
To study the in vivo effects of BIG3 ablation on islet function and consequent impact on systemic glucose homeostasis, we generated a BIG3‐knockout mouse (BKO) by targeting the 12th exon, the largest exon encoding the majority of BIG3's Sec7 domain (Fig 2A). BIG3 gene knockout was verified by PCR of the neo‐insertion from genomic DNA (Fig 2B), and protein absence was confirmed by Western blot in brain and islet lysates (Fig 2C and D) and by immunohistochemistry in pancreas (Supplementary Fig S3). Deletion of BIG3 had no effect on the expression of other BIG proteins and insulin (Supplementary Fig S4). BKO, generated initially on a C57BL/6J‐129/Sv mixed background, was backcrossed to 129/Sv mice for eight generations to generate a pure 129/Sv background. BKO mice were viable and fertile with normal body weight gain (Fig 2E). We also assessed the metabolic parameters by using the CLAMS system. No obvious difference was observed between BKO and control littermates in food intake, water consumption, respiratory exchange ratio, or energy expenditure (Fig 2F–I).
Increased insulin secretion in BKO islets
We next examined insulin secretion from isolated intact islets of BKO and control mice. Upon stimulation with high glucose, BKO islets exhibited a marked increase in first phase and total insulin secretion (Fig 2J and K). Depolarization‐evoked exocytosis from isolated BKO β‐cells increased approximately twofold compared to control at both early (immediately releasable pool) and late (releasable pool) time points (Fig 2L and M). EM analysis revealed that BKO β‐cells had approximately 35% increase in the number of insulin granules when compared to control cells, while the size of dense core was similar (Fig 2N–P). We also observed a 50% increase in islet insulin and proinsulin content when measured by ELISA (Fig 2Q and R). In contrast, no difference was observed in the overall islet morphology, islet area, islet number, or islet size between BKO and control mice (Fig 2S–V). Together, these results indicate that BIG3 deficiency leads to elevated insulin content, insulin granule number, and enhanced secretion upon stimulation at the cell and islet level.
Abnormal glucose homeostasis in BKO mice
We next analyzed whether the altered insulin secretion at the islet level affected proper glucose handling and systemic glucose homeostasis. When fed a normal chow diet, three‐month‐old BKO mice were hyperglycemic and hyperinsulinemic under postprandial conditions (both ad libitum and fasting‐refeeding), but not after 2‐h or overnight fasting (Fig 3A and B). Serum proinsulin levels and proinsulin‐to‐insulin ratio were the same (Fig 3C and D). BKO mice also displayed delayed glucose clearance during an oral glucose tolerance test (GTT) after overnight fasting (Fig 3E), even though insulin secretion was elevated (Fig 3F), suggesting reduced insulin sensitivity in BKO mice. Consistently, impaired glucose clearance was observed in BKO mice upon re‐feeding after overnight fasting (Fig 3G). To further ascertain the insulin sensitivity status, we performed hyperinsulinemic‐euglycemic clamp studies on two‐month‐old mice. BKO mice displayed approximately 1.5‐fold lower glucose infusion rate during clamping (Fig 3H), consistent with impaired glucose uptake. In response to hyperinsulinemia, hepatic glucose output (HGP) was reduced by > 75% in control mice, but only approximately 30% in BKO mice (Fig 3I). Overall glucose uptake rate doubled upon hyperinsulinemic clamping in control mice, though remained at basal levels in BKO mice (Fig 3J). Examination of peripheral tissues revealed that impaired glucose uptake was due to diminished muscle uptake, with approximately 80% decrease in soleus uptake and approximately 50% decrease in extensor digitorum longus uptake in BKO mice compared to controls (Fig 3K). Neither brown nor white adipose tissue was affected (Fig 3K). These results demonstrate that by 3 months of age, BKO mice have developed severe liver and muscle insulin resistance, possibly due to the chronic exposure of peripheral tissues to excessive insulin released by BKO islets. Our findings indicate that excessive insulin release may lead to severe peripheral insulin resistance and eventually disrupt glucose homeostasis in mice. This notion is consistent with the model that chronic hyperinsulinemia is the primary force driving insulin resistance, which has been well addressed by both in vitro and in vivo models.
BIG3 localizes to the TGN‐ISG compartment
To investigate how BIG3 modulates insulin secretion, we assessed BIG3 KD MIN6 cells. BIG3 predominantly co‐localized to insulin/chromogranin A (CGA)‐containing granules, in close proximity to the Golgi compartment, but distant from the plasma membrane (Fig 4A). CGA is a prohormone co‐sorted with proinsulin, and processed to β‐granin, which is co‐stored and secreted with insulin . Furthermore, BIG3 co‐localized with γ‐adaptin, a component of Adapter Protein I and a marker of clathrin‐coated vesicles. Consistently, there was little co‐localization of BIG3 with medial Golgi to TGN‐resident protein markers, such as Vti1a, Stx6, GS28, and GM130 (Supplementary Fig S5). These findings suggest that BIG3 may exert its inhibitory effect at the early stages of secretory granule biogenesis, where they remain physically close to the Golgi compartment with clathrin coats.
Granule biogenesis from the TGN can be promoted by stimuli and inhibited by cargo protein depletion . Therefore, we examined the redistribution of BIG3 upon treatment with high K+, which stimulates insulin release, and cycloheximide (CHX), which suppresses granule biogenesis, in reference to markers of various subcellular compartments (Fig 4B, C). After high K+ stimulation, the subpopulation of docked granules, characterized by CGA staining and distributed along the plasma membrane, were depleted. Conversely, freshly generated granules were still clustered close to the Golgi region, where BIG3 overlapped with CGA and γ‐adaptin (Fig 4B and C), suggesting that BIG3 is actively recruited to granules and influences secretory granule production. Following CHX treatment, CGA was mostly redistributed to the plasma membrane and did not overlap with BIG3 (Fig 4B). In contrast, BIG3 dispersed slightly, but remained in proximity to the Golgi region marked by γ‐adaptin (Fig 4C). The γ‐adaptin staining pattern did not change significantly after either treatment, indicating that the Golgi structure remained largely unaffected (Fig 4C). Taken together, these findings suggest that BIG3 is dynamically and predominantly recruited to the TGN where nascent immature granules are generated and whose abundance directly correlates with granule biogenic activity.
BIG3 negatively regulates insulin granule biogenesis
Both in vitro KD and in vivo KO experiments consistently demonstrate that regulated secretion is elevated in the absence of BIG3, consistent with the notion that BIG3 is a negative modulator of insulin secretion in β‐cells. Similar to other types of secretory granules, nascent insulin granules form from the TGN as ISGs with clathrin coats, followed by prohormone processing and removal of clathrin coats to generate a clathrin‐free mature secretory granule (MSG) . Therefore, there are typically two distinct populations of insulin/CGA‐marked granules: ISGs that cluster toward the Golgi region and MSGs that spread along the plasma membrane. The fact that BIG3 is predominantly associated with insulin granules concentrated to the clathrin‐patched TGN compartment suggests that BIG3 may be involved in insulin granule biogenesis. To explore the involvement of BIG3 in regulating nascent granule biogenesis, we assessed granule biogenic activity by measuring the prohormones and corresponding mature hormones. BIG3 KD in MIN6 cell resulted in a marked increase in cellular insulin (Fig 5A) and moderate increase in proinsulin (Fig 5B) under basal conditions. In addition, mature β‐granin markedly increased, along with a minor increase in the precursor CGA in BKD cells (Fig 5C and D). The proportional change of the two pairs of proteins indicates that BIG3 may have a general role in modulating secretory granules.
The decreased prohormone‐to‐mature‐hormone ratio may be due to enhanced prohormone processing. Therefore, we measured proinsulin processing at multiple time points in MIN6 cell lysates after 1‐h high K+ stimulation to promote and synchronize proinsulin synthesis, and another hour of CHX treatment to inhibit proinsulin synthesis and exocytosis. BKD cells had a significant decrease in proinsulin at 30 min, indicating that proinsulin was processed faster in BKD than in control cells (Fig 5E). The timing on proinsulin processing is in line with s previous study . We also determined CGA processing in BKD cells by pulse‐chase labeling of newly made proteins with 35S‐methionine and found faster conversion of CGA to β‐granin in BKD cells than in control cells (Fig 5F). Since granule production is coupled with prohormone production, these findings suggest that BIG3 is involved in a negative feedback mechanism to restrain granule biogenesis. The notion is further supported by the finding that BIG3 overexpression attenuated simulated exocytosis in MIN6 cells (Fig 1J). Collectively, these results suggest that BIG3 negatively regulates granule biogenesis: upon its removal, granule biogenesis and hormone production and storage are elevated. Consistent with preferential release of newly synthesized insulin granules , , , increased granule biogenesis leads to a larger pool of fresh insulin granules and consequently enhances secretion upon stimulation.
The molecular mechanism by which BIG3 modulates granule biogenesis is yet to be elucidated. It is well known that proinsulin transport requires Class I Arfs and their corresponding GEFs. GBF/BIG are the only known BFA‐sensitive Arf‐GEF, and multiple lines of evidence have shown that both secretory granule formation from the TGN and ISG maturation require Class I Arf activity in a BFA‐sensitive manner. This suggests that a yet to be identified BFA‐sensitive Arf‐GEF is required. It is worth noting that the single known functional domain in BIG3 by sequence analysis, the Sec7 domain, has a non‐functional catalytic motif . Thus, one of our hypotheses is that BIG3 acts as a competitive non‐functional Arf‐GEF, thereby negatively modulating granule production. The expression levels of BIG1/2 and GBF1 were not changed in BKO or BKD cells; however, whether their bioactivity in granule formation is affected remains to be studied. An alternative explanation could involve a mechanism similar to Arf4/Arf5, which binds to soma calcium‐dependent activator protein for secretion (CAPS) proteins at the TGN to promote dense core vesicle formation in a guanosine diphosphate (GDP)‐dependent manner, thereby bypassing the need for a functional GEF . It is not clear whether a similar mechanism applies to insulin granule formation and how BIG3 is involved. Currently, we are exploring direct evidence of BIG3 engagement in granule formation. It is worth noting that BIG3 is also expressed in the brain (Fig 1A). Whether BIG3 in the CNS contributes to the observed metabolic phenotypes remains to be determined.
In summary, our study shows that BIG3 is associated with immature insulin granules and negatively modulates insulin granule biogenesis in β‐cells and, in turn, insulin output by islets. Consequently, deletion of BIG3 in mice results in increased levels of circulating insulin and disrupted glucose homeostasis, including postprandial hyperglycemia and impaired glucose tolerance. Overall, our study identifies a novel regulatory mechanism for granule biogenesis and maturation that plays a significant role in modulating systemic metabolism.
Materials and Methods
All experiments involving animals were reviewed and approved by the Institutional Animal Care and Use Committee of A*STAR. Indirect calorimetry and other physiology tests were performed as previously described , , . For detailed information, see Supplementary Materials and Methods.
HL, WH, and WH designed the study; HL, SW, KC, and NG performed the experiments; HL, RE, AX, WH, and WH wrote the paper; all authors were involved in data analysis and approved the manuscript.
Conflict of interest
The authors declare that they have no conflict of interest.
Supplementary Figure S1
Supplementary Figure S2
Supplementary Figure S3
Supplementary Figure S4
Supplementary Figure S5
Supplementary Materials and Methods
Research in the laboratories of Hong and Han was supported by A*STAR Biomedical Research Council. A. Xu was supported by Hong Kong Research Grant Council (HKU4/CRF/10R).
FundingA*STAR Biomedical Research Council
- © 2014 Singapore Bioimaging Consortium, A*STAR