Comparative Biochemistry and Physiology, Part A 151 (2008) 180–184

Contents lists available at ScienceDirect

Comparative Biochemistry and Physiology, Part A

j ourna l homepage: www.e lsev ie r.com/ locate /cbpa

Drosophila ia2 modulates secretion of insulin-like peptide

Jihyun Kim, Hyojoo Bang, Syungkyun Ko, Inhee Jung, Haehee Hong, Jeongsil Kim-Ha ⁎Department of Molecular Biology, Sejong University, 98 Kunja-dong, Kwangjin ku 143-747, Seoul, South Korea

⁎ Corresponding author. Tel.: +82 2 3408 3644; fax: +E-mail address: jsha@sejong.ac.kr (J. Kim-Ha).

1095-6433/$ – see front matter © 2008 Elsevier Inc. Aldoi:10.1016/j.cbpa.2008.06.020

A B S T R A C T

A R T I C L E I N F O

Article history:

Islet antigen-2 (IA-2) is a m Received 17 March 2008Received in revised form 19 June 2008Accepted 19 June 2008Available online 25 June 2008

Keywords:AutoantigenDiabetesDrosophilaHexokinaseIA-2ia2GutPancreas

ajor autoantigen in type I diabetes. To throw light on the function of IA-2 weexamined the role of ia2, a Drosophila homologue, during Drosophila development. In situ hybridizationshowed that ia2 was expressed in the central nervous system and midgut region. The neuronal expressionpattern of ia2 was very similar to that of IA-2 in mammals. Disruption of gut-specific ia2 expression bydouble stranded RNA interference (dsRNAi) resulted in defects in gut development, and this phenotype wasrescued by overexpression of hexokinase. Until now the roles of IA-2 and hexokinase in insulin signaling havebeen described separately but we found that ia2 modulated the expression of both insulin and hexokinase.Moreover this modulation seems to be important for gut development during metamorphosis. As thepancreas develops from the gut during vertebrate development, our results suggest a possible role of IA-2 ininsulin and hexokinase regulation.

© 2008 Elsevier Inc. All rights reserved.

1. Introduction

Approximately 70% of newly diagnosed Type I diabetic patients haveautoantibodies to IA-2 (Lan et al., 1996; Leslie et al., 1999). IA-2 is amember of the protein tyrosine phosphatase family (Lan et al.,1994) butit is not clear whether IA-2 truly functions as a phosphatase because itfailed to show any phosphatase activity. Endogenous IA-2 is present inthe secretory vesicles of neuroendocrine cells throughout the body,including the alpha and beta cells of pancreatic islets (Solimena et al.,1996). Disruption of IA-2 inmice does not lead to a dramatic increase ofblood glucose level but it reduces glucose-dependent insulin secretion(Saeki et al., 2002; Henquin et al., 2008). Conversely, overexpression ofIA-2 results in increased glucose-induced insulin secretion (Harashimaet al., 2005). These observations point to a role of IA-2 in regulatinginsulin secretion and thus in the development of diabetes. However,knockout of IA-2 in the non-obese diabetic mouse, the most widelystudied animalmodel of type I diabetes, did not lead to the developmentof diabetes (Kubosaki et al., 2004). Therefore, thephysiological role of IA-2 in the pathogenesis of Type I diabetes is still not clear.

Drosophila melanogaster contains probable orthologs of human dis-ease genes that match 75% of the 1378 human disease loci examined(Lasko, 2002). Recent research on the Drosophila insulin-signalingpathway has raised the possibility that certain features of diabetesmellitus are shared by both Drosophila and humans. In the Drosophilagenome, seven insulin-like peptides (ilps) are found that are expressed

82 2 312 8834.

l rights reserved.

in brain and other tissues, such as imaginal discs and gut. The moststudied region that produces the ilps is an area of seven neuroendocrinecells in each brain hemisphere. The expression of ilp-2, -3, and -5was detected in these insulin-producing cells (IPCs); additionally, IPCablation was shown to cause an elevation of carbohydrate levels inhemolymph (Rulifson et al., 2002). In addition to carbohydrateregulation, the roles of the ilps were demonstrated in growth andreproduction (Claeys et al., 2002). However, the specificity of each ilp, aswell as additional roles of the ilps, has not been investigated seriously.

A homolog of the human IA-2 gene, ia2, was found in Drosophila.As the possible use of Drosophila in the study of diabetes has beensuggested from the ilp results, we expected to find another regulatorymechanism from the study of ia2 inDrosophila. Therefore, we examinedthe tissue-specific expression of ia2 and analyzed its function in vivo.

2. Materials and methods

2.1. Tissue preparation and in situ hybridization

Flieswere embedded inOCTcompoundand frozen in liquidnitrogen.Serial 15 µm cryosectionswere placed on polylysine-coated slides, fixedwith4%paraformaldehyde in 1× phosphate-buffered saline (PBS) for 1 h,and washed in PBST (0.1% Triton X-100 in PBS). After incubation inacetylation buffer (0.25% acetic anhydride in 0.1M triethanolamine),theywerewashed inPBST, andprehybridized inhybridizationbuffer (5XSSC, 50% formamide, 100 µg/mL salmon sperm DNA, 50 µg/mL heparin,and 0.1% Tween-20) at room temperature for 10 min. Hybridizationswere performed at 55 °C overnight. After washing with PBST, the slideswere incubated with alkaline phosphatase-conjugated anti-DIG

Fig. 1. Expression of the ia2 gene during development. A. Developmental RT-PCRanalysis of ia2 gene expression. L1, L2, L3 indicates 1st, 2nd and 3rd larvae, respectively,and P1, P2, P3 are one day white, two day yellow and three day-old pupae, respectively.The ribosomal protein gene rp49 was used as a loading control. B. In situ hybridizationof sectioned larvae and adult flies with a Drosophila ia2 riboprobe. The riboprobe wasmade from a DNA template corresponding to nucleotides 3728 to 4257 in GenBankaccession number NM134718. ia2 expression was detected in the brain and midgut(arrows) of first and third instar larvae (top and middle), and expressionwas conservedin adult flies (bottom).

181J. Kim et al. / Comparative Biochemistry and Physiology, Part A 151 (2008) 180–184

antibody (Roche, Mannheim, Germany), and hybridization signals werevisualized by incubation with BCIP and NBT. Incubation was continueduntil thehybridization signalswere sufficiently strong, and the reactionswere stopped by washing in 1× PBS.

2.2. Transgenic flies harboring ia2 dsRNA and a hexokinase-overexpressing construct

To generate a construct expressing a stem-loop structured ia2dsRNA, ia2 cDNA fragments corresponding to nucleotides 3728–4257and 3728–4397 (nucleotide numbers according to GenBank acces-sion number NM134718) were incorporated into the pCaSpeR-UASvector in opposite orientation. A full length 1493 nucleotide Hexoki-nase (Hex) cDNA (GenBank accession number AF237469) was clonedinto pCaSpeR-UAS in the sense orientation. Transgenic flies were gen-erated by injecting embryos with these constructs together with ahelper plasmid, delta 2-3. UAS-InR and UAS-Akt-expressing flies wereobtained from Dr. JK Jeong at KAIST.

2.3. Hexokinase antibody production and western analysis

Polyclonal antibodies against Hex protein corresponding to the Cterminal half (nucleotides 599 to 1339 of GenBank AF237469) wereraised in rats. Affinity purified Hex antibody was tested for specificityby transfecting Schneider cells with full length Hex cDNA cloned intopCaSpeR-hs vector (Thummel et al., 1988) and examining extracts ofthe transfected cells for reactivity with the Hex antibodies. Elevatedamounts of Hex protein were detected in transfected and 2 h heat-induced cells (data not shown). For Western analysis, larval extractswere incubated with anti-Hex and anti-GFP (Santa Cruz, USA)antibodies at 1: 4000 and 1:1000 dilutions, respectively.

2.4. RT-PCR

RNA was extracted from 3rd instar larvae of the wild type and theKr-GAL4NUAS-ia2RNAi strain. From 5 µg of total RNA, single strandcDNAwas synthesized with oligo dT primer and the final volume wasadjusted with D.W to 200 µL at the end of the reaction. Five µL of thecDNA was used for PCR which consisted of 95 °C for 1 min, 55 °C for1 min and 72 °C for 1 min, repeated twenty eight times (Master cyclergradient 5331, Eppendorf Inc. Germany) for all the insulin-like-peptide (Ilp) genes except for Ilp4; 35 cycles were performed for thelatter. Primers used were: Ilp1 (5′ CGGAGCAGGAGGTGCAGGAT 3′/ 5′CTATTTCGGTAGACAGTAGA 3′), Ilp2 (5′ ACCTAAGCAGTAAACCCATA 3′/5′ TCCAGATCGCTGTCGGCACC 3′), Ilp3 (5′ ATGGGCATCGAGATGAGGTG3′/ 5′ GGAACGGTCTTCGAAGCCAT 3′), Ilp4 (5′ ATGAGCCTGATTAG-ACTGGG 3′/ 5′ GGTCTCGCACTCTAGCATCC 3′), Ilp5 (5′ ATGTTCCG-CTCCGTGATCCC 3′/ 5′ GGAGCTATCCAAATCCGCCA 3′), Ilp6 ( 5′ATGGTTCTCAAAGTGCCGAC 3′/ 5′ CCTGCGCTTCCCGAAACTGT 3′). Ilp7(5′ TCGGACTGGGAGAACGTGTG 3′/ 5′ GGAGTGTTTCCATCCGATCG 3′)and rp49 (5′ CAGTCGGATCGATATGCTAAGCTGT 3′/ 5′ TAACCGATG-TTGGGCATCAGATACT 3′). For real-time PCR analysis, 5 µL of thesynthesized cDNA was used per reaction. SYBR green dye (Invitrogen,USA) and EF-Taq polymerase (Solgent, Korea) was used. PCR consistedof 50 °C for 2 min, 95 °C for 1 min followed by 95 °C for 10 s, 55 °Cfor 25 s and 72 °C for 30 s, repeated forty times. The amounts of Ilpsrelative to the rp49 control gene product were measured using theddCt relative quantitation study program in the 7300 Real Time PCR(Applied Biosystems, USA) system SDS software.

3. Results

3.1. Drosophila homologs of the human IA-2 gene

The 979 amino acid sequence of human IA-2 was used to search aDrosophila protein database. Two hypothetical proteins, ia2-PA and

ia2-PB, from a single genomic region (CG31795) were found with ahigh degree of similarity to human IA-2. ia2-PA and ia2-PB are 1144and 1288 amino acids in length, respectively. ia2-PB (GenBank num-ber Q59E11) differs only slightly from ia2-PA: it contains a 163 aminoacid N-terminal extension, and its final C-terminal 12 amino acidsequence is different from that of ia2-PA (GenBank number Q9VPV8).Most of the extracellular N-terminal region of human IA-2 is notconserved in the Drosophila ia2 proteins (Grumbling and Strelets,2006). The greatest similarity is found in the intracellular proteintyrosine phosphatase domain, which shares 65% identity and 77%homology to IA-2. Inwhat follows the two homologues will be treatedas one protein and referred to as Drosophila ia2.

3.2. Expression of Drosophila ia2

In humans and the mouse, IA-2 is exclusively expressed in brainand pancreas. In Drosophila, small clusters of neuroendocrine cells inthe brain called insulin producing cells (IPCs) have been shown tocontribute to pancreas function (Rulifson et al., 2002). No other tissuesthat function as pancreas equivalents have been identified in Droso-phila. To compare the expression pattern of Drosophila ia2with that ofhuman and mouse, we examined ia2 expression in various develop-mental stages. Drosophila ia2 transcripts were detected from the larvalstage and persisted throughout development to the adult (Fig. 1). To

Fig. 3. Hexokinase is reduced when ia2 expression is suppressed. Expression of Hexwasexamined using Hex antibody. A dramatic decrease of Hex protein was observed in Kr-GAL4NUAS-ia2RNAi 3rd instar larvae compared to Kr-GAL4 strains. As the Kr-GAL4chromosome also contains UAS-GFP, GFP protein was used as a loading control.

182 J. Kim et al. / Comparative Biochemistry and Physiology, Part A 151 (2008) 180–184

examine the tissue specific expression of ia2 we sectioned wild typelarvae and adult flies and carried out in situ hybridization with an ia2probe. In larval and adult stages ia2 was exclusively expressed in thecentral nervous system (CNS), which consists of brain and thoracicganglion (Fig. 1B). The neuronal expression pattern of ia2 is verysimilar to that of IA-2 in the mouse, where transcripts are found inbrain, pituitary and pancreas (Shimizu et al., 2001). ia2 was alsoexpressed in midgut cells (Fig. 1B). The Drosophila larval midgutconsists of two types of cells: polytene enterocytes and diploid gutimaginal gland cells (Yee and Hynes,1993). Duringmetamorphosis theformer cells die and the latter expand to form the adult gut. ia2 wasexpressed in both the polytene and diploid cells (data not shown).

3.3. Disruption of ia2 expression in the abdominal region causes a defectin abdomen formation during metamorphosis

Sincewe could not locate any existing ia2mutant flies we inhibitedia2 function by double stranded RNA interference (dsRNAi). We madea construct consisting of a 529 nucleotide inverted repeat sequencefrom the C- terminus of the ia2 coding region, with a UAS sequence inthe promoter, and generated transgenic flies harboring this construct.The dsRNA was expressed specifically in neurons and gut cells, usingtissue specific GAL4 drivers. When an elav-Gal4 driver was used tosuppress ia2 expression in the CNS region we did not detect anyreduction of ia2 mRNA expression (data not shown), and the flies didnot show any defects. On the other hand, Kr-Gal4 driven lines (we willrefer to these as Kr-Gal4NUAS-ia2RNAi lines) showed a tissue-specificreduction of ia2 expression: CNS expression was normal, but gutexpression was significantly reduced (Fig 2A, B). Since ia2 expression

Fig. 2. Tissue specific disruption of ia2. Inwild type adult flies ia2 is expressed in midgutcells (A). In flies expressing ia2 dsRNA under the Krüppel-GAL4 (Kr-GAL4) driverexpression of ia2 was reduced only in tissues where Kr-GAL4 is expressed: expressionin the midgut (arrows) was dramatically reduced (B). Larvae harboring both Kr-GAL4and UAS-dsRNA (ia2) also had a reduced rate of eclosion (E). Eighty percent of theflies of the expected genotype arrested at the 3-day pupa stage and died withouteclosion (C, E). Flies harboring both Kr-GAL4 and UAS-dsRNA (ia2) had darker eyepigment because of the white gene inserted in the P element vector (bottom in C and D)whereas flies with either one of the transposons had lighter eye pigment (top in Cand D). Overexpression of Hex together with the ia2 dsRNAi construct reversed thepupal stage lethal phenotype, as can be seen by the empty pupal case (F).

in the CNS remained normal, we were only able to confirm theefficiency of RNAi in the gut region by in situ hybridization. 80% of theflies did not eclose (Fig. 2C-E); they had eyes and wings but did notcomplete development. To examine the cause of this lethality weremoved the bodies from the pupal cases. Although most of the bodyparts were complete, the abdominal regionwas covered by a very thinfilm-like layer instead of hard cuticle (Fig. 2D).When this thin filmwasremoved almost no abdominal structures could be seen. We concludethat the internal organs of the abdominal region did not developproperly in these flies at metamorphosis.

3.4. Overexpression of hexokinase rescues the abdominal defect causedby dsRNA interference of ia2

Since IA-2 is reported to be involved in the secretion of insulin(Saeki et al., 2002), we tested whether the lethal RNAi phenotypecould be rescued by overexpressing genes encoding proteins of theinsulin signaling pathway. We overexpressed InR, Akt and Hex inmidgut cells using the same Gal4 driver that was used to drive ia2dsRNA expression. Overexpression of InR or Akt did not rescue thelethal phenotype of ia2 dsRNA flies (data not shown) but over-expression of Hex did (Fig. 2 E and F). We therefore asked whetherHexexpression is lower in the dsRNA interference strains. We examinedHex protein levels by Western blot analysis and found a markedreduction in the Hex band when ia2 was knocked down (Fig. 3).

3.5. Insulin-like-peptide expression is reduced by knockdown of ia2

In Drosophila, there are seven insulin-like-peptides (ilps). Weexamined the expression of all seven ilps in wild type and Kr-GAL4NUAS-ia2RNAi larvae. A reduction of mRNA expression was detectedonly in ilp6 (Fig. 4). While no change of expression of any of theother ilps was detected by reverse transcription-polymerase chainreaction (RT-PCR), a reduction of about 40% of the ilp6 PCR productwas observed (Fig. 4A). The same result was obtained by real time PCR(Fig. 4B).

4. Discussion

Human IA-2 and IA-2 beta share 74% identity and are majorautoantigens in type I diabetes. The IA-2 gene is conserved in manyspecies such as macaco, rat, mouse, C. elegans, zebrafish and Droso-phila (Cai et al., 2001). The function of IA-2 has only been tested inmice and an attempt to determine its function in C. elegans usingdsRNA was unsuccessful (Cai et al., 2001).

Drosophila has a homologous ia2 gene located at 21E3 on thesecond chromosome. The corresponding two transcripts from thissingle genomic region generate two protein isoforms (Grumbling and

Fig. 4. Expression of insulin-like peptides in ia2 RNAi strains. RNA was extracted from Kr-GAL4NUAS-ia2RNAi and control Kr-Gal4 third instar larvae and examined for all the seveninsulin-like peptides. A. Reduction of expressionwas only observed for the ilp6 transcript. The amount of PCR product was quantitated using the GelQuant program provided by DNRBio-imaging Systems Ltd (Jerusalem, Israel). B. To confirm the data, expression was re-examined by real-time PCR. The relative quantification method was used and a reduction ofabout 40% of the Ilp6 transcript was observed.

183J. Kim et al. / Comparative Biochemistry and Physiology, Part A 151 (2008) 180–184

Strelets, 2006). Our findings and those of other workers show thatexpression of the IA-2 family has similar tissue specificity: in humanand mouse there is a high level of IA-2 expression in brain andmoderate levels in the pancreas (Shimizu et al., 2001). We also found ahigh level of Drosophila ia2 expression in brain and the thoracicganglion region. Although a test of pancreas-specific expression is notpossible since there is apparently no pancreas in Drosophila, weobserved expression of ia2 in the gut region, which is the organ fromwhich the pancreas develops in vertebrates. Therefore, not only theirsequence conservation but also the conservation of their tissueexpression pattern makes it worthwhile to compare the function ofIA-2 and its homologues in various organisms.

Insulin stimulates the expression of genes that encode glycolyticenzymes, including glucokinase/hexokinase, while inhibiting theexpression of those genes that encode gluconeogenic enzymes (Saltieland Kaha, 2001). The selective stimulation of hexokinase mRNA andprotein synthesis by insulin has been reported (Osawa et al., 1995),and reduced hexokinase activity is observed in diabetes (Vestergaard,1999). We also observed a dramatic decrease in Hex expressionwithinia2-knockdown Drosophila strains. In addition, a reduction in iIp-6expressionwas observed. Since ia2 knockdownwas shown to result inthe down-regulation of both ilp6 and Hex, we next examined theblood sugar levels in these flies. Since the major sugar in Drosophilahemolymph is trehalose, wemeasured hemolymph carbohydrate levelsas the sum of both trehalose and glucose according to the method ofRulifson et al. (2002). Unexpectedly, we observed a decrease in carbo-hydrate levels (data not shown).

During metamorphosis, larval gut cells die and diploid gutimaginal gland cells expand to form the adult gut structure. Theabsence of ia2 does not appear to interfere with the hydrolysis of thelarval cells but does prevent the expansion of new adult-specific gutcells. Certain roles in growth and reproduction have been demon-strated for certain ilps but the role of ilp6 has not yet been defined.Various ilp-expression patterns were examined by Brogiolo et al.(Brogiolo et al., 2001) who found that an ilp6 signal could be detectedin the gut regions of Drosophila larvae. One of the best-documentedeffects of insect ilp genes in Manduca sexta and Bombyx mori was thestimulation of ecdysteroidogenesis (Hayes et al., 1995; Nagata et al.,1992). Ecdysone is a key regulatory molecule in metamorphosis. Thesevere defective phenotype that we observed in abdominal develop-mentmight be caused by an ecdysone deficiency.We observed normalmetamorphosis in both the head and thorax but metamorphosisin the abdomen appeared to be defective. To determine whether

ia2 functions in the gut imaginal gland to either regulate hormonalsignaling or promote cell survival for adult gut tissue formation re-quires further analysis.

Evidence for the inter-regulation of IA-2/ia2, insulin, and hexoki-nase is increasing. An association of glucokinase/hexokinase withinsulin secretory vesicles as a means to regulate glucokinase/hexo-kinase stability by preventing its degradation has been suggested(Saltiel and Kaha, 2001). Therefore, the destabilization of insulinsecretory vesicles by ia2 knockdown could lead to decreases in bothinsulin and hexokinase levels. Although the mechanism for the rescueof the ia2 deficient, abdominal morphogenesis-defective phenotypeby Hex is not clear at this point, it is clear that ia2 and Hex function inthe same pathway during gut development. It would be interesting todetermine whether increased hexokinase activity can also compen-sate for a reduction of IA-2 in the mammalian pancreas.

We show that Drosophila ia2 shares many regulatory features withmammalian IA-2. A further extension of our study would lead to abetter understanding of the regulatory components in the IA-2pathway.

Acknowledgements

This work was supported by grant 02-PJ1-PG3-21002-0001 fromthe Korean Ministry of Health and Welfare, and R01-2006-000-10783from the Korea Science and Engineering Foundation to J. K-H.

References

Brogiolo, W., Stocker, H., Ikeya, T., Rintelen, F., Fernandez, R., Hafen, E., 2001. Anevolutionarily conserved function of the Drosophila insulin receptor and insulin-like peptides in growth control. Curr. Biol. 11, 213–221.

Cai, T., Krause, M.W., Odenwald, W.F., Toyama, R., Notkins, A.L., 2001. The IA-2 family:homolog in Caenorhabditis elegans, Drosophila and zebrafish. Daibetologia 44, 81–88.

Claeys, I., Simonet, G., Poels, J., Loy, T.V., Vercammen, L., De Loof, A., Broeck, J.V., 2002.Insulin-related peptides and their conserved signal transduction pathway. Peptides23, 807–816.

Grumbling, G., Strelets, V., 2006. The FlyBase consortium, FlyBase: anatomical data,images and queries. Nucleic Acids Res. 34, D484–D488.

Harashima, S., Clark, A., Christie, M.R., Notkins, A.L., 2005. The dense core transmem-brane vesicle protein IA-2 is a regulator of vesicle number and insulin secretion.Proc. Natl. Acad. Sci. U. S. A. 102, 8704–8909.

Hayes, G.C., Muehleisen, D.P., Bollenbacher, W.E., Watson, R.D., 1995. Stimulation ofecdysteroidogenesis by small prothoracicotropic hormone: role of calcium. Mol.Cell. Endocrinol. 115, 105–112.

Henquin, J.-C., Nenquin, M., Szollosi, A., Kubosaki, A., Notkins, A.L., 2008. Insulinsecretion in islets from mice with a double knockout for the dense core vesicleproteins islet antigen-2 (IA-2) and IA-2β. J. Endocrinol. 196, 573–581.

184 J. Kim et al. / Comparative Biochemistry and Physiology, Part A 151 (2008) 180–184

Kubosaki, A., Miura, J., Notkins, A.L., 2004. IA-2 is not required for the development ofdiabetes in NOD mice. Diabetologia 47, 149–150.

Lan, M.S., Lu, J., Goto, Y., Notkins, A.L., 1994. Molecular cloning and identification of areceptor-type protein tyrosine phosphatase, IA-2, from human insulinoma. DNA CellBiol. 13, 505–514.

Lan, M.S., Wasserfall, C., Maclaren, N.K., Notkins, A.L., 1996. IA-2, a transmembraneprotein of the protein tyrosine phosphatase family, is a major autoantigenin insulin-dependent diabetes mellitus. Proc. Natl. Acad. Sci. U. S. A. 93, 6367–6370.

Lasko, P., 2002. Diabetic flies? Using Drosophila melanogaster to understand the causesof monogenic and genetically complex diseases. Clin. Genet. 62, 358–367.

Leslie, L.D., Atkinson, M.A., Notkins, A.L., 1999. Autoantigens IA-2 and GAD in Type I(insulin-dependent) diabetes. Diabetologia 42, 3–14.

Nagata, K., Maruyama, K., Nagasawa, H., Urushibata, I., Isogai, A., Ishizaki, H., Suzuki, A.,1992. Bombyxin-II and its disulfide bond isomers: synthesis and activity. Peptides13, 653–662.

Osawa, H., Printz, R.L., Whitesell, R.R., Granner, D.K., 1995. Regulation of hexokinase IIgene transcription and glucose phosphorylation by catecholamines, cyclic AMP, andinsulin. Diabetes 44, 1426–1432.

Rulifson, E.J., Kim, S.K., Nusse, R., 2002. Ablation of insulin-producing neurons in flies:growth and diabetic phenotypes. Science 296, 1118–1120.

Saeki, K., Zhu, M., Kubosaki, A., Xie, J., Lan, M.S., Notkins, A.L., 2002. Targeted disruptionof the protein tyrosine phosphatase like molecule IA-2 results in alterations inglucose tolerance tests and insulin secretion. Diabetes 51, 1842–1850.

Saltiel, A.R., Kaha, C.R., 2001. Insulin signaling and the regulation of glucose and lipidmetabolism. Nature 414, 799–806.

Shimizu, S., Saito, N., Kubosaki, A., Seo, S.W., Takeyama, N., Sakamoto, T., Matsumoto, Y.,Saeki, K., Matsumoto, Y., Onodera, T., 2001. Developmental expression and localizationof IA-2mRNA inmouse neuroendocrine tissues. Biochem. Biophys. Res. Commun. 288,165–171.

Solimena, M., Dirkx Jr., R., Hermel, J.M., Pleasic-Williams, S., Shapiro, J.A., Caron, L.,Rabin, D.U.,1996. ICA512, an autoantigen of typeI diabetes, is an intrinsic membraneprotein of neurosecretory granules. EMBO J. 15, 2102–2114.

Thummel, C.S., Boulet, A.M., Lipshitz, H.D., 1988. Vectors for Drosophila P-element-mediated transformation and tissue culture transfection. Gene 74, 445–456.

Vestergaard, H., 1999. Studies of gene expression and activity of hexokinase, phos-phofructokinase and glycogen synthase in human skeletalmuscle in states of alteredinsulin-stimulated glucose metabolism. Dan. Med. Bull. 46, 13–34.

Yee, G.H., Hynes, R.O., 1993. A novel tissue-specific integrin subunit beta nu, expressedin the midgut of Drosophila melanogaster. Development 118, 845–858.