Molecular mechanisms of biphasic insulin secretion

(1)

UNIVERSITATISACTA UPSALIENSIS

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1118

Molecular mechanisms of biphasic insulin secretion

NIKHIL R. GANDASI

ISSN 1651-6206 ISBN 978-91-554-9281-6

(2)

Dissertation presented at Uppsala University to be publicly examined in B41, BMC, Husargatan 3, Uppsala, Friday, 11 September 2015 at 09:15 for the degree of Doctor of Philosophy (Faculty of Medicine). The examination will be conducted in English. Faculty examiner: Christien Merrifield (Laboratoire d'Enzymologie et Biochimie Structurales CNRS Bâtiment 34, France).

Abstract

Gandasi, N. R. 2015. Molecular mechanisms of biphasic insulin secretion. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine 1118.

42 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-554-9281-6.

Pancreatic beta-cells secrete insulin in response to increase in blood glucose concentration with a rapid first phase and slower, sustained second phase. This secretion pattern is similar in entire pancreas, isolated islets of Langerhans and single beta-cells and it is disrupted in type 2-diabetes.

Insulin stored in secretory vesicles has to undergo preparatory steps upon translocation to the plasma membrane which include docking and priming before being released by exocytosis.

A better understanding of the molecules involved in these steps is required to determine the rate limiting factors for sustained secretion. Here these processes were studied in real time using total internal reflection fluorescence microscopy, which enables observation of insulin granules localized at the plasma membrane. A pool of granules morphologically docked at the plasma membrane was found to be depleted upon repeated stimulations. Recovery of the docked pool of granules took tens of minutes and became rate limiting for sustained secretion.

Shorter depolarization stimuli did not deplete the docked pool and allowed rapid recovery of releasable granules. When a new granule arrived at the plasma membrane, docking was initiated by de novo formation of syntaxin/munc18 clusters at the docking site. Two-thirds of the granules which arrived at the plasma membrane failed to recruit these proteins and hence failed to dock. Priming involved recruitment of several other proteins including munc13, SNAP25 and Cav1.2 channels. Exocytosing granules were in close proximity to Ca²⁺ influx sites with high degree of association with Cav1.2 channels. This is because of the association of these channels to exocytosis site through syntaxin and SNAP25. During exocytosis the assembled release machinery disintegrated and the proteins at the release site dispersed. Syntaxin dispersal was initiated already during fusion pore formation rather than after release during exocytosis.

This was studied using a newly developed red fluorescent probe - NPY-tdmOrange2 which was the most reliable pH sensitive red granule marker to label insulin granules. Overall these data give new insights into the molecular mechanisms involved in biphasic insulin secretion.

Disturbances in the secretion at the level of granule docking and fusion may contribute to the early manifestations of type-2 diabetes.

Keywords: Exocytosis, Insulin secretion, Membrane trafficking, Microscopy, Diabetes Nikhil R. Gandasi, Department of Medical Cell Biology, Box 571, Uppsala University, SE-75123 Uppsala, Sweden.

ISBN 978-91-554-9281-6

urn:nbn:se:uu:diva-259071 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-259071)

(3)

Dedicated to my family and friends

(4)

(5)

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Gandasi, N.R., Yin, P. and Barg, S. (2015) Granule docking and priming are both rate limiting for insulin secretion in human β-cells. Manuscript

II Gandasi, N.R. and Barg, S. (2014) Contact-induced clustering of syntaxin and munc18 docks secretory granules at the exocy- tosis site. Nature Communications, 5:3914

III Gandasi N.R.^#, Yin P.^#, Riz M., Cortese G., Chibalina M., Rorsman P., Sherman A., Pedersen M.G. and Barg S. (2015).

Dynamic clustering of L-type Ca²⁺-channels during priming en- ables fast synchronized insulin secretion. Submitted manuscript IV Gandasi N.R., Vestö K., Helou M., Yin P., Saras J. and Barg S.

(2015). Survey of red fluorescence proteins as markers for se- cretory granule exocytosis. PLoS ONE 10(6): e0127801.

Papers not included in the thesis.

Krus U., King B.C., Nagaraj V., Gandasi N.R., Sjölander J., Buda P., Garcia-Vaz E., Ottosson-Laakso E., Storm P., Gomez M.F., Fex M., Vikman P., Zhang E., Barg S., Blom AM., and Renström E. (2014). The complement inhibitor CD59 plays a fundamental role in insulin secretion by controlling recycling of exocytotic fusion proteins. Cell Metabolism, 19:883- 90

Reprints were made with permission from the respective publishers.

(6)

(7)

Abbreviations

RRP readily releasable pool

IRP immediately releasable pool

SNARE soluble NSF attachment protein receptor

Syx syntaxin-1A

SM sec1/munc18-like

RIM rab3 interacting molecule

NPY neuropeptide Y

PIP2 phosphatidylinositol 4,5-bisphosphate NSF N-ethylmaleimide-sensitive factor

Ca²⁺ calcium ion

TIRF total internal reflection fluorescence EGFP enhanced green fluorescent protein

CD caging diameter

FWHM full width at half maximum

(9)

Introduction

Exocytosis is a fundamental cell biological process where vesicle cargo is released by fusion with the plasma membrane. There are two types of exocytosis namely. regulated and constitutive exocytosis. Regulated exocytosis is triggered by Ca²⁺and constitutive exocytosis is triggered without any external stimulus. Transport or delivery of newly synthesized proteins in many cell types occurs through constitutive exocytosis. Regulated exocytosis requires an external stimulus and occurs in mammalian cell types like pancreatic islet cells, neurons and immune cells. They secrete hormones, neuro- transmitters, and cytotoxins that make up the vesicle cargo in these cells. In the present study, pancreatic beta-cells that release the blood-glucose lower- ing hormone insulin was used as a model to study regulated exocytosis.

Insulin synthesis, processing and exocytosis

Insulin is produced in beta-cells of the islets of Langerhans in the pancreas and secreted according to the metabolic demands of the body. Insulin is synthesized at the rough endoplasmic reticulum in the form of pre-pro-insulin.

The amino terminal signal is removed before the export from the endoplasmic reticulum to the Golgi to form proinsulin. It is then packaged into imma- ture secretory granules which acidify gradually in the trans-Golgi network.

In granules proinsulin is cleaved by several proteases to form insulin [2].

Mature secretory granules are the most abundant organelles in beta-cells and fill out much of the cytosol. The granules translocate to the plasma membrane where they can be released by regulated exocytosis.

Insulin is released in response to an increase in the blood glucose concentration. Glucose is taken up by beta-cells through GLUT transporters and then metabolized in the cells to yield ATP. The resultant increase in the ATP/ADP ratio closes ATP sensitive K⁺- (K-ATP) channels, leading to gradual depolarization of the plasma membrane. Ca²⁺ then enters through voltage dependent calcium channels resulting in increase of the cytosolic Ca²⁺ concentration which triggers exocytosis of insulin granules [3].

(10)

Insulin secretion in response to a step increase in glucose has a biphasic time course, with a rapid, pronounced first phase lasting about 5-10 min followed by a slower but less pronounced second phase, which can be mostly resolved into regular pulses. This secretion pattern is characteristic of the entire pancreas, isolated islets and single beta-cells alike. The pattern is altered in early stages of the development of type-2 diabetes with a blunted first phase and irregular pulsatility [4,5]. However, the underlying mechanisms are not un- derstood. Theoretical average insulin granule release rates during the first phase is approximately 20 insulin gr*min^-1 (granules per min) and 6 gr*min^-1 during second phase [6]. Although these rates have been confirmed in murine beta-cells [7] the rates remain inconclusive in human beta-cells. Capaci- tance measurements of exocytosis in murine beta-cells demonstrated that only about 100 of the nearly 10,000 granules can be released immediately in response to high Ca²⁺[8,9,10]. The number of releasable granules in human beta cells is approximately 300 [11]. The Ca²⁺ concentration required for exocytosis of these granules is 10 times higher than the average Ca²⁺ concentration seen in stimulated beta-cells [12,13,14]. Such high Ca²⁺ concentrations are believed to be achieved only at the mouth of the voltage dependent calcium channels. Ca²⁺entry from the voltage dependent calcium channels generates a spatial Ca²⁺gradient where there is at least 3 fold higher Ca²⁺ at the exocytosis sites compared to bulk of the cytoplasmic Ca²⁺. The granules that release immediately in response to voltage dependent elevation of intracellular Ca²⁺ concentration are defined as the readily releasable pool of granules (RRP). These granules account for less than 1% of the total granules in a beta cell. The RRP is quickly exhausted, but gradually replenished from a

“reserve pool” of granules [8,9,10,15,16]. These granules are in the vicinity of the plasma membrane but have to undergo several steps known as priming before they become release ready (Fig 1). The availability of release ready/primed granules thus determines the amplitude of first phase and lim- its the rate of secretion during second phase.

A small subset of RRP granules which account for the first few granules released upon influx of Ca²⁺ are assumed to be located in the immediate Fig 1. Life cycle of a secretory granule at the plasma membrane. Secretory vesicles translocate to the plasma membrane where they dock. Docked granules undergo ATP dependent steps referred to as priming to form a readily releasable pool (RRP) of granules. In beta-cells entry of Ca²⁺ through voltage gated calcium channels trigger exocytosis with resulting release of insulin (modified from [1]).

(11)

vicinity of the Ca²⁺ channels and are referred to as immediately releasable pool (IRP)[17]. These granules undergo exocytosis with rates that are transi- ently as high as 650 granules*s^-1 compared to 500 granules*s^-1 for RRP granules. This suggests that IRP granules experience Ca²⁺ concentrations well above the average observed in the bulk of the cytosol (_˜2µM) [12,14].

Nanodomains, where Ca²⁺ reaches tens to hundreds of µM, are expected to develop near the mouth of open Ca²⁺-channels. This is consistent with a model where IRP granules are situated in the vicinity of voltage-dependent Ca²⁺-channels, where domains of high Ca²⁺ develop [17,18,19]. However, direct evidence for these channels clustering at the release site is lacking. L- type Ca²⁺-channels conduct the majority of the Ca²⁺ that enters the beta-cells.

Mouse β-cells express predominantly the LC-type channel (CaV1.2) [17,20]

while rats and humans express L_D-type channel (CaV1.3) [21,22,23]. These channels have been reported to bind proteins of the exocytosis machinery, such as syntaxin, synaptotagmin, RIM and RIM binding protein [24,25,26,27,28,29,30]. The interaction site of neuronal Ca²⁺-channels is a loop between its transmembrane domains II and III, which is often referred to as synprint site. A peptide derived from the corresponding region of the L_C-type channel ablates fast secretion [17] in beta-cells and in neurons [31,32,33]. The Synprint peptide from the LC but not LD-type channels binds to the SNARE (soluble N-ethylmaleimide sensitive factor attachment protein receptor) proteins syntaxin and SNAP25 [30,34]. This suggests that the LC- type channels expressed in mouse beta-cells interact with the SNARE proteins and overexpression of synprint peptide blocks this interaction by in- creasing the distance of the exocytotic machinery to the Ca²⁺ influx site.

Secretory granules visualized in electron micrographs of beta-cells [35] and other secretory cells [36,37,38,39] show a subset of granules which are in immediate vicinity to plasma membrane and are referred to as “morphologically docked” granules. These granules are observed to have restricted mo- bility in the fluorescent field in live cells [40] suggesting a physical binding to the membrane [41]. Upon stimulation, more than half of the granules in the vicinity of plasma membrane were lost in chromaffin cells [39]. This shows that the exocytosis occurs from the pool of docked granules [15,42,43], although this has been challenged by reports of granules that fuse immediately after arrival at the plasma membrane [44,45,46,47].

At least 10% of the granules in EM micrographs of beta-cells of the approximately 10,000 granules are docked. Not all docked vesicles are fusion competent since the primed pool is limited to a only few hundred granules [48].

Hence granules can be classified to docked and primed pools based on their functionality and distance to the plasma membrane (fig 1) [49,50,51].

Primed granules in the vicinity of the plasma membrane are released during the first phase in beta cells stimulated with glucose. During the second phase

(12)

the release rate is slowed down due to replenishment of primed granules in an ATP-dependent process [52]. This biphasic time course observed during glucose stimulation results in exocytosis from pre-docked granules [8,15].

Repeated stimulations by patch clamp depolarizations result in a larger fraction of pre-docked granules being released, despite this a large fraction of docked granules are left at the plasma membrane [8,53]. A small fraction of docked granules (<15%) remain resistant to repeated depolarizations since they are in dead-end docked state [53]. The requirement of docking and priming (sequential model) as a prerequisite for exocytosis remained un- doubted [8,15] until it was reported that granules could fuse immediately upon arrival at the plasma membrane [7,46,47,54,55]. This pattern of exocytosis called crash fusion is mostly observed during the second phase of glucose stimulated insulin secretion owing to limited availability of docked granules during this phase (Fig 2). In concert with this observation there have also been reports of docking being a constraint for granules which are release ready [56]. In contrast granules in the sequential model stay at the plasma membrane and hence have a chance to recruit priming factors and become release ready assuming that they have not been primed before arrival at the plasma membrane. Both models are based on results from rodent primary beta-cells and chromaffin cells but have never been reported in human beta-cells.

Proteins involved in exocytosis

The molecular and physiological mechanisms of membrane fusion is similar in different types of cells. Most of the proteins involved in neurotransmitter release have also been identified in beta cells and perform a similar role during exocytosis [57]. SNARE proteins drive the membrane fusion. Syntaxin and SNAP25 are the SNARE proteins localized in the plasma membrane (target membrane), referred to as t-SNAREs. VAMP2 is a SNARE protein localized to the vesicle membrane and is thus called a v-SNARE. The SNARE domains of syntaxin and SNAP25 in the target membrane and that of VAMP2 on the vesicle membrane form a zipper which can progressively overcome the energy barrier of membrane fusion [58,59]. Several studies suggest that about 3 SNARE complexes are sufficient to drive fusion of a Fig 2. Granule availability for fusion. 1.

An incoming granule sequentially follows docking and priming steps before fusion 2. An incoming granule undergoes fusion without passing through the docking (crash fusion) as suggested in the alter- nate model for fusion.

(13)

vesicle [60,61,62,63,64,65]. This is consistent with the energy requirements of the fusion reaction and the energy released during SNARE complex formation [66].

Synaptotagmin is the Ca²⁺ sensor of exocytosis and has a dual role, acting both during priming and fusion. Synaptotagmin acts as a clamp such that the zipper formed by the trimeric SNARE complex initially is blocked but become released upon binding of Ca²⁺ to the C2 domains of synaptotagmin [67]. In addition, it has been proposed that Ca²⁺ independent binding of synaptotagmin to the membrane phospholipid PIP2 (phosphatidylinositol 4,5- bisphosphate) can position the vesicles such that they can interact with syntaxin and form a SNARE complex [68,69]. This view has been supported by the observation that PIP₂ can form microdomains in the plasma membrane.

These microdomains partially co-localize with syntaxin clusters, which increases upon increase in PIP₂ at the plasma membrane [68,70,71]. Syntaxin clusters are made up of 50-70 molecules of syntaxin and these molecules in the cluster are in rapid equilibrium with unbound molecules [72,73,74].

Clusters of SNAREs are associated with docked granules and facilitate exocytosis [75].

All SNARE proteins contain at least one SNARE motif, VAMP2 and syntaxin also have a membrane anchoring transmembrane region. Syntaxin is the only SNARE protein possessing an additional N-terminal domain, which folds over the SNARE domain keeping syntaxin in a closed conformation [76,77]. The N-terminal domain, which accounts for 60% of the protein is composed of a short N-peptide and a Habc domain [78,79]. Munc18 binds to syntaxin through its N-terminal domain and the SNARE domain [80,81,82,83,84] by forming an intramolecular coiled coil between the SNARE domain, N-terminal Habc domain [85,86]. This results in conformation change at Habc domain of syntaxin making it accessible for formation of the SNARE complex. Munc18 is stabilized by binding to the Habc domain and munc18-SNARE complexes mediate different functions during synaptic vesicle exocytosis. Deletion of the Habc domain resulted in decrease in the size of the RRP at the synapse [87]. Furthermore, deletion of the N-terminal domain of syntaxin abolishes clustering of protein at the granule sites [75]. The SNARE-munc18 protein complex disassembles after fusion and is thought to be recycled for future fusion processes. The recycling is done by proteins such as N-ethylmaleimide sensitive factor (NSF) and soluble NSF attachment protein alpha (αSNAP) so that the individual SNARE components are made available for the future fusion processes [51].

SNARE proteins further coordinate with accessory proteins to facilitate exocytosis [59,88,89].

(14)

Munc13 is another SM (sec1/munc18-like) protein implicated in neurotransmitter release [90]. Expression of the MUN domain of munc13 alone can rescue exocytosis in munc13 knockout mice [89]. Munc13 is involved in the transfer of vesicles to RRP during priming [91,92,93]. This function is regulated by Ca²⁺ and a Rab3a interacting protein called RIM [94]. Knock- down of RIM leads to severe defects in priming, which can be rescued by a mutant of munc13 [95]. The MUN domain of munc13 increases the transi- tion rate between syntaxin-1/munc18 complex to SNARE complex [96,97].

Rab3a is a G-protein that plays a role in late steps of exocytosis. Rab3a cy- cles between an active form when bound to GTP and an inactive form when bound to GDP [98]. Hydrolysis of GTP on rab3a by the intrinsic GTPase activity leads to dissociation of rab3a from the vesicle membrane during exocytosis [99]. Rab3a binds directly to munc18 and has been shown to have a major role during docking [100,101]. Increase in its expression leads to increase in the number of docked granules [102] and a decrease in the number of docked granules in rab3a knockdown [103]. Consistent with the role of rab3a in membrane trafficking its expression is highest in brain and endocrine cells [104]. Rab3a interacting molecule (RIM) has been implicated in vesicle fusion [105]. RIM is involved in tethering of voltage gated calcium channels to the release site. As discussed above this is critical for coupling of influx of Ca²⁺ to fusion [106]. The stoichiometry and time course of recruitment of these proteins to the release site remains unknown.

The SNARE and SM proteins syntaxin 1 [7,107,108,109], SNAP25 [37,108], VAMP2 [37], Munc 18 [37,110] and synaptotagmin [37] have a role in granule docking apart from their role in exocytosis. Knock down of any of these proteins leads to severe defects in docking. Knockdown of syntaxin abolished exocytosis and markedly reduced the number of docked granules [107,110]. Syntaxin without the Habc domain does not co-localize with docked granules [75] compared to full-length syntaxin which has a very high degree of co-localization [75,111].

Kinetics of exocytosis

A transient pore is formed when vesicle membrane and plasma membrane fuse during exocytosis [8,112,113,114,115,116,117]. This initial pore has been shown to be too narrow to allow release of the relatively large insulin molecule [8,116,118]. In endocrine cells including beta cells, the pore can open and close several times before release of the granule content [8,119,120,121]. Dilation of the pore can lead to complete cargo release and the fusion pore dynamics may be influenced by several external stimuli

(15)

[116,122]. Fusion pore lifetime and size have been studied by patch clamp techniques in beta-cells [115,116,118,123].

Total Internal Reflection Fluorescence (TIRF) microscopy of labeled insulin granule proteins has enabled the study of formation and dissipation of the fusion pore [118,124,125]. A better understanding of molecules involved in the regulation of the fusion pore is limited by the availability of probes to study the pore kinetics. The pH of the vesicle lumen is acidic but changes to neutral when the pore establishes contact between the luminal content and the extracellular fluid. This pH change can be used to follow the opening and collapse of the pore with pH-sensitive fluorescent proteins with near neutral pKa. PHluorin with a pKa of 7.6 [126,127], is often used to study the kinetics of exocytosis process [128,129,130,131,132]. Although fluorescent proteins with near neutral pKa are readily available in the green spectrum, red- emitting pH-sensitive fluorescent proteins are rare. They are also often not well characterized as reporters of exocytosis. Red-emitting pH-sensitive fluorescent proteins offer better spectral separation during dual and triple color applications with blue and green fluorescent proteins. This is beneficial in the triple color imaging of the behavior of proteins involved in exocytosis and would give a better understanding of the exocytotic machinery, since the fusion pore is thought to be formed due to conformational changes in the SNARE complex [133,134].

In order to follow the life cycle of the insulin granule during docking, priming and fusion, we use TIRF microscopy. TIRF microscopy allows studies of fluorescent molecules within 100 nm of the plasma membrane. The spatial organization of various proteins at the granule site was studied to understand the mechanisms involved in biphasic insulin secretion.

(16)

Methods

Cells and human islets for studies of insulin secretion

Murine beta cell lines are commonly used to study insulin secretion. Even though there are many murine beta cell lines, very few of them stably mimic the insulin secretion achieved in freshly isolated primary beta-cells. In response to glucose rat insulinoma - INS1 cells exhibit 2 to 4 fold increase in insulin secretion [135] which in comparison to the 15 fold increase in primary cells [136] is small. A clone of INS1 cells called 832/13 has been reported to secrete 13 fold increase in insulin secretion in response to glucose stimulation [137]. Hence this was used as a model for studying insulin secretion.

We used human pancreatic islets (paper I and III) kindly provided through the Nordic Network for Clinical Islet Transplantation, Uppsala facility. INS1 cells were transfected with plasmid expression vectors and the dispersed human islet cells were infected with adenoviral expression systems.

Microscopy

TIRF microscopy

- Insulin granule exocytosis can be visualized in real time using fluorescence microscopy. Exocytosis of insulin can be visualized at the plasma membrane of the cell which is >10nm in thickness. Detection of fluorophores labelled to monitor exocytosis improves with better signal to noise ratio and elimination of fluorescence from the background. TIRF microscopy uses an evanescent wave to selectively illuminate the fluorophores in a 100-150 nm thin layer of the cell surface adhered to the glass coverslip [138]. Cells transfected with fluorescently labelled proteins of interest were imaged using a custom-built lens-type TIRF microscope with a resolution of 0.16 μm per pxl with red (excitation 561 nm/emission 610 nm) and green channels (excitation 488 nm/emission 523 nm).

Single molecule imaging

– TIRF microscopy can be used to visualize live cells with a resolution of 10-20 nm. This technique was exploited to study the behavior of single molecules of syntaxin and L-type calcium channels in relation to insulin granules. Single molecules were identified based

(17)

on characteristic single step bleaching observed in cells expressing very low levels of the EGFP tagged protein.

Confocal microscopy

– The distribution of soluble proteins with respect to granules localized in the cytoplasm was visualized using confocal microscopy. In this technique elimination of out of focus light results in higher contrast images of the focal plane. Confocal imaging was done with a Zeiss LSM780 microscope using a ×63/1.40 objective (Zeiss) with sequential scanning of the red (excitation 561 nm, emission 578–696 nm) and green channel (excitation 488 nm, emission 493–574 nm).

Fluorescently tagged proteins for imaging

– Dual or triple color fluo- rescence microscopy depends on well-defined spectral separation of the fluorophores used. The combination of EGFP and mCherry are widely used for two-color simultaneous imaging due to their well separated excitation and emission spectra. Therefore neuropeptide Y (NPY) tagged to mCherry and EGFP tagged to proteins involved in exocytosis machinery were transi- ently co-transfected for dual color fluorescence microscopy applications.

Fluorescent proteins with near neutral pKa are beneficial in sensing the change in pH between acidic and neutral during exocytosis. EGFP has a pKa of 6 and is hence very beneficial to study the kinetics of exocytosis [8,129,139,140,141]. Red-emitting pH-sensitive fluorescent proteins are rare and often not well characterized for dual color applications as reporters of exocytosis although some of them have been characterized for endocytosis [142,143].

Image analysis

Visualizing insulin granule dynamics

– Cells with NPY-tagged fluo- rescent protein containing granules were imaged with TIRF microscopy.

Granules appeared punctate in the TIRF field and most of them were station- ary. A small fraction of granules appeared in the TIRF field and became arrested after reaching their highest brightness, likely because they became bound to the plasma membrane. Granules that approach the TIRF field and become laterally confined for at least 2 frames can either dock or just visit the plasma membrane. We define docking as granules that remain confined at least 60-90 s; visitors were those granules that remained for <60 s after appearing at the plasma membrane. Granules that are not stable at the plasma membrane and lose its fluorescence over a time of 1-2 s and disappear from the TIRF field are referred to as undocking. Exocytosis events in comparison were characterized by rapid loss of the granule marker fluorescence within 0.2s and were visualized when the cells were depolarized with high K⁺. Exo- cytosis with NPY-Venus and -EGFP as granule markers were preceded by an

(18)

increase in fluorescence lasting up to several seconds due to the pH- dependent dequenching that occurs when the fusion pore opens [8]. This increase was not seen in cells expressing pH insensitive constructs such as NPY-mCherry, although the rapid loss of the granule fluorescence remained a reliable marker of exocytosis.

Quantifying insulin secretion

- Insulin secretion was measured by de- tecting the frequency of exocytosis events from at least 10-15 cells per con- dition. Insulin secretion is often measured through enzyme-linked assays, which measures the secreted insulin. They are limited by temporal resolution and require a large number of cells. Capacitance measurements provides very high temporal resolution and allow detection of individual vesicle fusion events in single cells [1,144]. However, this technique is somewhat invasive and only allows recordings from one cell at a time. In the present work, insulin secretion was measured by monitoring the frequency of exocytosis events in the TIRF field over time. The optical imaging technique mim- ics capacitance measurements in providing high temporal resolution. Exocy- tosis was evoked by high K⁺ applied by computer-timed local pressure ejec- tion through a glass pipette. This protocol bypasses the metabolic aspects of glucose stimulation and is generally thought to mimic the 1^st phase of insulin secretion [145]. Indeed, after K⁺ stimulation many cells initially show a high frequency of exocytosis that decreased after the first few minutes, somewhat reminiscent of biphasic insulin secretion [124].

Quantification of proteins at the release site

- To quantify the degree of co-localization of proteins to the granule site INS1 cells were cotransfect- ed with granule marker NPY-mcherry and various EGFP-tagged protein constructs. Syntaxin 1 (Syx1) clusters that colocalize with morphologically docked granules at the plasma membrane is observed in a variety of cells [75,146,147,148]. We confirmed that Syx1-EGFP aggregated in small clusters that appeared diffraction limited in size (<0.2µm) and that they were overlaid on a hazy background of diffuse Syx1-EGFP fluorescence. To quantify this specific association, we applied “location-guided averaging”

[74,75], somewhat analogous to classical binding assays. Fluorescence from granule-bound syntaxin1-EGFP (∆F) and free syntaxin molecules in the surrounding membrane area (S) was measured; their relationship follows a simple one-site binding model in the form ∆F = Bmax*S/(Kd + S), where the initial slope ∆F/∆S (approximated by ∆F/S for S<<k_D) is proportional to the binding affinity (Bmax/kD) to the granule site. Small images spatially aver- aged clearly showed a syntaxin1 spot at the granule site, but not at random locations [75]. Similarly, other proteins enriched at the granule site to vary- ing degrees. Rab3a, Rab27a, RIM, syntaxin3 was similar to syntaxin1 whereas munc18 and munc13 resulting in about half of this affinity.

Granuphilin and rabphilin accumulated much more than syntaxin1. Ca²⁺-

(19)

channel subunit CaV1.2 and NSF accumulated to one third of the level of syntaxin1. SNAP 25, rab7a, NCX1, alpha-SNAP, and a probe for phosphatidylinositol 4,5-biphosphate (PIP₂) [149] accumulated to one fourth of the level of syntaxin1, but they were all significantly enriched at the granule site when compared with random regions in the PM. We conclude that a number of proteins, probably not limited by those tested here, accumulate at the docking site (Fig 3).

Fig 3. Quantification of proteins at the docking site - A Quantification of granule associated fluorescence (c-a=∆F), and nonspecific fluorescence (a-bg=S). The ratio

∆F/S is proportional to the proteins affinity to the docking site for the proteins me- tioned in the figure. B Average images of the same constructs at the granule site or at random locations (rand). C Images of cells expressing the constructs indicated in A.

(20)

Aims

This thesis addresses the following questions:

I. How does granule docking and priming relate to phasic insulin secretion?

The availability of release ready granules determines the rate of biphasic secretion. Therefore the recovery of this pool is crucial for sustained secretion. The aim of this study was to investigate if the preparatory steps which include docking and priming; become rate limiting for the recovery of release ready granules.

II. How is the exocytosis machinery assembled during docking and priming?

Binary acceptor complexes of syntaxin and SNAP25 have been proposed as docking receptor [37,148]. Granules that bind to these docking receptors are assumed to recruit other proteins involved in exocytosis upon docking (Fig 4). Direct evidence for this model of docking is lacking. The aim of this part of the work was to study the mechanism of docking and recruitment of these proteins before exocytosis in real time using total internal reflection fluorescence microscopy.

Fig 4. Alternative mechanisms of docking tested. 1. The incoming granule docks randomly to form its own nanodomain which recruits other proteins before its release or 2. The incoming granule should dock at a preexisting release site with SNARE proteins

(21)

III. What is the basis for synchronization between electrical activity and exocytosis?

Electrophysiological and biochemical data suggest that L- and P/Q-type channels functionally couple to the proteins which are part of the exocytosis machinery [24,25,26,27,28,29,30]. Using high resolution microscopy com- bined with spatiotemporal mathematical modeling we aim to see how the vicinity of exocytosing granules to localized Ca²⁺ influx sites influence the granule’s probability to undergo exocytosis. We also study the association of these Ca²⁺ influx sites to Ca²⁺ channels.

IV. Which is the best granule marker to study the kinetics of exocytosis and release?

NPY tagged fluorescent proteins are valuable tools for studying vesicle trafficking processes like exo and endocytosis. There are very few red-emitting pH sensitive fluorescent proteins that sense the change in pH during these processes and therefore beneficial in dual color applications. We test a panel of red fluorescent proteins to identify the best granule marker for study of exocytosis.

(22)

Results and Discussion

Depletion of the docked granule pool becomes a rate limiting factor for sustained secretion (Paper I and II).

A large fraction of ~10,000 total granules in beta-cells are fusion incompe- tent at any given point of time [35]. To determine what fraction of granules at the plasma membrane is fusion competent, we stimulated human islet cells with glucose (10mM) or high K⁺(75mM). Stimulation with glucose resulted in high frequency of exocytosis (0.12 gr*s^-1) followed by lower frequency of exocytosis (0.02 gr*s^-1) for 200 s after stimulation. This is reminiscent of pronounced insulin secretion observed initially in response to glucose in mouse beta-cells [52,150]. Similar to previous studies in human beta cells [151] short bursts of exocytosis events interrupted by intervals without secretion were observed in a large fraction of the cells. More than a third of the docked granules were lost at the end of the experiment. To bypass the metabolic aspects of glucose stimulation, high K⁺ stimulation was applied resulting in shorter stimulation time. This resulted in kinetics similar to first phase during glucose stimulation but half of the docked granules were lost already within 60 s. These results are consistent with data from patch clamp depolarization stimuli in some of the previous studies [35,152] where it took 60-120 s to recover the high frequency of exocytosis in subsequent depolarization stimuli [48]. Here high K⁺ stimulation was repeated on the same cell instead of patch clamp depolarizations in previous studies. A larger pool of docked granules were present before stimulation which resulted in high frequency of exocytosis (0.9 gr*s^-1) during the initial stimulation. In contrast subsequent stimulation of the same cell resulted in decreased frequency of exocytosis (0.1 gr*s^-1). In concert with this there were fewer docked granules due to its depletion during first stimulation which could not be recovered in time (2 min) for subsequent stimulation. This would mean that the rate of docking (0.05 gr*s^-1) cannot keep pace with the rate of exocytosis (0.9 gr*s^-1). In agreement with this prediction, the rate of exocytosis was still not complete- ly recovered 20 min after the initial stimulation (0.6 vs 0.9 gr*s^-1). Hence recovery of the docked pool of granules requires tens of minutes and is rate limiting for sustained secretion.

These results indicate that docking is a pre-requisite for sustained secretion and this conclusion is consistent with previous studies [8,15]. However, this

(23)

view has been challenged by evidence that new incoming granules can release without docking, a process referred to as “crash fusion” [7,54,55] (fig 2). To further clarify whether docking is a pre-requisite for fusion we tracked granules in INS1 cells with TIRF microscopy for 120 s to check if they were confined to their position under non-stimulatory conditions. This was followed by stimulation with high K⁺to trigger exocytosis. This experiment demonstrated that granules which underwent exocytosis in response to K⁺ had been docked and immobile for at least 25 s before fusion. In contrast, granules which showed movement before stimulation did not undergo exocytosis. New incoming granules (granules which were not seen in the TIRF field in the beginning of the observation period) showed a high degree of movement in the sub-membrane space but did not undergo exocytosis within the time frame of the experiment. These data suggest that docking indeed is a prerequisite for exocytosis. Crash fusion was reported mostly during glucose stimulation [7], which typically requires extended observation periods.

One possible explanation for the discrepant results may be that the granule label progressively bleaches making the granules difficult to detect. Howev- er, when the granules exocytose, they would increase in brightness due to the dequenching that occurs when the label is exposed to the neutral extracellular pH. In summary, granule docking is a prerequisite for exocytosis and recovery of the docked granule pool may therefore become a rate limiting factor for sustained insulin secretion.

Conversion of docked granules to RRP is slow (Paper I).

Upon stimulation with high K⁺ a subset of docked granules responded resulting in exocytosis of these granules (responders) in human beta-cells. Long repeated depolarizations resulted in fusion of about half of the docked granules. The remaining granules underwent exocytosis during subsequent depolarizations but not during the first stimulation. Although there was availability of docked granules in subsequent stimulations there was a decrease in frequency of exocytosis. This observation is consistent with the results from INS1 [8] and PC12 cells [53] and suggests that docked granules are not necessarily fusion competent but they have to be primed before they can undergo exocytosis. The rate of priming was determined by the time taken for a set of newly docked granules to attain release competence. The release probability was calculated and expressed as a function of residence time of the newly docked granules at the membrane before exocytosis. Half- maximal release probability was reached after 3 min. This rate is slower than the previous estimates of ~1 min for RRP refilling in mouse beta cells [48,153]. This difference might be due to differences in the depolarization protocols. Thus, long depolarizations used in the present study might be

(24)

more efficient in depleting the docked and primed granule pool than shorter depolarizations used in previous studies. Short depolarizations (<2 s) neither depleted the docked granules nor reduced the frequency of exocytosis during subsequent stimulations. Hence the supply of primed granules may be sufficient during short stimulations [10]. A small fraction of the docked granules never undergo fusion, which would mean that they never get primed for release and hence are in a dead-end docked state [53]. In spite of this, docking and priming determine the availability of release ready granules. From the previous experiments recovery of docked pool would take tens of minutes, in addition to this they have to be primed to become release competent. Therefore during long stimulations availability of release ready granules becomes rate limiting for sustained secretion.

Mechanism of docking (Paper II)

Two-thirds of the granules that arrived at the plasma membrane did not re- side at the membrane for more than 20 s and only one third of the granules had a longer residence resulting in stable docking. Currently it is believed that granules that arrive at the plasma membrane bind to pre-determined docking sites. These sites are thought to be binary acceptor complexes made up of SNARE proteins localized to plasma membrane, syntaxin and SNAP25 which act as docking receptors. Incoming granules with VAMP2 has been proposed to bind to these docking receptors to form a stable docking complex [37,69,148]. When we simultaneously imaged granules and syntaxin with TIRF microscopy we did not find pre-existing syntaxin clusters to which incoming granules docked. Instead, syntaxin clusters were formed upon granule arrival at the plasma membrane. When syntaxin clustering was prevented by overexpression of the Habc fragment, stable docking was in- hibited. In neuronal cell types the Habc domain of syntaxin directs the protein to granules [75,154] and deletion of this domain results in fewer releasable vesicles [87], this might be due to the strong docking phenotype similar to our observation in INS1 cells. Munc18 clusters were formed on a similar timescale as syntaxin and these proteins co-localize at the granule site. Only one third of the granules, which visited the plasma membrane, docked and induced de novo formation of syntaxin/munc18 clusters. No clusters were observed at granules that visit the plasma membrane and they failed to dock.

This observation suggests that granules may carry molecular cues that initi- ate syntaxin/munc18 recruitment and which confer weak interactions between the granule and the plasma membrane during the tethering phase [41].

Candidates for such molecules are, the small GTPases Rab3 and Rab27.

They have been demonstrated to be involved in docking [103] and to interact with the rab3 effector protein RIM [155]. Consistent with this idea, at least some of the granules arriving at the membrane carried munc18 and rab3

(25)

which could bind syntaxin to bridge the two membranes. Another candidate is synaptotagmin, which can bridge vesicle and target membranes in a lipid- and Ca²⁺-dependent manner [156,157]. However, since synaptotagmin does not bind the Habc-domain of syntaxin [158], the present observation that the Habc domain prevents granule docking is inconsistent with direct binding between synaptotagmin and syntaxin initiating the docking process [37]. In conclusion, these data show that incoming granules induce de novo for- mation of syntaxin/munc18 clusters, which is a prerequisite for stable docking.

Organization of exocytosis machinery during granule priming (Paper II and III).

The assembly of the exocytosis machinery requires recruitment of SNAREs and many other proteins to the release site. We followed the recruitment of some of the key proteins, including munc13 and SNAP25, to the release site.

The bulk of munc13 and SNAP25 were not recruited to the release site until about 40 s after docking. The calcium channel Cav1.2 was recruited to a subset of docked granules on a similar timescale. The timing of recruitment of munc13, SNAP25 and Cav1.2 in our experiments is consistent with a role in priming rather than docking. Munc13 binding to the plasma membrane during priming [91,96,159] is dependent on PIP₂ and Ca²⁺[160] and directly binds syntaxin and munc18 [96,161]. Rab3 interacting protein RIM1 is known to bind munc13 [159] and SNAP25 [24] and this interaction might be responsible for driving munc13 and SNAP25 to the release site in a Ca²⁺

dependent manner [162]. Consistent with this idea, rab3a co-localized with granules before they docked resulting in rab3a being recruited with the gran-

Fig 5. Interaction via syntaxin's Habc domain recruits clusters to the docking site.

A) A granule approaching the plasma membrane tethers and forms a stable docking complex upon syntaxin clustering. B) Addi- tion of a soluble N-terminal Habc fragment of syntaxin competes with native syntaxin. The granule can dock as long as sufficient wildtype syntaxin is available to the incoming granule. C) Increased expression of N-terminal Habc fragment competitively binds to incoming granules and hence prevents granule docking.

(26)

ule to plasma membrane during granule docking. RIM1 is also known to bind Cav1.2 channels [24,27,163,164,165]. Such binding might provide a structural framework for channel recruitment and further interaction with SNARE proteins syntaxin 1A and SNAP25 [28,34,166]. The late recruitment of these proteins during granule priming explains the slow rate of organization of the exocytosis machinery. The time taken for these proteins to be recruited to the granules in part explains why not all docked granules are fusion competent.

Disintegration of the exocytosis machinery during membrane fusion (Paper II and III).

Although it is clear that the fusion competent granules have the exocytosis machinery in place [167,168,169,170], it remains unclear how big this fraction is and what happens to the proteins after fusion. We could detect clusters of syntaxin, munc18, munc13, rab3a and SNAP25 at a subset of granules which underwent fusion. Clusters of these proteins were found to co- localize to some extent with non-responding granules too, but the accumula- tion was significantly higher at the responding granules. Within seconds after exocytosis the complex disintegrated with the exception of SNAP25, which remained clustered even 2 s after exocytosis. 20% of the syntaxin also remained at the granule site 2 s after exocytosis. This might reflect the pres- ence of some SNARE complexes remaining intact due to delayed recruitment of recycling factors, such as NSF, which is required to release the SNARE complex in to its constituents [171].

The exocytosis machinery binds to Cav1.2 channels via the synprint site of the channel [30,172]. The Cav1.2 channels conduct the majority of the Ca²⁺

that enters the beta-cell during electrical activity. They are localized to a subset of docked granules. Granules undergoing fusion were associated with stronger EGFP-CaV1.2 signals than those that failed. Most of the exocytosing granules were released within 5 s resulting in a high frequency exocytosis after the application of the depolarization stimulus. Cav1.2 clusters at these granules disintegrated within 4 s after exocytosis. When the granule association with Cav1.2 channels was blocked by co-expression of synprint fragment the number of responding granules markedly declined.

There was a low frequency of exocytosis when cells were stimulated with Ach to release intracellular Ca²⁺. During depolarization-dependent Ca²⁺- influx, a subset of RRP is faster and well synchronized with the depolarization stimulus and called as the immediately releasable pool (IRP). Our finding that Ca²⁺-channel-associated granules undergo rapid, synchronized exocytosis suggests strongly that these granules correspond to the IRP which is observed in capacitance measurements [17]. The IRP granules are release

(27)

ready granules with the exocytosis machinery in place allowing immediate fusion upon Ca²⁺ influx. The exocytosis machinery disintegrates within a few seconds after exocytosis and cannot be reused by newly incoming granules (fig 6).

The release probability increases with the proximity of a granule to Ca

²⁺

influx sites (Paper III).

Higher density of Cav1.2 channels does not necessarily result in increased Ca²⁺-influx at these sites since all the channels are not open. To test for localized Ca²⁺-influx, we visualized Ca²⁺-influx sites using a low affinity Ca²⁺

indicator Fluo5FF in dispersed human islet cells and the genetically encoded Ca²⁺ sensor R-GECO targeted to plasma membrane in INS1 cells. Spatio- temporal averaging of the Fluo5FF and R-GECO images following a depolarization stimulus revealed fast and localized Ca²⁺-influx at responding granules, while the signal was slower and more diffuse at non-responding granules. When the cells were instead stimulated with a short pulse of ace- tylcholine (ACh, 250 µM), which causes release of Ca²⁺ from intracellular stores, the GECO signal was not localized to the granule site. Likewise, expression of the synprint fragment derived from CaV1.2 prevented granule- localized Ca²⁺-influx. Mathematical modeling of Ca²⁺-influx confirmed that the observed rates of Ca²⁺-entry and distribution are consistent with localized Ca²⁺-influx near granules. A spatiotemporal mathematical model with time- to-event statistical analysis [173] allowed us to calculate the release probability. The results demonstrated that if a Ca²⁺ channel moves closer to a granule and reduces the distance between them by 50%, the release probability per time unit increases ~4 fold. This finding is consistent with the responding granules showing a higher density of closely associated Cav1.2 channels.

Kinetics of exocytosis and release (Paper IV).

A fusion pore forms when the vesicle and plasma membranes are melting together during exocytosis, and this process is thought to require conformational changes in the SNARE complex [133,134]. To investigate how the SNARE proteins relate to the dynamics of fusion, we resorted to dual color imaging with a fluorescently tagged SNARE protein and a pH sensitive fluorescent protein to label granules. There are few pH sensitive red fluorescent proteins available and many of them have not been tested for their suitability as secretory granule markers. The present survey identified NPY tagged to tandem dimer (td)-mOrange2 as a suitable pH-sensitive red fluorescent pro-

(28)

tein for dual-color applications to study the dynamics of exocytosis, pore formation and vesicle cargo release. This protein was well-targeted to granules, exhibited bright fluorescence, which bleached slowly, and it showed pH-sensitivity similar to EGFP. Moreover, it allowed detection of single exocytosis events, which were observed as a transient rise of fluorescence caused by dequencing, followed by rapid decline as the marker was released from the vesicle. Experiments with simultaneous imaging of SNARE protein syntaxin-EGFP and td-mOrange2 showed that, the loss of syntaxin coincided with the rising phase of td-mOrange2 intensity, suggesting that syntaxin dispersal is initiated already during pore formation rather than after release of the granule content. This new approach will extend the possibilities to study the kinetics of exocytosis under various conditions.

Fig 6. Model of granule docking and release.

A Approaching granule induces clustering of syntaxin and munc18. The site then matures by recruiting additional proteins, including munc13, and disintegrates during exocytosis. B Undocking as consequence of syntaxin/munc18 cluster dispersal.

(29)

Conclusions

Although there are more than 10,000 granules in a beta cell, few of these are readily available for secretion. Granules which eventually release first have to dock at the plasma membrane and then go through a priming process. A fraction of the docked granules makes up the RRP, which is quickly depleted during the 1^st phase of insulin secretion. Depletion of the RRP results in a slower rate of exocytosis. Replenishment of the RRP takes tens of minutes.

The slow recruitment of granules and subsequent conversion of granules into the RRP during priming becomes a rate limiting factor for sustained secre- tion. Granule docking is associated with de novo formation of syntax- in/munc18 clusters at the plasma membrane. We found no evidence for preexisting granule docking receptors. Granules recruit munc13, SNAP25 and Cav1.2 after docking and the timing of recruitment of these proteins is consistent with a role in priming. Release of primed granules is accelerated due to opening of the calcium channels visualized by Ca²⁺ nanodomains at responding granules, rather than its high density after recruitment. These nanodomains are absent when the interaction between the channel and the release site is blocked. A similar effect can be observed due to decreased expression of SNARE proteins in islets from diabetic patients [174]. This should directly affect docking and could account for secretory defects in patients with type-2 diabetes. The docking step is therefore an attractive target for drug development.

To study the kinetics of exocytosis we tested a panel of ten orange/red and two green FPs in fusions with neuropeptide Y (NPY) for use as secretory granule marker. NPY tagged to td-mOranage2 was well-targeted to granules, bright, bleached slowly and showed pH-sensitivity which is best suited for study of kinetics during exocytosis. The use of this as granule marker in dual color applications with SNARE protein syntaxin fused to EGFP lead us to discover that the syntaxin dispersal is initiated already during pore formation rather than after release. Therefore this vesicle marker is likely suitable also for the study of many other membrane fusion or fission reactions, such as endocytosis and constitutive exocytosis.

(30)

Acknowledgements

A lot people have helped me immensely during my Phd. This is an opportunity to specially express my gratitude for what you have done.

I would like to express my sincere gratitude to my mentor and supervisor, Sebastian Barg for his valuable guidance and support. I only knew I wanted to do science when I joined you but over the years your constant encouragement has made me develop a passion for it. This was also made possible because of your vision, never give up attitude, constant inputs and belief in me. Your patience with me and healthy criticisms helped me grow and develop to become independent. This made us explore a lot of possibilities and solve some challenging projects. I had fun and learnt a lot from this experience. I am very thankful that you sponsored me to attend several big scientific meetings which have given me new ideas. This thesis would not have been possible without your timely inputs for honing my writing skills. Apart from the science I have really enjoyed sharing office with you and having many friendly talks. Thanks to Rebecca for all the publicity we got when our papers came out! I am very happy that I got this opportunity; a big thanks to you for everything.

Thanks to my co-supervisor Anders Tengholm for fruitful collaborations and useful suggestions. Thanks for guiding me throughout with your inputs and encouragement. Your critical comments made this thesis better.

Thanks to Erik Gylfe and Nils Welsh for all the facilities in the department.

Special thanks to Erik Gylfe for sharing your ideas and discussing about my work during the group meetings.

Thanks to Olof who has been the go to person since I started here. I enjoyed all the scientific and non-scientific talk with you. I hope we get to do more science together. Thanks for comments on my thesis.

Thanks to all the people who have collaborated in the projects. Jan for all the molecular biology and constructs which has made many projects possi- ble. You have been a great technician to work with. Per-Eric for valuable guidance on the thesis and for sharing the vast knowledge regarding patch clamp. Peng for being a nice officemate, projectmate and help in the cell lab.

Omar who has brought new life in to the lipid project. You have been very

(31)

helpful with all the islet preparations and big thanks for that. Emma for or- ganizing all the stuff in the cell lab within and practical talk. I hope to get more time on the project that we embarked on. Maria and Kim for taking the granule marker project forward and give it a nice end. Hope the new projects takes shape soon too. Swati for nice times in and out of the office. It was great to get in touch with you in Amsterdam. Good luck to Misty on the new projects, hope we get a breakthrough soon. Ida for embarking on Epac project together and nice times in the department. Anne for wonderful col- laborations and nice time together both in and out of the department. Erik Renström, Vini and all collaborators in Lund for nice work together.

Morten and Michela for all the input on modeling and fruitful collaboration.

Patrik Rorsman and his lab for timely availability of constructs and nice collaborations. Bryndis Birnir and her group for all the help with confocal and some nice scientific discussions and to all previous members and project students in the lab.

Thanks to Bo Hellman for your suggestions and constant encouragement, Eva Grapengiesser for organizing nice Friday fika. Thanks to Ingela for nice suggestions. Thanks to my histology teacher Jia for helping me with the islets, tips about the thesis and nice times together. Thanks to Yuinjian for all the suggestions in making constructs and nice lunches. Thanks to Par- ham, Chenxiao, Antje, Tian, Qian, Hongyan and Vishal for nice times in and out of the lab. Thanks to Oleg, Helene, Parvin, Mariane and Ing- marie for all the technical help and suggestions. Thanks to Daniel, Sara and Monica for constant supply of islets.

Special thanks to Kailash for sharing teaching and being very caring and dependable person. Hope we get to work together at some point of time.

Thanks to all the teaching colleagues Björn, Sara, Marie, Azazul, Qian, Ye, Xuan, Karin, Rikard, Ulrika, Gustaf and Gu. Thanks to Azazul for useful tips about the thesis. Thanks to Ernie, Levon, Hjalti, Jing, Johan, Hannes, Peter for nice times during fika. Thanks to Camilla and Schumin for handling all the administrative work in time. Thanks to all the senior researchers Gunilla Westermark, Michael Welsh, PO Carlsson, Ulf Eriksson, Mia Phillipson, Mats Hjortberg and Stellan Sandler for all the useful suggestions. Thanks to Lina for handling the course credits and stuff.

Thanks to everyone in the department who have made it a nice place to work.

I am very thankful to my small family in Uppsala which include: Maanvi, Chirag, Rashmi, Sarosh for all the care and always being there whenever required. Times spent with you guys were special. Our trip to Norway with you guys, Uncle and aunty was a memorable one. Chetan for great support, sharing interests and insights on scientific and non-scientific stuff and being

(32)

there both in tough and fun times. Seema and Kailash who have been great friends and a reliable pair during tough times. You both have been a very dependable people. Naveen for always being there since I was in Uppsala, now with Divya joining you it is even nicer. Harisha for being there with some valuable tips, adventure and great food. It is nice to see you and Shruthi together. Megha and Ani for being a great pair to spend great time with games, photography and everything. We have to do those trips soon, may be now is the time! Aarav, Navya and Sandeep for all the care and nice food and times hope we get the license soon! Aadrish and Sonchita for nice times. I am happy to be part of the new collaboration with you. Thanks to Amol, Priya, Adi, Smitha, Nimesh, Anubha, Naren, Sujatha, Raghuveer, Satwik, Yash and Megha B for all the great times. Kali for final suggestions on the thesis formatting. Jayanth, Vinutha, Harsha, Shreya, Swaroop, Pragna and Samartha for all the nice get togethers in Stockholm. Madhu Uncle and aunty for providing me accommodation when I was in real need. Thanks to all my friends in Uppsala, Stockholm, Malmö and Umeå.

Special thanks to all my dear friends in Mysore which include Hemanth, Manasa, Chaitra, Aparna, Roopa, Rashmi, Swapna, Prashanth, Vibha and Bhatta. Thanks to boss (sunil sir) who inspired to find a Ph.D in Swe- den. Thanks to all the friends back in India who has been in one way or the other with me all the time.

Thanks to all my teachers including CDS, MKK, Gururaj, HNY and PRN sir; Devaki and Gayathri madam who have encouraged me through my formative years.

Special thanks to my dear sister Shruthi who has been supportive all the time. Time spent during your visit to Europe with Vishwas was fantastic.

Thanks to all my cousins who have been very supportive through all these years. Thanks to all the relatives who have been a great source of encouragement over the years.

Thanks to my family in Kothegala for all the encouragement. My sincere gratitude goes to my parents without whom it would have been impossible to be what I am today. Special thanks to my brother Neeraj for being a great source of encouragement.

Most of all thanks to my dear wife Lakshmi for the understanding, encour- agement, patience, care……and always being my better half…

Special thanks to SSSD for recognizing this work for Young Investigator Award and all other funding agencies which has been supporting this work.

(33)

References

1. Barg S (2003) Mechanisms of exocytosis in insulin-secreting B-cells and glucagon-secreting A-cells. Pharmacol Toxicol 92: 3-13.

2. Hutton JC, Rhodes CJ (1991) The enzymology of proinsulin conversion. Cell Biophys 19: 57-62.

3. Ashcroft FM, Rorsman P (1989) Electrophysiology of the pancreatic beta-cell.

Prog Biophys Mol Biol 54: 87-143.

4. Cerasi E, Luft R (1967) "What is inherited--what is added" hypothesis for the pathogenesis of diabetes mellitus. Diabetes 16: 615-627.

5. Tengholm A, Gylfe E (2009) Oscillatory control of insulin secretion. Mol Cell Endocrinol 297: 58-72.

6. Straub SG, Sharp GW (2002) Glucose-stimulated signaling pathways in biphasic insulin secretion. Diabetes Metab Res Rev 18: 451-463.

7. Ohara-Imaizumi M, Fujiwara T, Nakamichi Y, Okamura T, Akimoto Y, et al.

(2007) Imaging analysis reveals mechanistic differences between first- and second-phase insulin exocytosis. J Cell Biol 177: 695-705.

8. Barg S, Olofsson CS, Schriever-Abeln J, Wendt A, Gebre-Medhin S, et al. (2002) Delay between fusion pore opening and peptide release from large dense-core vesicles in neuroendocrine cells. Neuron 33: 287-299.

9. Barg S, Huang P, Eliasson L, Nelson DJ, Obermuller S, et al. (2001) Priming of insulin granules for exocytosis by granular Cl(-) uptake and acidification. J Cell Sci 114: 2145-2154.

10. Barg S, Eliasson L, Renstrom E, Rorsman P (2002) A subset of 50 secretory granules in close contact with L-type Ca2+ channels accounts for first-phase insulin secretion in mouse beta-cells. Diabetes 51 Suppl 1: S74-82.

11. Braun M, Ramracheya R, Johnson PR, Rorsman P (2009) Exocytotic properties of human pancreatic beta-cells. Ann N Y Acad Sci 1152: 187-193.

12. Ammala C, Eliasson L, Bokvist K, Larsson O, Ashcroft FM, et al. (1993) Exocytosis elicited by action potentials and voltage-clamp calcium currents in individual mouse pancreatic B-cells. J Physiol 472: 665-688.

13. Bokvist K, Eliasson L, Ammala C, Renstrom E, Rorsman P (1995) Co- localization of L-type Ca2+ channels and insulin-containing secretory granules and its significance for the initiation of exocytosis in mouse pancreatic B-cells. EMBO J 14: 50-57.

14. Dryselius S, Grapengiesser E, Hellman B, Gylfe E (1999) Voltage-dependent entry and generation of slow Ca2+ oscillations in glucose-stimulated pancreatic beta-cells. Am J Physiol 276: E512-518.

15. Parsons TD, Coorssen JR, Horstmann H, Almers W (1995) Docked granules, the exocytic burst, and the need for ATP hydrolysis in endocrine cells. Neuron 15: 1085-1096.

16. Eliasson L, Renstrom E, Ding WG, Proks P, Rorsman P (1997) Rapid ATP- dependent priming of secretory granules precedes Ca(2+)-induced exocytosis in mouse pancreatic B-cells. J Physiol 503 ( Pt 2): 399-412.

(34)

17. Barg S, Ma X, Eliasson L, Galvanovskis J, Gopel SO, et al. (2001) Fast exocytosis with few Ca(2+) channels in insulin-secreting mouse pancreatic B cells. Biophys J 81: 3308-3323.

18. Ishihara H, Maechler P, Gjinovci A, Herrera PL, Wollheim CB (2003) Islet beta- cell secretion determines glucagon release from neighbouring alpha-cells. Nat Cell Biol 5: 330-335.

19. Neher E (1998) Vesicle pools and Ca2+ microdomains: new tools for understanding their roles in neurotransmitter release. Neuron 20: 389-399.

20. Nitert MD, Nagorny CL, Wendt A, Eliasson L, Mulder H (2008) CaV1.2 rather than CaV1.3 is coupled to glucose-stimulated insulin secretion in INS-1 832/13 cells. J Mol Endocrinol 41: 1-11.

21. Braun M, Ramracheya R, Bengtsson M, Zhang Q, Karanauskaite J, et al. (2008) Voltage-gated ion channels in human pancreatic beta-cells:

electrophysiological characterization and role in insulin secretion. Diabetes 57: 1618-1628.

22. Iwashima Y, Pugh W, Depaoli AM, Takeda J, Seino S, et al. (1993) Expression of calcium channel mRNAs in rat pancreatic islets and downregulation after glucose infusion. Diabetes 42: 948-955.

23. Reinbothe TM, Alkayyali S, Ahlqvist E, Tuomi T, Isomaa B, et al. (2013) The human L-type calcium channel Cav1.3 regulates insulin release and polymorphisms in CACNA1D associate with type 2 diabetes. Diabetologia 56: 340-349.

24. Coppola T, Magnin-Luthi S, Perret-Menoud V, Gattesco S, Schiavo G, et al.

(2001) Direct interaction of the Rab3 effector RIM with Ca2+ channels, SNAP-25, and synaptotagmin. J Biol Chem 276: 32756-32762.

25. Han Y, Kaeser PS, Sudhof TC, Schneggenburger R (2011) RIM determines Ca(2)+ channel density and vesicle docking at the presynaptic active zone.

Neuron 69: 304-316.

26. Kaeser PS, Deng L, Wang Y, Dulubova I, Liu X, et al. (2011) RIM proteins tether Ca2+ channels to presynaptic active zones via a direct PDZ-domain interaction. Cell 144: 282-295.

27. Kiyonaka S, Wakamori M, Miki T, Uriu Y, Nonaka M, et al. (2007) RIM1 confers sustained activity and neurotransmitter vesicle anchoring to presynaptic Ca2+ channels. Nat Neurosci 10: 691-701.

28. Sheng ZH, Rettig J, Takahashi M, Catterall WA (1994) Identification of a syntaxin-binding site on N-type calcium channels. Neuron 13: 1303-1313.

29. Takahashi T, Momiyama A (1993) Different types of calcium channels mediate central synaptic transmission. Nature 366: 156-158.

30. Wiser O, Bennett MK, Atlas D (1996) Functional interaction of syntaxin and SNAP-25 with voltage-sensitive L- and N-type Ca2+ channels. EMBO J 15:

4100-4110.

31. Mochida S, Sheng ZH, Baker C, Kobayashi H, Catterall WA (1996) Inhibition of neurotransmission by peptides containing the synaptic protein interaction site of N-type Ca2+ channels. Neuron 17: 781-788.

32. Rettig J, Heinemann C, Ashery U, Sheng ZH, Yokoyama CT, et al. (1997) Alteration of Ca2+ dependence of neurotransmitter release by disruption of Ca2+ channel/syntaxin interaction. J Neurosci 17: 6647-6656.

33. Szabo Z, Obermair GJ, Cooper CB, Zamponi GW, Flucher BE (2006) Role of the synprint site in presynaptic targeting of the calcium channel CaV2.2 in hippocampal neurons. Eur J Neurosci 24: 709-718.

Molecular mechanisms of biphasic insulin secretion