The Role of RNA Binding Proteins in Insulin Messenger Stability and Translation

Full text

(1)

(2)

(3)

(4)

(5)

(6) .

(7)

(8)

(9)

(10) ! "#$ %".

(11)

(12) . &&'( )*&+)*** ,, ,,-&.('.+.

(13) .

(14)

(15) ! "

(16) #$%#% # &#$'

(17) (

(18) '

(19)

(20)

(21) ' )(

(22)

(23) (*+(

(24) ,

(25) - (* ! ./*%#%*+(.

(26)

(27) '.01 )

(28) 2 3 +

(29) *1 . *

(30)

(31) .

(32)

(33)

(34) $45*6* *2304784#$$6788$8* 1 (

(35) ( (

(36) '

(37) ''

(38) '

(39) (

(40) ' (

(41) '

(42) ( '

(43) ( (

(44)

(45) ' ( * +( .01 (( ( .01

(46)

(47)

(48) ' .01

(49) * 3 ( ( ( (

(50)

(51)

(52) '

(53)

(54) (,

(55) 9 ( (. ( '

(56)

(57) ' ( (

(58)

(59) ' .01

(60) (

(61)

(62)

(63) ' .01 9

(64) , :

(65) '

(66)

(67) (* +

(68) ' (

(69) ,

(70) .01

(71) , (

(72)

(73) ,

(74)

(75) ;)+<

(76) .01*2 , '

(77) ( )+.01

(78) .01*2 ,

(79) '

(80) ( )+

(81) ($=

(82)

(83) ' ( .01 ( .01 (

(84) ( . ( (

(85) '> (

(86)

(87) (

(88) , ()+*! ( . ,

(89) ( (

(90) ( (

(91)

(92)

(93)

(94)

(95) ' (

(96) .01.# *+(

(97)

(98)

(99) ')+

(100) ( ( *! ,,

(101) ' ,

(102)

(103) ( .0) +21. ( '

(104) ( .01

(105) (

(106) ,. ( (

(107)

(108) ' .01

(109)

(110) '' (

(111) * 2

(112)

(113)

(114)

(115)

(116)

(117)

(118)

(119) ( ,(( (

(120) ' '

(121)

(122) ( . (

(123)

(124) *1.

(125) ' (

(126)

(127)

(128) ('

(129)

(130)

(131) ' * !

(132) . "

(133)

(134) !"

(135) #$%&! !'%$&() !* ?.9 /! %#% 2330#5$#5%5 2304784#$$6788$8 & &&& # % 6;(. &@@ *9*@

(136) A B & &&& # % 6<.

(137) To my family. "The best way to become acquainted with a subject is to write a book about it." Benjamin Disraeli (1804-1881).

(138)

(139) List of Papers. This thesis is based on the following papers, which are referred to in the text by their Roman numerals. I. Fred, R.G. and Welsh, N. (2005) Increased expression of polypyrimidine tract binding protein results in higher insulin mRNA levels. Biochem. Biophys. Res. Commun. 328(1):38-42. II. Fred, R.G., Bang-Berthelsen, C.H., Mandrup-Poulsen, T.M., Grunnet, L.G. and Welsh, N. (2010) High glucose suppresses human islet insulin biosynthesis by inducing miR-133a leading to decreased polypyrimidine tract binding protein expression. PLoS ONE 5:e10843. III. Fred, R.G., Sandberg, M. and Welsh, N. Cap-dependent and cap-independent insulin biosynthesis in human pancreatic islets. Manuscript. IV. Fred, R.G., Islam, T., Adams, C. and Welsh, N. Identification of insulin mRNA binding proteins; implications for insulin biosynthesis and messenger stability. Manuscript. Reprints were made with permission from the respective publishers..

(140)

(141) Contents. Introduction ................................................................................................... 11 Background ................................................................................................... 12 General gene expression........................................................................... 12 Insulin gene expression ............................................................................ 13 Messenger stability and translation .......................................................... 14 Translational regulation ....................................................................... 14 Processing bodies and stress granules ................................................. 15 MicroRNAs and gene regulation ......................................................... 15 Control of insulin mRNA levels and translation ...................................... 16 Regulation of insulin mRNA stability ................................................. 17 PTB - The polypyrimidine tract binding protein ................................. 17 Alternative Translation ............................................................................. 18 Hippuristanol ....................................................................................... 19 Nitric oxide .......................................................................................... 19 Islets of Langerhans ................................................................................. 19 Aims .............................................................................................................. 21 Methodology ................................................................................................. 22 Cell cultures.............................................................................................. 22 Human islets of Langerhans ..................................................................... 22 Downregulation of protein levels using siRNA ....................................... 22 Modification of miR-133a levels ............................................................. 23 Preparation of cytoplasmic extract ........................................................... 23 Total RNA isolation and cDNA synthesis ............................................... 23 Semiquantitative real-time PCR ............................................................... 23 Insulin and total protein biosynthesis ....................................................... 24 RNA-oligonucleotide affinity binding ..................................................... 24 SDS-PAGE and immunoblotting ............................................................. 24 Identification of RNA binding proteins.................................................... 25 1-D PAGE and Digestion .................................................................... 25 Nanoflow LC/MS/MS ......................................................................... 25 Peptide and modification identification ............................................... 26 Results and discussion .................................................................................. 27 Paper I ...................................................................................................... 27.

(142) Paper II ..................................................................................................... 28 Paper III .................................................................................................... 30 Paper IV ................................................................................................... 32 Conclusions ................................................................................................... 34 Paper I ...................................................................................................... 34 Paper II ..................................................................................................... 34 Paper III .................................................................................................... 34 Paper IV ................................................................................................... 34 Acknowledgements ....................................................................................... 35 References ..................................................................................................... 36.

(143) Abbreviations. ADP ATP cAMP DPM ECL eIF ER ERK FCS GAPDH hnRNP IBMX IL-1 IRAK1 IRES ITAF KRBH miRNA mRNA mRNP mTOR NES NLS NO nPTB P-body PAGE PBS PCR PKA Poly(A). Adenosine diphosphate Adenosine triphosphate 3',5-cyclic adenosine monophosphate Disintegrations per minute Enhanced chemiluminescence Eukaryotic initiation factor Endoplasmic reticulum Extracellular signal-regulated kinase Fetal calf serum Glyceraldehyde-3-phosphate dehydrogenase Heterogeneous nuclear ribonucleoprotein 3-isobutyl-1-methylxanthine Interleukin-1 Interleukin-1 receptor-associated kinase 1 Internal ribosome entry site IRES trans acting factor Krebs Ringer Bicarbonate HEPES buffer microRNA Messenger RNA mRNA-ribonucleoprotein Mammalian target of rapamycin Nuclear export sequence Nuclear localization sequence Nitric Oxide Neuronal PTB Prosessing body Polyacrylamide gel electrophoresis Phosphate buffered saline solution Polymerase chain reaction Protein kinase A Polyadenylated.

(144) PTB RBP RISC RNA RNAi RNP RRM SDS siRNA SR SRP TIAR TRAF6 UPR UTR. Polypyrimidine Tract Binding protein RNA binding protein RNA induced silencing complex ribonucleic acid RNA interference ribonucleoprotein RNA recognition motif Sodium dodecyl sulphate Small interfering RNA SRP receptor Signal recognition particle TIA-1-related protein TNF receptor associated factor 6 Unfolded protein response Untranslated region (of mRNA).

(145) Introduction. Diabetes mellitus is a diverse group of metabolic disorders with the common consequence of hyperglycemia through defective insulin secretion or defective insulin action. These metabolic disorders are commonly divided into type 1 and type 2 diabetes, where both types are characterized by a progressive beta cell failure. In type 1 diabetes the insulin producing beta cells within the islets of Langerhans are destroyed, rendering the patient incapable of insulin production. Type 2 diabetes is more variable and the onset is influenced by genetic predisposition, age, weight and physical activity. A patient suffering from type 2 diabetes experiences an increased insulin resistance followed by different degrees of beta cell failure [1, 2]. Although the reason for insufficient insulin secretion may vary depending on the type and stage of the disease, an amplified insulin biosynthesis is an absolute prerequisite to meet the demand during extended periods of insulin hypersecretion [3]. It is therefore probable that disruption of the insulin biosynthesis could contribute to the insulin deficiency observed in diabetes and warrant further investigations of the mechanisms that control insulin biosynthesis. Insulin biosynthesis is dependent on several important mechanisms including proper transcription of the insulin gene, generating sufficient levels of primary transcript [4], correct splicing and transport to the cytosol of the insulin mRNA [5] and the ability of the endoplasmic reticulum both to synthesize insulin and to keep the protein load in equilibrium with the unfolded protein response [6]. Glucose is the main stimulator of insulin biosynthesis [7-9] and induces both increased translation of the insulin mRNA [10-12] and increased levels of insulin mRNA available for translation [13]. Thus, both insulin gene expression and translational control are required to meet the long-term need for insulin biosynthesis. Due to the high copy number of the insulin messenger in beta cells, rates of transcription is of lesser importance compared to messenger stability and translation during short periods of insulin biosynthesis [14]. Since the mechanisms behind the control of insulin messenger stability and translation are still largely obscure, the work in the present thesis aimed to investigate the potential function of insulin mRNA binding proteins and how they modulate both insulin mRNA stability and translation.. 11.

(146) Background. General gene expression All gene expression starts with DNA being copied into RNA, the process known as transcription. Initiation of transcription relies on the specific recognition of cis-regulatory regions in the DNA, by transcription factors. These general and/or specific transcription factors will then recruit the transcriptional machinery, including RNA polymerase II that will synthesize a pre-messenger RNA. The following elongation of the transcript is closely coupled to, and in parts regulates, processing of the pre-messenger RNA. The processing includes capping, splicing and polyadenylation of the premessenger RNA before transcription is terminated and the mature messenger RNA is released. All of these posttranscriptional steps require RNA binding proteins (RBPs) in the context of ribonucleoprotein (RNP) complexes and thus what is released from the site of transcription is not merely a mature mRNA but rather a messenger RNA-ribonucleoprotein (mRNP) complex. After production of the mature mRNA the mRNP complex is exported to the cytosol. Here, the specific and highly dynamic composition of RBPs binding to the mRNA will target it for a number of different outcomes including translation, degradation or localization. If the mRNA is targeted for translation, assembly of the 80S ribosome will initiate translation and protein synthesis. [15-17] Most secretory proteins, such as insulin, are synthesized with a hydrophobic N-terminal signal sequence. When this signal sequence emerges from the ribosome it will induce binding of the signal recognition particle (SRP) that will induce a halt in translation and target the ribosome to the endoplasmic reticulum (ER). At the ER the SRP receptor (SR) facilitates the transfer of the peptide to the translocation pore and translation resumes with the protein being synthesized directly into the ER lumen through the translocation pore. [18]. 12.

(147) Insulin gene expression Insulin gene transcription is controlled by a 340 base pair promoter region, located upstream of the insulin gene. This promoter region contains several cis-elements for binding of both general and beta cell specific transcription factors. In comparison with humans and most other species rodents have two insulin genes, which are equally expressed [19]. Although there are major differences between the rodent and human insulin promoter regions it seems that the regulatory elements for glucose induced gene expression are conserved between species [4]. There are multiple factors that affect insulin gene expression including glucose, fatty acids, hormones and certain amino acids, of which glucose is the most important. Although later steps in glucose signaling may diverge, the early steps are due to changes in intracellular status due to increased glucose metabolism. These changes include increased ratio of ATP/ADP, increased redox state, increased levels of malonyl CoA, increased phospholipid metabolism and changes in intracellular pH. The increase in ATP/ADP ratio also leads to increased levels of intracellular calcium through the closure of ATP-dependent potassium channels and subsequent activation of voltage dependent calcium channels [20]. Downstream of these events glucose promotes the binding of the beta cell specific transcription factors Pdx-1, MafA and Beta2/NeuroD1 to the insulin promoter. These transcription factors collectively activate insulin gene transcription [21]. The mechanisms behind the glucose induced insulin gene transcription by these transcription factors have been partially discovered. In the case of Pdx-1 glucose regulates its subcellular localization and interactions with coregulators. The regulation seems to be phosphorylation dependent and a number of signaling pathways and kinases have been implicated in the phosphorylation of Pdx-1, including p38/SAPK, PI3K, MAPK and PASK. NeuroD1 requires both phosphorylation by ERK and glycosylation to translocate into the nucleus where it heterodimerizes with E47, which is also phosphorylated by ERK, before binding to the insulin promoter. NeuroD1 also interact with co-regulators such as p300. Compared with Pdx-1 and NeuroD1, the transcription of MafA itself is upregulated by glucose. As the transcription of MafA increases it enters the nucleus and binds to the insulin promoter activating insulin gene transcription. [21] In addition to its transcriptional regulation, glucose also affect the splicing of the insulin pre-mRNA [22, 23] and, as mentioned in the introduction, increases the insulin mRNA stability [10].. 13.

(148) Messenger stability and translation The regulation of mRNA in mammalian cells depends in principal upon two factors, transcription and degradation of mRNA. The variations in mRNA decay rate i.e. stability of the messenger can therefore contribute to gene induction or gene repression. This regulation of mRNA stability is controlled by interactions between regulatory cis sites in the mRNA and RNA binding proteins. Changes in half-life can occur in response to many different developmental and environmental stimuli, for example cytokines, hormones and nutrients. In addition to this, environmental stress such as hypoxia or hyperglycemia can also affect the mRNA stability [24]. When talking about the half-life of mRNA it is important to notice that this is not necessarily the same as mRNA stabilty. Due to the fact that mRNA degradation is not a random event occuring at a fixed rate, opposit to the decay of radioactive isotopes, the time it takes for half the mRNA to decay may be an underestimate of the average life span of a mRNA [25]. The mRNA molecule can be degraded by different pathways usually coupled to the different components of the mRNA molecule. The two most important components are the 5'-UTR cap and the 3'-UTR polyadenylated (poly(A)) tail, which protect the mRNA from exonucleolytic decay. The 5'cap and the poly(A) tail stabilize the mRNA by recruiting proteins to these components and thereby protecting them from exonucleases [26]. Another possible mean of degradation is endonucleolytic decay, which cleaves the mRNA enabeling degradation of mRNAs with intact 5'- and 3'-ends [24, 27]. Recent discoveries have shown that mRNAs that are being translated are spatially separated from those that are being degraded. It seems that mRNAs engaged in translation are protected from degradation but as translation is repressed the mRNAs are located to cytoplasmic particles, known as processing bodies and stress granules, either for temporary translational arrest or degradation [28, 29].. Translational regulation The initiation of translation is in principle the process of the 80S ribosome forming on the mature mRNA. Once the 80S ribosome is in place, translation starts with the base pairing of the Met-tRNAMet with the initiation codon and continue with translational elongation i.e. protein biosynthesis. In principle that is, in practice this process requires at least nine different eukaryotic initiation factors, the forming of ribosomal subunits and preinitiation complexes and in most cases scanning of the 5’-UTR, a process that in turn requires the relaxation of the secondary structure. Regulation of translation is important because it allows for rapid changes of protein levels. The step of regulation is usually the initiation of translation, which can be either unspecific, affecting almost all translation or specific, affecting certain mRNAs. 14.

(149) depending on sequence specificity. All regulation affecting either the eukaryotic initiation factors or the ribosomes will fall under the first category and affect virtually all translation in the cell. An example of this can be seen in beta cells. When there is an accumulation of unfolded protein in the endoplasmic reticulum the unfolded protein response will both lead to increase BiP expression and, if necessary, phosphorylation of eIF2 alpha thereby inhibiting translation [30]. Regulation by sequence specific RNA binding proteins on the other hand will only affect mRNAs containing the correct motif in the correct location. This type of regulation is almost always inhibitory, so that an initiation or increase in translation will require degradation or inhibition of the repressor proteins [17, 31]. In addition to repression of translation there is a specific transit of repressed or non-translating mRNAs from the polysomes to the processing bodies were the mRNA molecules can either be stored or degraded.. Processing bodies and stress granules Processing bodies, also known as P-bodies are naturally occurring cytoplasmic foci containing non-translating mRNAs as well as proteins involved in both translational repression and mRNA decay. The number of P-bodies increases in response to stress or when the number of mRNAs not located to the polysomes increase. A second cytoplasmic foci similar to the P-bodies are stress granules. Stress granules are formed in response to environmental stresses such as heat shock, oxidative stress and UV irradiation. These mRNPs share several features and some protein components [32, 33]. New advances in the field of stress induced translational repression shows that in response to stress, mRNAs bound to the ER does not aggregate into stress granules like their cytoplasmic counterparts. The mechanism behind this protective function seems to be that ribosomes do not disengage from ER associated transcripts in response to stress [34]. Another recent discovery shows that both P-bodies and stress granules contain RNA-induced silencing complexes (RISC) suggesting that these RNA granules are coupled to the microRNA-induced translational silencing pathways as well [35].. MicroRNAs and gene regulation MicroRNAs (miRNAs) are endogenous, short (about 23 nt), non-coding RNAs that by pairing to the mRNAs of protein-coding genes can play a direct role in their posttranscriptional regulation. The number of certified miRNA genes in humans are about 400 so far but together with the proposed candidates there might be as many as a thousand [36, 37]. MicroRNAs are transcribed as long primary transcripts that may give rise to several different miRNAs. In humans this pri-miRNA is first processed by the Rnase Drosha resulting in a shorter stem-loop structure that is then exported to the cytop-. 15.

(150) lasm. This precursor miRNA (pre-miRNA) is then further processed by Dicer, a second RNase that is also involved in the maturation of siRNA, to produce the mature miRNA. The mature miRNA is subsequently loaded on the Argonaute protein and incorporated into the silencing complex. Similar to protein coding genes, genes that codes for miRNA are under strict regulation and can be both transcriptional and posttranscriptional regulated [38] . The mechanism of repression can occur both via repression of the mRNA translation and by an accelerated rate of mRNA degradation. In some cases both these events occur simultaneously. The mechanism for miRNA mediated translational repression is unclear with several suggested pathways while the miRNA-mediated degradation occurs through the normal deadenylation-dependent pathway [39, 40]. The target site for the miRNA is usually in the 3’-UTR possibly due to the fact that the miRNA and associated proteins might be dislocated by the translational machinery if present in the 5’UTR or within the open reading frame [41]. In specific cases it has also been reported that the interaction with miRNA have increased the rate of mRNA translation. In this case proteins associated with RISC have bound to the 3’UTR and stimulated translation [42].. Control of insulin mRNA levels and translation Insulin mRNA is highly abundant in beta cells ranging from 20,000, at a low glucose concentration, to 100,000 molecules per cell, at a high glucose concentration [14], representing approximately 30% of the total mRNA pool in the beta cell. This high level of insulin mRNA in beta cells is mainly achieved by a low degradation rate. It has been shown that the half-life of insulin mRNA is between 29 hours, at low glucose, and 77 hours at high glucose [13]. As mentioned in the introduction the high copy number of insulin mRNA together with its very long half-life results in the fact that de novo transcription of the insulin gene contributes very little to the pool of pre-existing insulin mRNA molecules in a short-term context and thus it is likely that insulin mRNA stability mediates the increase observed during the first 24 hours after glucose stimulation [43]. As well as stabilizing the insulin mRNA glucose also increases its translation rate [12]. The regulation of translation of insulin mRNA in response to glucose is a combination of both a global and specific upregulation. The global upregulation of translation in response to glucose in beta cells is due to several factors including increased activity of general translation initiation factors, upregulated translational elogongation and by inducing the release of SRP from transcripts targeted for the ER [9, 12, 44]. The specific upregulation of insulin mRNA translation by glucose has been shown to be mediated by interaction of the two untranslated regions [45]. It has recently been pro-. 16.

(151) posed that a short sequence in the 5’ UTR mediates this glucose responsive translational regulation [46]. It deserves to be mentioned that glucose also has been reported to induce a selective translocation of insulin mRNA to the ER [47] and that this might be connected to the increased initiation rate of insulin mRNA translation [44]. It could also be speculated that the increased association with the ER in itself leads to increased stability of the insulin mRNA.. Regulation of insulin mRNA stability The stability of the insulin mRNA, as well as other mRNAs, is more then likely regulated by several different trans-acting factors interacting with cisacting elements within the UTRs of the insulin mRNA. The co-operation or competition of these factors will then be responsible for the regulation of both the constitutive as well as the glucose induced insulin mRNA stability. Previous research on this topic has shown that it is primarily the 3'-UTR that direct the stability of the messenger [45]. At first it was believed that this was due to a conserved UUGAA sequence located in the 3'-UTR [45] but later research has shown that there is a pyrimidine rich cis-acting sequence further upstream that is able to increase the messenger stability via the interaction with the trans-acting factor polypyrimidine tract binding protein (PTB) [14]. Interestingly, it has also been shown that PTB by binding to this sequence of other mRNAs also stabilizes mRNAs coding for insulin granule proteins (ICA512, PC1/3 and PC2), indicating that there is a unifying mechanism for the simultaneous and coordinated up-regulation of the biosynthesis of insulin and other granule proteins [48]. It has since then been shown that PTB preferentially associate with transcripts from several gene ontology groups, including intracellular protein transport, vesicle-mediated transport, Golgi apparatus and Golgi trans face [49]. Taken together this suggest that PTB is one of the trans-acting factor that regulate insulin mRNA stability and that it might stabilize mRNAs of proteins involved in the entire secretory pathway.. PTB - The polypyrimidine tract binding protein PTB is a RNA binding protein that was first identified as the heterogeneous nuclear ribonucleoprotein I (hnRNP I) [50, 51], and has since then been implicated in RNA splicing [52], cap-independent translation [53], RNA polyadenylation [54], RNA localization [55], mRNA stabilization [14] and focal adhesion [56]. There are two closely related genes encoding for PTB located on different chromosomes. The first one is PTBP1, located at 19p13.3, and the second is PTBP2, located at 1p22.1. The second gene is primarily expressed in neurons producing a protein known as neuronal PTB (nPTB) [57]. The primary transcript of PTBP1 produces four isoforms of. 17.

(152) PTB through alternative splicing. These isoforms are named PTBP1aPTBP1d and are also known as PTB1, PTB2, PTB3, PTB4 and PTB-T [5861]. The RNA binding capacity of PTB arises from four different RNA recognition motifs (RRMs) [62], which all contribute to RNA binding [63]. The amount of PTB in the cell is controlled in part by an auto-regulatory feed back loop. Overexpression of PTB promotes an alternative splicing that targets the transcript for degradation via the nonsense-mediated decay pathway [64]. Although PTB can be found both in the nucleus and the cytoplasm it is under normal conditions predominately nuclear. Supporting nuclecytoplasmic shuttling PTB contains both a nuclear localization sequence (NLS) [62] and a nuclear export signal (NES) [65]. The import of PTB seems to follow the classic nuclear import pathway by interaction between the NLS and importin-D [66]. While the export mechanism is still somewhat unclear it has been shown that it is energy dependent and not affected by transcription. One factor that has been shown to facilitate the export of PTB is phosphorylation by protein kinase A (PKA) [67]. This have since been shown in insulin producing cells as well, where PTB phosphorylation by PKA was mediated by increased cAMP levels [48, 68]. As mentioned previous glucose signaling is secondary i.e. there are no receptors for glucose. Instead the initial signaling is due to changes in intracellular status in response to increased glucose metabolism. The interaction between PTB and the insulin mRNA can be further stimulated with a high glucose concentration or hypoxia [14, 43]. Recent investigations showed that the glucose induced binding of PTB to the insulin mRNA was inhibited by rapamycin, resulting in reduced levels of insulin mRNA [69]. Rapamycin inhibits a protein known as the mammalian target of rapamycin (mTOR), which has previously been linked to both insulin gene transcription [70] and insulin biosynthesis [71]. It is therefore plausible that mTOR is located downstream of the initial glucose signaling, functioning as both a nutrient sensor and regulator of beta cell function.. Alternative Translation The most common form of translation requires the interaction of the eukaryotic initiation factor complex (eIF-4F) with the 7-methyl-G cap (5’-cap) for binding of the 40S ribosomal subunit to the mRNA molecule [72]. The alternative to this is referred to as cap-independent translation or internal ribosome entry site (IRES) translation. This process does not require the eIF4F complex but rather a structural element in the 5’-UTR of the mRNA for the direct binding of the 40S ribosome subunit to the initiation codon. Some 10-15% of all mRNAs are thought to be capable of being translated via capindependent mechanisms. This pathway ensures protein synthesis of critical proteins during conditions of cellular stress, such as hypoxia, cytokine-. 18.

(153) induced nitric oxide production and apoptosis [73]. This alternative translation is facilitated by proteins known as IRES trans-acting factors (ITAFs). The binding of ITAFs to the mRNA causes structural changes that enable the 40S ribosome to be recruited [74]. Interestingly, one of the ITAFs that not only facilitates the IRES translation of eukaryotic mRNA, but also has been suggested to function as a universal ITAF, is PTB [75, 76] Cap-independent translation can is commonly measured using a dicistronic pRF vector [77]. In addition to be difficult to use in primary tissue there are also doubts regarding the strength of the IRES-mediated translation using these vectors [78, 79]. An alternative approach is to measure protein synthesis rates in the presence of an inhibitor of eIF-4F. One such inhibitor is hippuristanol, which was recently shown to specifically inhibit cap-dependent translation [80].. Hippuristanol The common gorgonian Isis Hippuris, a soft coral that grows in the western Pacific Ocean, produces a cytotoxic polygenated steroid, presently known as hippuristanol. This steroid has the ability of inhibiting the RNA helicase eIF4A, a subunit of eIF-4F [81]. Experiments have shown that through this activity hippuristanol is able to specifically inhibit cap-dependent translation without affecting cap-independent translation [82]. This action is much appreciated since most inhibitors of protein synthesis targets ribosomes, which makes them difficult to use for specific inhibition of cap-dependent translation [80].. Nitric oxide Nitric oxide (NO) is a small molecule and although it is fairly unreactive towards biomolecules it has a broad range of direct and indirect biological effects, including changing the oxidative environment of the cell, inducing endoplasmic reticulum (ER) stress, activating the unfolded protein response (UPR) and promoting apoptosis [83-85]. One biological effect suggested for NO is the capability to induce proteolysis of eIF4G, also a subunit of eIF-4F, thus stimulating a shift from cap-dependent to cap-independent protein biosynthesis. Another biological effect that might seem to be in contrast with this is the reported ability to induce proteolysis of the 60S ribosome and eIF2D, down regulating total protein biosynthesis [86, 87].. Islets of Langerhans The islets of Langerhans are clusters of endocrine cells within the otherwise exocrine pancreas. They constitute approximately 1-2% of the mammalian. 19.

(154) pancreas and are spherical in the range of 25-300 Pm in diameter. These clusters, or islets, are miniature organs consisting of not only of endocrine cells but also endothelial cells, fibroblasts and macrophages/dendritic cells. The endocrine cells are divided in five different cell types that together regulate metabolism. These cell types are: glucagon producing alpha cells, somatostain producing delta cells, pancreatic polypetide producing PP cells, ghrelin producing epsilon cells and of course the insulin producing beta cells [88]. Although there are vital differences between rodent and human islets in islet structure [89], cell composition [90, 91] and sensitivity to cytokines and stress [92] there is far more known about, for example, insulin biosynthesis in rodent islets [93-95] than there is about human islets [96, 97]. This is in part due to the scarce availability of human islets but also the lack of cell lines of human origin, leaving most of the in vitro research in this field to be performed on rodent islets and/or rodent cell lines.. 20.

(155) Aims. The general aims of this thesis were to further explore the mechanisms involved in the control of insulin messenger stability and to investigate how insulin mRNA binding proteins, with emphasis on PTB, affect insulin mRNA stability and translation. The specific aims for the different papers were:. I. To clarify the correlation between PTB gene expression and insulin mRNA levels and investigate if increased PTB levels can affect long-term insulin production.. II. To investigate whether pathogenic factors in type 2 diabetes increase the expression of PTB-targeting microRNAs, and to explore the putative influence of such microRNAs on insulin biosynthesis rates.. III. To study whether PTB binds to the 5'-UTR of insulin mRNA and to investigate whether insulin mRNA can be translated by a cap-independent mechanism in human islets.. IV. To identify insulin mRNA binding proteins and elucidate their role in insulin biosynthesis and insulin mRNA stability.. 21.

(156) Methodology. Cell cultures The mouse insulinoma ETC-6 cell line [98] was cultured in DMEM supplemented 10 % FCS, 2mM L-glutamine and antibiotics. The rat insulinoma cell line INS-1 832/13 [99] was cultured in RPMI 1640 supplemented with 10 % fetal calf serum (FCS), 2mM L-glutamine, 10mM HEPES, 1mM sodium pyruvate and 50 PM E-mercaptoethanol and antibiotics. All cells were passaged approximately twice a week and kept at 37qC in a humified air incubator with 5 % CO2.. Human islets of Langerhans Isolated human islets were provided by Professor Olle Korsgren (the Department of Radiology, Oncology and Clinical Immunology at Uppsala University Hospital, Uppsala, Sweden) through the Uppsala facility for the isolation of human islets from Scandinavian brain-dead individuals using collagenase digestion and Biocoll gradient centrifugation [100]. The islets were precultured for 1-5 days in CMRL 1066 medium containing 5.6 mM glucose, 10 % FCS, and 2 mM L-glutamine and antibiotics. To evaluate islet beta cell content and quality, islets were routinely stained with Newport green followed by fluorescence microscopy and batches of islets were perfused to determine the response to glucose. The beta cell percentages were routinely 30-60 % and for 11 batches of human islets used in this study the mean stimulation index in response to glucose was 7.8±2.4.. Downregulation of protein levels using siRNA In paper I we used a pre-designed siRNA oligonucleotide directed against PTB. This siRNA was introduced into the cells using Lipofectamine¥. Due to the unexpected upregulation of PTB that was observed a second, custom designed, siRNA oligonucleotide against PTB was acquired. The second siRNA oligonucleotide was introduced into the cells using Dharmafect I and, as predicted, downregulated PTB (paper IV). All transfections were. 22.

(157) performed using non-targeting siRNA as negative control and were performed according to the manufacturers recommendations.. Modification of miR-133a levels To study the effects of miR-133a in human islets we used a pre-designed miR-133a precursor and a pre-designed miR-133a inhibitor, these are both small, double-stranded and chemically modified RNA molecules. The microRNA constructs, designed to respectively mimic or cancel out the effects of miR-133a, together with a negative control microRNA oligonucleotide were transfected into human islets using Dharmafect I. To improve transfection efficiency, human islets were dispersed using trypsin prior to transfection. The resulting islets cell suspension was then transfected according to the manufacture’s recommendations.. Preparation of cytoplasmic extract After incubation the cells or islets were washed in cold PBS and resuspended in solution A, a 10 mM HEPES buffer (pH 7.9) supplemented with 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT and protein/phosphatase inhibitor cocktail. The samples were incubated on ice for 10 min and following a brief centrifugation and again resuspended in solution A before they were lysed with an electric homogenizer. The lysates were then centrifuged for 5 min at 800 x g at 4C after which the supernatants were collected. At this point 10 % of the cytoplasmic extract was saved for total lysate control.. Total RNA isolation and cDNA synthesis Cells or islets were washed in PBS and total RNA was obtained using the Ultraspec¥ RNA Isolation system according to the instructions of the manufacturer. The RNA samples were either used for cDNA synthesis using MMulV reverse transcriptase Rnase H (papers I and IV) or TaqMan Reverse Transcription Reagents (paper II). For microRNA isolation (paper II), RNA precipitation was prolonged at -20 qC and the TaqMan microRNA Reverse Transcription kit was used for cDNA synthesis.. Semiquantitative real-time PCR Real time PCR for mRNA was performed using SYBR Green Taq Readymix (papers I and IV) or TaqMan Gene Expression Assays (paper II) and the. 23.

(158) TaqMan microRNA Assay was used for microRNA (paper II). The genes of interest were normalized against PPIA, beta actin, GAPDH or let7c (an internal control for microRNA).. Insulin and total protein biosynthesis To measure protein synthesis in human islets and INS-1 cells they were incubated for 2 and 3 hours respectively in Krebs-Ringer bicarbonate buffer supplemented with HEPES (KRBH), bovine serum albumin (BSA) and radioactive leucine (L-[3,4,5-3H] leucine). The longer incubation time for the INS-1 cells was due to the cell line's lower rate of insulin biosynthesis. After incubation, and incorporation of radioactive leucine, the islets or cells were washed, harvested and homogenized by sonication. For measurement of insulin biosynthesis by immunoprecipitation guinea pig anti-human insulin serum or normal guinea pig serum (control) was added to an aliquot of the samples. After incubation the antibodies together with any bound protein was precipitated using protein A-Sepharose. Total protein was precipitated by the addition of trichloroacetic acid (TCA), which induces the precipitation of macromolecules. To determine the radioactivity i.e. disintegrations per minute (DPM), the samples were solubilized in Ultima Gold scintillation fluid and analyzed with liquid scintillation spectrometry (Wallac System 1409) using external standardization.. RNA-oligonucleotide affinity binding The affinity of RNA binding proteins was investigated using biotinylated oligonucleotides corresponding to different regions within the rat and human insulin mRNA UTRs. The biotinylated oligonucleotides were bound to streptavidin coated Dynabeads£ and resuspended in cytosolic extracts, from INS-1 cells or human islets, to allow RNA-protein complexes to form. For control, a scrambled non-specific oligonucleotide was used. After incubation the beads were washed and resuspended in 1x SDS-sample buffer. To elute the proteins the beads were boiled for 5 min in the SDS-sample buffer.. SDS-PAGE and immunoblotting Protein samples in SDS-sample buffer were separated on SDS-PAGE. Proteins were electrophoretically transferred to Hybond-P filters (Amersham Biosciences), which were then blocked for one hour using a 5 % BSA protein solution. The filters were probed with primary antibodies for PTB,. 24.

(159) hnRNP L, hnRNP U, TIAR and ERK. The horseradish peroxidase conjugated anti-mouse, anti-rabbit or anti-goat antibodies were used as secondary antibodies and the immunodetection was performed as described for the ECL immunoblotting detection system by Amersham Biosciences, using the Kodak Image station 4000MM. The resulting bands were quantified using Kodak Digital Science ID software.. Identification of RNA binding proteins 1-D PAGE and Digestion Eluted proteins were separated using SDS-PAGE. The gels were fixed in 40 % (v/v) ethanol and 10 % (v/v) acetic acid after which the protein bands were visualized using colloidal Coomassie staining. The protein bands specific to the sample lane (i.e. not present in control lane) were excised and samples were in-gel reduced, alkylated, and digested with modified sequence-grade trypsin as previously described [101]. Samples were vacuumcentrifuged and reconstituted in HPLC water prior to analysis.. Nanoflow LC/MS/MS All experiments were performed on a 7-tesla hybrid linear ion trap Fourier transform mass spectrometer using a modified nanoelectro-spray ion source. The HPLC setup consisted of a solvent degasser, nanoflow pump, and thermostated microautosampler. A 15-cm fused silica emitter was used as analytical column. The emitter was packed in-house with a methanol slurry of reverse-phase, fully end-capped Reprosil-Pur C18-AQ 3-μm resin using a pressurized "packing bomb" operated at 50–60 bars. Mobile phases consisted of 0.5% acetic acid and 99.5% water (v/v) (buffer A) and 0.5% acetic acid and 10% water in 89.5% acetonitrile (v/v) (buffer B). 8 μl of prepared peptide mixture was automatically loaded onto the column over 20 minutes at 5% buffer B, flow rate of 500 nl/min followed by a 90-min gradient from 5 to 45% buffer B at a constant flow rate of 200 nl/min. MS analysis was performed in data-dependent acquisition mode in which the mass spectrometer automatically switches between a high resolution survey scan (resolution = 100,000; m/z range, 200–1600) followed by lower resolution fragmentation spectra (electron capture dissociation [102, 103] followed by collisionactivated dissociation resolution = 25,000) of the two most abundant peptides eluting at any given time.. 25.

(160) Peptide and modification identification Acquired RAW files were converted to dta files using Extract_msn through BioWorks Browser, and complementary pairs were identified as described previously described [104]. Base peptides were identified by searching against the NCBI (National Center for Biotechnology Information) database using the Mascot search engine. Searches were performed permitting 2 trypsin miscleavages and the mass tolerance for monoisotopic peptide identification was set to 10 ppm and ±0.02 Da for fragment ions. The instrument setting was "ESI-FTICR". The peptide mass tolerance, mass accuracy window for fragment ions, and enzyme specificity as well as the instrument settings were kept unchanged. Parsing of data and statistical analysis of the search results reported by Mascot were performed using the open-source software MSQuant [105].. 26.

(161) Results and discussion. Paper I To clarify the correlation between PTB and insulin mRNA levels we stimulated E-TC-6 cells with glucose and IL-1E, two substances known to affect insulin mRNA levels. Exposure to glucose increases insulin mRNA levels rapidly, in one study already after a two hour incubation [10]. Previous studies have indicted a role for PTB in this short-term stabilization of the insulin mRNA [14, 48, 67]. To explore the possibility of a long-term role for PTB we decided to study the effects after 24 hours of incubation with glucose. The cytokine IL-1E has previously been shown to increase insulin mRNA levels after a 24 hour incubation [106]. The mechanism for this has not been clarified and therefore we decided to study the effects after a 24 hour incubation with IL-1E as well. The results showed that stimulation with both glucose and IL-1E increased PTB mRNA levels. This was paralleled by a three to four fold increase in insulin mRNA levels. These findings support the view that PTB increases insulin mRNA stability [14, 48] and indicate a long-term role for PTB. The observation that also IL-1E increased PTB and insulin mRNA levels suggests that the stimulatory role of IL-1Eon beta cell insulin production [106]in cells not inhibited by iNOS expression, involves, at least in part, an enhanced PTB gene expression. Since insulin mRNA levels are mainly controlled by messenger stability over a 24 hour period it is to be expected that the present results reflect changes in insulin mRNA stability, and not changes in transcription. To confirm that the increase in PTB resulted in increased insulin levels we attempted to downregulate PTB by means of siRNA. Surprisingly, the siRNA treatment had the opposite effect on the PTB mRNA levels. Western blot experiments showed that siRNA mediated increase in PTB mRNA levels also enhanced PTB protein levels. Although this experiment does not explain the result it excludes any artificial errors originating from the PCR measurements. Since the unexpected increase in PTB protein levels was paralleled by an increase in insulin mRNA levels this experiment supports the role for PTB in control of insulin mRNA stability. With both a short-term and a long-term mechanism (both through glucose and IL-1E) for PTB in insulin mRNA stability control it becomes increasing-. 27.

(162) ly motivated to find possible mediators of the PTB effects. It has recently been shown that phosphorylation of PTB on ser-16 and the subsequent relocalization from the nucleus to the cytoplasm [48, 67] could promote PTBmediated insulin mRNA stabilization. There are several reports on possible kinases for such a mechanism [107, 108]. Although we have not been able to consistently reproduce the glucose- and cAMP-induced effect on PTB ser-16 phosphorylation and PTB-subcellular localization (results not shown) it is possible that the affinity of cytoplasmic PTB for insulin mRNA is regulated by interaction with several other proteins involved in the control of insulin mRNA stability.. Paper II Previous studies have demonstrated that binding of PTB to the 3'-UTR of the insulin mRNA [14] and other mRNAs important for insulin granule formation [48] stabilizes these mRNAs leading to improved insulin biosynthesis. Recent discoveries in the field of microRNAs have implied that expression of certain microRNAs might contribute to pathogenesis of diabetes [109]. We therefore wanted to investigate whether factors known to impair beta cell function might do so by upregulation of microRNAs that target PTB. We chose to investigate the effect of high glucose, palmitate and cytokines on miR-133a, miR-124a and miR-146 levels. miR-133a is one microRNA that has been implicated in the pathophysiology of diabetes. In type 2 diabetes patients the normal downregulation of miR-133a in skeletal muscle by insulin is reduced leading to increased miR133a levels [110], and cardiomyocytes from type 2 diabetes patients also show increased miR-133a levels [111]. The consequences of this upregulation and its possible link to insulin resistance and heart failure is unknown but may be due to effects on the uncoupling protein (UCP) -2, a miR-133a target [112]. PTB is another possible target of miR-133a since PTB mRNA contains a possible miR-133a binding site and because miR-133a has been shown to downregulate nPTB [113], a paralog of PTB. Our results showed that a prolonged exposure to high glucose led to a significant increase in miR-133a levels. Neither cytokines nor palmitate significantly altered miR-133a levels although there was a tendency toward increased miR-133a levels in response to palmitate. This response was however not consistent between different batches (donors) of human islets. Consistent with the notion that microRNAs are able to downregulate protein levels without affecting mRNA levels the following experiments showed that although PTB mRNA levels were unaffected in response to high glucose, PTB protein levels were significantly lower. After establishing that high glucose downregulates PTB expression we went on to measure insulin mRNA levels in response these treatments. The results showed that there. 28.

(163) was no significant change in insulin mRNA levels with either treatment. This was somewhat surprising since we expected the lowered PTB expression to result in a corresponding downregulation of insulin mRNA. Measurements of the insulin biosynthesis in response to high glucose resulted nevertheless in a significantly decreased insulin biosynthesis. The inconsistency between insulin mRNA and biosynthesis data could be due to an upregulation of insulin gene transcription that obscured any destabilization of the insulin mRNA. Thus, it can be speculated that PTB is important not only for the stability of the insulin transcript but also involved more directly in its translation. This could be explained by the fact that not all mRNA in the cytosol is available for translation but can instead be stored in translantionally inactive pools such as stress granules or processing bodies and that binding of PTB to the insulin mRNA is necessary for translation to occur [114]. A consequence of this might be that PTB protein levels, and not insulin mRNA levels, are rate limiting for insulin biosynthesis. To verify the effect of miR-133a on PTB levels and insulin biosynthesis we also showed that a miR-133a inhibitor counteracted the effects of high glucose on both PTB levels and insulin biosynthesis and that the effects of high glucose was mimicked by a miR-133a precursor under normal conditions. Taken together these data demonstrate that miR-133a, by downregulating PTB expression is in part responsible for the decrease in insulin biosynthesis in response to high glucose. In addition to this it remains to be evaluated whether miR-133a also targets UCP-2 in beta cells, thereby increasing ROS production and contributing to the miR-133a induced glucotoxicity. The second microRNA that we measured was miR-124a. This microRNA has previously been shown to target PTB in neuronal cells [115] and is reported to be upregulated in a mouse beta cell line in response to high glucose [116]. miR-124a may also target rab27A, a secretory vesicle protein, [117] and FoxA2, a transcription factor, thereby leading to alterations in beta cell differentiation, glucose metabolism and insulin release. Over expression of miR-124 has previously been shown to affect calcium levels and glucoseinduced insulin release [117, 118]. Our results did not show any significant changes in response to glucose, palmitate or cytokines. Again this might be due to variations in the material and it is still possible that a glucoseincreased expression of miR-124a might negatively affect islet function. An interesting finding was that miR-146 was significantly decreased in human islet in response to high glucose. miR-146 levels has previously been shown to increase in response to cytokines in human islets [119], which was confirmed by our results as well. This induction of miR-146 by cytokines mediates a feed back inhibition of Toll-like receptor/IL-1 receptor signaling by downregulating the expression of IRAK1 and TRAF6, which are both downstream signaling mediators [119]. The finding that high glucose downregulated miR-146 could help us understand the islet inflammation and beta cell dysfunction seen in type 2 diabetes since hyperglycemia has been sug-. 29.

(164) gested to increase IL-1 production [120]. A possible mechanism for this negative effect of hyperglycemia could therefore be that prolonged exposure to high glucose concentrations increase islet susceptibility to Toll-like/IL-1 receptor activation by counteracting the miR-146-induced downregulation of IRAK1 and TRAF6. Taken together we have found a microRNA-regulated mechanism in human islets that mediates a suppressive effect on insulin biosynthesis. Even though the effect on insulin biosynthesis of the miR-133a precursor was equivalent to that of high glucose it is likely that this is one of several mechanisms that in combination lead to beta cell failure in type 2 diabetes.. Paper III Based on the idea that PTB, if binding to the 5’-UTR of the insulin mRNA, could induce cap-independent translation, we investigated whether PTB binds to the 5’-UTR of the insulin mRNA. Both human and rodent insulin mRNA contain a polypyrimidine rich sequence located in the 5'-UTR in addition to the one previously described in the 3'-UTR [14]. The sequence located in the 5'-UTR is similar to the PTB consensus sequence [45, 76], thus indicating a possible second binding site for PTB. Using RNA oligonucleotides corresponding to either the first 29 or the following 30 nucleotides of the 5’-UTR sequence we performed pull-down experiments that clearly showed that PTB binds specifically to the second part of the insulin mRNA 5’-UTR. This was supported by the fact that a very weak signal was detected with the oligonucleotide corresponding to the first 29 nucleotides that does not contain any polypyrimidine rich sequence. The sequence specificity of the PTB-RNA interaction was further supported by the result that no PTB binding was observed with a control oligonucleotide consisting of a scrambled insulin mRNA 5'-UTR nucleotide sequence. After having established that PTB binds to the 5'-UTR of the insulin mRNA we wanted to measure any possible cap-independent translation. This is commonly measured using a dicistronic pRF vector [77] but there are several drawbacks using these vectors. The results from previous measurements of cap-independent translation using monocistronic and dicistronic vectors have been questioned and there is an on-going debate whether they might give false-positive results [78, 79]. Another problem is that islets of Langerhans are notoriously difficult to transfect even when using viral mediated gene transfer [121]. Instead we decided to treat human islets with hippuristanol, an inhibitor of cap-dependent translation. Hippuristanol has recently been shown to block the activity of eIF4A and thus inhibit cap-dependent protein biosynthesis without affecting cap-independent protein biosynthesis [81]. To induce cap-independent translation we also treated human islets with DETAnonoate, a NO donor, since NO have been shown to stimulate. 30.

(165) cap-independent translation by inducing proteolysis of eIF4G [86, 87]. The experiments were performed both in the presence of high glucose and low glucose to mimic stimulatory as well as non-stimulatory conditions. The results show that hippuristanol preferentially blocked high glucoseinduced insulin biosynthesis, but not so much basal insulin biosynthesis and thus we suggest that insulin biosynthesis occurs at a low rate by capindependent translation. The rates of insulin biosynthesis in the presence of the inhibitor were similar at low and high glucose, indicating that glucosestimulated insulin mRNA translation occurs exclusively via an increase in cap-dependent translation. The effects of hippuristanol on total protein synthesis were somewhat more pronounced than those on insulin biosynthesis at high glucose, resulting in an increased percentage insulin/total protein biosynthesis. Our results also show that DETAnonoate inhibits most of the insulin biosynthesis at high glucose and that the DETAnonoate treated groups are less sensitive to inhibition by hippuristanol. Subsequent pull-down experiments showed that there were no differences in affinity of PTB for the 5'UTR oligonucleotide after incubation with glucose and DETAnonoate suggesting that the cap-independent insulin biosynthesis is constitutive. Taken together our findings indicate that glucose stimulates insulin biosynthesis in human islets by increasing cap-dependent translation, and that nitrosative stress counteracts this effect. In addition, it is possible that PTB by binding to the 5’-UTR of the insulin mRNA ensures basal rates of translation, which might be of importance for the human body to always maintain a certain minimum of insulin biosynthesis. Such a mechanism would hypothetically ensure a basal production of insulin even at conditions of starvation and stress, thereby promoting survival of the beta cell, as insulin has been demonstrated to function as an autocrine beta-cell survival factor [122]. In addition to the discovery of cap-independent insulin biosynthesis the present investigation also confirms previous findings that the percentage insulin of total protein synthesized in response to glucose is considerably lower in human islets as compared to rodent islets [96, 97] and that the biosynthesis of both insulin and total protein is markedly increased in response to glucose, indicating that human islets, as opposed to rodent islets, respond to a high glucose concentration by an equally strong enhancement of both insulin and total protein biosynthesis.. 31.

(166) Paper IV PTB usually exerts its effect on mRNA stability and translation in cooperation with additional proteins [123]. In that context it becomes of great interest to first identify additional proteins that bind to the insulin mRNA and then investigate whether they have any function in the stability and translation of the insulin mRNA. We have consequently investigated this and can now report that in addition to PTB, both hnRNP U and TIAL bind specifically to sequences within the insulin mRNA UTRs, and that their respective affinity for these sequences correlates with changes in insulin mRNA stability and biosynthesis. To identify proteins that may be involved in stability and translation we used a biotinylated construct corresponding to the previously reported sequence that was shown to facilitate insulin mRNA stability [14]. Proteins interacting with this sequence were then eluted and identified by mass spectrometry. At first we identified several candidates including hnRNP K, hnRNP E and Tia-1 but due to several reasons including high background and weak immunoblot signals we continued with hnRNP U, hnRNP L and TIAR as our main candidates. Secondly, we wanted to investigate whether these proteins binds to the 5'-UTR as well as the 3'-UTR of the insulin mRNA. This would be in line with previous work reporting that the glucose induced translation of insulin mRNA is induced by the interaction between the 3'-UTR and the 5'-UTR of the insulin mRNA [45]. For this purpose and to be able to observe differences in binding capacity we designed several constructs. Since a possible interaction between the two UTRs could either subsist due to direct interaction of the two UTRs, i.e. complementary binding forming a double stranded secondary structure, or by interaction of one or more proteins that bind to the insulin mRNA UTRs, we included one group with oligonucleotides from both UTRs. Based on the predicted secondary structure of the insulin mRNA both modes of interactions should be possible by this combination. To begin with our results showed that all four proteins binds to the 5'UTR of the insulin mRNA. The discovery that PTB binds to the 5'-UTR of the insulin mRNA eventually led to the findings presented in Paper III. Although the strength with which PTB and hnRNP U bound to the constructs differed, both TIAR and hnRNP L bound with similar affinity to all constructs. It was later decided that hnRNP L binds unspecifically to our oligos and was therefore subsequently used as internal binding control. Further, the results showed that hnRNP U had less affinity for the short 3' oligo than for the other constructs. The signal was significantly weaker than the unspecific signal in the control, indicating that possible binding sites were occupied by other proteins. Finally PTB showed a high affinity to the 3' oligos and a significantly lower affinity to the 5' oligos. In addition to this, two experiments on human islets indicate that PTB binds with a higher affinity to the 5' se-. 32.

(167) quence in human cells compared to rodent cells. This result is supported by the longer pyrimidine rich sequence found in the human insulin mRNA 5'UTR. This could indicate a difference in the mechanism for insulin messenger stability and translation between humans and rodents. After establishing that hnRNP U and TIAR in addition to PTB bind our oligonucleotides, we wanted to investigate how their respective affinity would change and correspond to cellular processes. The results show that the affinity of PTB and TIAR to the two longer constructs was significantly increased in response to high glucose. This finding both corresponds with previous data for PTB and indicates a role for TIAR in the glucose mediated insulin mRNA stability. In response to DETAnonoate the affinity of both hnRNP U and TIAR were significantly upregulated. Despite previous findings by other groups’, treatment with IBMX did not significantly change the affinity for any of the proteins. The reason for this discrepancy is unclear but might be due to experimental differences. The affinities of the different proteins to our constructs were correlated to the effects on insulin mRNA by measurement of both insulin mRNA levels and insulin biosynthesis in response to the previous treatments. For these experiments a low glucose group was added as control for high glucose. The results show that while high glucose significantly increases both insulin mRNA levels and insulin biosynthesis, treatment with IBMX leads to significant decreases. Since no effect was seen on the affinities of these proteins the effect of IBMX might be through other insulin mRNA binding proteins or unspecific effects. Treatment with DETAnonoate significantly decreased insulin mRNA levels, insulin biosynthesis and total protein biosynthesis. The response to DETAnonate correlates with previous findings showing that DETAnonoate downregulates translation and induces stress in beta cells [92]. To validate the effect of these proteins on insulin mRNA stability, PTB and TIAR were downregulated using siRNA. The results confirmed our previous findings as downregulation of PTB decreased insulin mRNA levels and downregulation of TIAR increased insulin mRNA levels. Further studies in this field are warranted due to the many questions that still remain. The fact that TIAR is involved in the formation of stress granules and thereby can affect both the stability and the translation of mRNA on a global level is one such question [32]. Previous findings that both hnRNP U and TIAL might stabilize mRNAs [124, 125] together with our results suggests that the involved proteins by themselves may have opposite functions and that it is the composition of the RNPs that regulates the stability and translation of insulin mRNA. In summary our findings show that increased insulin biosynthesis is, in part, mediated through RNA binding proteins and that it is the affinity of these proteins that mediate their effect.. 33.

(168) Conclusions. Paper I x x. An increase in PTB mRNA and protein levels is paralleled by an increase in insulin mRNA levels. The expression of the PTB gene participates in the regulation of insulin production.. Paper II x x x. Prolonged exposure to high glucose increases miR-133a levels in human islets. Increased levels of miR-133a downregulates PTB levels and insulin biosynthesis rates. miR-146 levels are decreased in response to high glucose in human islets.. Paper III x x x. Insulin mRNA can be translated at low rates via a capindependent mechanism. Insulin biosynthesis is mainly cap-dependent at a high glucose concentration, but not during conditions of stress. PTB binds to the insulin mRNA 5’-UTR in vitro and might be the mediator of cap-independent insulin biosynthesis.. Paper IV x x x. 34. In addition to PTB both hnRNP U and TIAR bind specifically to the insulin mRNA in vitro. PTB binding to insulin mRNA results in increased stability and translation. During stress the binding of TIAR promotes insulin mRNA degradation..

(169) Acknowledgements. This work was carried out at the Department of Medical Cell Biology, Uppsala University, Sweden. I would like to thank everyone at the department for creating an open and friendly atmosphere. It has always made me feel welcome. Especially, I would like to express my sincere gratitude to the following people: My supervisor, Prof. Nils Welsh, who not only accepted me as a Ph. D. student and introduce me to the field of diabetes but also encourage and guided me through these years. Thank you for everything! My co-supervisors, Prof. Michael Welsh and Prof. Stellan Sandler for scientific discussions, stimulating conversations and challenging questions. The former and present head of the department, Prof. Arne Andersson and Prof. Erik Gylfe, for running a smooth and efficient department. Former and present members of the group, Andreea Barbu, Ole Forsberg, Robert Hägerkvist, Natalia Makeeva, Dariush Mokhtari, Johan Olerud and Xuan Wang for good times in and out of office. My co-writers, for valuable and fruitful collaborations. The technical and administrative staff, for their expertise, help and advise on all things practical. Former and present PhD students at the department. My parents, without whom I would never have made it this far. My brother and sister for always supporting me. My own little family, Syrina and Samuel for your endless love and for reminding me of what is important in life.. 35.

(170) References. 1. 2. 3. 4. 5. 6.. 7. 8. 9. 10. 11. 12. 13.. 36. Tisch, R. and H. McDevitt, Insulin-dependent diabetes mellitus. Cell, 1996. 85(3): p. 291-7. Cnop, M., et al., Mechanisms of pancreatic beta-cell death in type 1 and type 2 diabetes: many differences, few similarities. Diabetes, 2005. 54 Suppl 2: p. S97-107. Wajchenberg, B.L., beta-cell failure in diabetes and preservation by clinical treatment. Endocr Rev, 2007. 28(2): p. 187-218. Hay, C.W. and K. Docherty, Comparative analysis of insulin gene promoters: implications for diabetes research. Diabetes, 2006. 55(12): p. 3201-13. Minn, A.H., et al., Increased insulin translation from an insulin splice-variant overexpressed in diabetes, obesity, and insulin resistance. Mol Endocrinol, 2005. 19(3): p. 794-803. Kaneto, H., et al., Role of oxidative stress, endoplasmic reticulum stress, and c-Jun N-terminal kinase in pancreatic beta-cell dysfunction and insulin resistance. Int J Biochem Cell Biol, 2005. 37(8): p. 1595-608. Permutt, M.A. and D.M. Kipnis, Insulin biosynthesis. I. On the mechanism of glucose stimulation. J Biol Chem, 1972. 247(4): p. 1194-9. Permutt, M.A. and D.M. Kipnis, Insulin biosynthesis. II. Effect of glucose on ribonucleic acid synthesis in isolated rat islets. J Biol Chem, 1972. 247(4): p. 1200-7. Permutt, M.A., Effect of glucose on initiation and elongation rates in isolated rat pancreatic islets. J Biol Chem, 1974. 249(9): p. 273842. Brunstedt, J. and S.J. Chan, Direct effect of glucose on the preproinsulin mRNA level in isolated pancreatic islets. Biochem Biophys Res Commun, 1982. 106(4): p. 1383-9. Itoh, N., et al., Glucose stimulation of the proinsulin synthesis in isolated pancreatic islets without increasing amount of proinsulin mRNA. FEBS Lett, 1978. 93(2): p. 343-7. Itoh, N. and H. Okamoto, Translational control of proinsulin synthesis by glucose. Nature, 1980. 283(5742): p. 100-2. Welsh, M., et al., Control of insulin gene expression in pancreatic beta-cells and in an insulin-producing cell line, RIN-5F cells. II. Regulation of insulin mRNA stability. J Biol Chem, 1985. 260(25): p. 13590-4..

No results found