Molecular Brain Research 63 Ž1998. 155–162
Research report
Alternative splicing of mouse IL-15 is due to the use of an internal splice site in exon 5 Marco Prinz, Uwe-Karsten Hanisch, Helmut Kettenmann, Frank Kirchhoff
)
Max Delbruck 10, 13092 Berlin, Germany ¨ Center for Molecular Medicine (MDC), Cellular Neurosciences, Robert-Rossle-Straße ¨ Accepted 13 October 1998
Abstract IL-15 is a pleiotropic cytokine modulating growth and differentiation of several hematopoietic cell types. Recently, we have demonstrated that mouse microglial cells, the brain macrophages, express both IL-15 and IL-15rIL-2 receptors. Based on single-cell RT-PCR data, we describe here an alternatively spliced IL-15 mRNA variant found in a small subpopulation of mouse microglia Ž5%, 3 out of 60 cells expressing IL-15 transcripts.. PCR cycle sequencing of this larger transcript revealed the mouse homologue of the alternatively spliced exon A as it is known from the human IL-15 gene. Analysis of the corresponding mouse IL-15 gene region shows that the larger IL-15 transcript contains an yet unidentified 5X sequence of exon 5 while the shorter transcript uses an internal splice acceptor site. The mouse exon 5A segment has a length of 136 nt Ž17 nt longer than the human exon A.. It contains five in-frame stop codons at its 5X end and a new translation initiation site at its 3X end. This new start site is surrounded by a favourable Kozak consensus sequence suggesting a more efficient translation rate. Further translational control by stem–loop binding factors is inferred by a predicted RNA stem–loop structure around the start site. Insertion of exon 5A would lead to an IL-15 polypeptide with a shortened leader sequence of 26 amino acids, as compared to the 48 amino acid leader sequence encoded by the transcript lacking exon 5A. Thus, the final IL-15 protein of the two splice variants is identical; different leader sequences could, however, lead to differences in the intracellular sorting, processing andror secretion of IL-15. q 1998 Elsevier Science B.V. All rights reserved. Keywords: Brain macrophages; Single-cell RT-PCR; Alternative splicing; Interleukin-15
1. Introduction Microglial cells are viewed as the major immunocompetent elements of the CNS. During pathological conditions such as inflammation, stroke or trauma, they transform from a resting, quiescent state into activated forms. Activated microglia differ from their resting counterparts in terms of their morphology, cell surface antigen expression and electrophysiological properties w20,25x. Microglial cells are thought to be the predominant source and target of cytokines in the nervous system. They have been shown to synthesize andror respond to IL-1, IL-3, IL-6, interferon-g, tumor necrosis factor a and granulocyte-macrophage colony stimulating factor at different stages of activation. However, less is known about the role of microglial IL-2 and IL-2 receptors w15x.
) Corresponding author.
[email protected]
Fax:
q49-30-9406-3819;
E-mail:
We have recently characterized the expression of the cytokine interleukin-15 ŽIL-15. and its receptor signaling system in brain microglia w14x. IL-15 represents a member of the four-helix bundle family of cytokines sharing with IL-2 many biological properties and signaling through receptor subunits IL-2Rb and IL-2Rg w1,12,28x. IL-15 mRNA has been detected in a number of tissues and cell types, such as heart, liver, placenta, skeletal muscle, adherent peripheral blood mononuclear cells and in various cell lines. In the CNS, microglial cells express both IL-15 and its functional trimeric IL-15 receptors composed of IL-15Ra , IL-2Rb and IL-2Rg subunits suggesting that IL-15 could act as an autocrine growth factor w14x. In addition, microglia may serve as an IL-15 source for other neural cell populations throughout the brain as well. Furthermore, in vivo, IL-15 could be responsible for several CNS effects currently ascribed to IL-2 w15x. IL-15 cDNA has been isolated and characterized from human, green monkey, rhesus macaque, bovine, porcine,
0169-328Xr98r$ - see front matter q 1998 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 9 - 3 2 8 X Ž 9 8 . 0 0 2 8 4 - 8
156
M. Prinz et al.r Molecular Brain Research 63 (1998) 155–162
rat and mouse tissues w2,8,9,12,29,31x. IL-15 is substantially conserved among mammalian species. It shares 82% homology between rodents and primates. Within each order, the homology score is even higher Žin average 97%.. In all species, the IL-15 mRNA shows an open reading frame of 162 amino acids Žaa. as a precursor protein containing a 48 aa leader sequence that can be cleaved to generate the mature IL-15 polypeptide of 114 aa. In human cell lines of small cell lung cancer, an alternatively spliced variant of IL-15 has been found which forms a truncated protein isoform with a leader sequence of 21 aa residues w22x. However, the functional implications of this shorter precursor protein is still unknown w21x. Here, we show by single-cell RT-PCR, that cultured microglia derived from mouse brain express an alternatively spliced IL-15 mRNA variant ŽIL-15A.. We present the IL-15A mRNA sequence containing an additional, yet unidentified 5X sequence of exon 5. Mechanisms of translational control and consequences for functional IL-15 expression of both IL-15 transcripts are suggested.
2. Materials and methods 2.1. Cell cultures Microglia cells were prepared from cortices of embryonic Žgestation day 14. and newborn Žpostnatal day 0 to 1. mice ŽNaval Medical Research Institute, NMRI, Charles River, Sulzfeld, Germany. as described previously, but with minor modifications w11x. In brief, cortical tissue was carefully freed of blood vessels and meninges. The tissue was trypsinized for 2 min, carefully dissociated with a fire-polished pipette and washed twice. The cortical cells were cultured in Dulbecco’s modified Eagle’s medium ŽDMEM. supplemented with 10% fetal calf serum, with change of serum every third day. After 9 to 12 days, microglia were separated from the underlying astrocytic monolayer by gentle agitation. Microglial cells in the supernatant were washed once and plated on poly-L-lysine ŽPLL.-coated glass coverslips at a density of 10 5 cellsrcm2 . Cells were allowed to grow for two days before they were used for experiments.
The purity of microglial cultures was confirmed by staining sister cultures with biotinylated Griffonia simplicifolia isolectin B4. Briefly, cultures were fixed in 4% paraformaldehyde in phosphate-buffered saline Ž150 mM NaCl, 10 mM sodium phosphate pH 7.4. for 5 min followed by two washes in phosphate-buffered saline, then incubated with biotinylated isolectin B4 Ž1:50, Vector Labs, Burlingame, CA. for 20 min followed by incubation with Cy3-conjugated streptavidin ŽSigma, Deisenhofen.. Cultures routinely contained ) 98% microglia. In addition, all cells from which individual cytoplasmic RNA was harvested, were identified as microglia by their characteristic electrophysiological properties. 2.2. RT-PCR and single-cell RT-PCR of microglia Amplification of mRNA transcripts by RT-PCR was performed with total RNA obtained from virtually bloodfree Žbuffer-perfused. brain of adult mice or cultured microglia Ž10 6 to 10 7 cells. using the Trizol method ŽLife Technologies, Eggenstein, Germany. w10,14x. cDNAs were prepared using Superscript RTe ŽLife Technologies. and amplified in a Thermocycler 9600 ŽApplied Biosystems, Weiterstadt, Germany. by hot-start PCR Ž5 min, 948C. in 35 to 40 cycles Ž30 s, 948C denaturation; 30 s, 55 to 608C annealing; 30 s q 1 s per cycle, 728C elongation. and 10 min at 728C for elongation. Subsequently, the tubes were cooled to 48C and the products analyzed by agarose gel electrophoresis. Primers for mouse IL-15 were derived from the published sequence delivered to the EMBL database: U14332, sense: mmil15-492s ŽIL15-1s., 5X- GAA TAC ATC CAT CTC GTG CTA CT-3X , antisense: mmil15-651as ŽIL15-2as., 5X-GCT TTC AAT TTT CTC CAG GTC-3X , mmil15-913as ŽIL15-1as., 5X-TTT GCA AAA ACT CTG TGA AGG-3X . The primer pairs were selected to cross exon borders distinguishing genomic contaminations. For single-cell RT-PCR, microglial cells were electrophysiologically identified by patch-clamp recording in the whole-cell mode w13,14x. The cytoplasm of individual cells was harvested through the patch pipette containing Žin mM. KCl 130, EGTA 5, CaCl 2 0.5, MgCl 2 3, ATP 3, HEPES 10, pH 7.2, by applying negative pressure. Subsequently, cDNA was synthesized from the
Fig. 1. Analysis of IL-15 expression in mouse microglia. ŽA. IL-15 transcript expression was analyzed by RT-PCR of cytoplasmic RNA individually harvested from cultured microglial cells after patch-clamp recording Žlanes 1–3.. In the majority of cells, cDNA fragments with the predicted length of 160 bp ŽIL-15. were detected Žlane 1.. In few microglial cells, cDNA fragments of 296 bp were detected ŽIL-15A., either alone or in addition to the 160 bp fragment Žlanes 2 and 3.. M, 100 bp molecular weight markers, Br, cDNA fragments from virtually blood-free adult mouse brain homogenate, MG, cDNA fragments derived from microglial culture. ŽB. Whole-cell currents of cultured mouse microglial cell obtained by patch-clamp recording. A series of deand hyperpolarizing voltage steps was applied from a holding potential of Vh s y70 mV. Microglial membrane conductance is dominated by the presence of an inwardly rectifying Kq current. ŽC. Genomic organization of the mouse IL-15 gene. The mouse IL-15 gene is composed of 8 exons. Below the exon–intron structure, the coding sequences ŽCDS. of the two different transcripts are depicted. The shorter transcript is obtained by use of the internal X splice site of exon 5r5A, while the longer transcript Ždescribed here. uses the 5 acceptor site of exon 5A Žhatched box.. In addition, the different polypeptides are shown which can be derived from the transcripts. The vertically hatched areas demarcate the leader sequences. Note that the exon 5A containing transcript may encode two different polypeptides, the IL-15 with a shortened leader sequence and a probably non-functional N-terminal part of 49 amino acids.
M. Prinz et al.r Molecular Brain Research 63 (1998) 155–162
obtained cytoplasmic RNA by reverse transcription. The fragments of interest were amplified by semi-nested PCR approaches. Amplimer identity was confirmed by DNA sequencing carried out in either sense or antisense direction using the Taq DyeDeoxy Terminator Cycle sequencing kit on an automated ABI DNA sequencer Žmodel 373A, Applied Biosystems, Weiterstadt, Germany.. For
157
negative control purposes, solution was expelled from pipettes without patch-clamp recording and treated in the very same way as the experimental samples. PCR cycle sequencing was also applied for analyzing genomic PCR fragments obtained from 1 mg of mouse genomic DNA ŽPromega, Heidelberg, Germany. using primers mmil15492s and mmil15-651as.
158
M. Prinz et al.r Molecular Brain Research 63 (1998) 155–162
2.3. Sequence analysis
3. Results
For DNA sequence analysis the HUSARrGCG program package available at http:rrgenome.dkfz-heidelberg.de was used. RNA secondary structure prediction and free energy calculations were performed using mfold at http:rrwww.ibc.wustl.edur;zuker w16x. The same software was used to recalculate free energies of published RNA stem–loops. mRNA sequences are represented with T instead of U. Leader sequence and splice junction analyses were performed by use of the software packages SignalP w24x and NetStart w5x, respectively, available at http:rrwww.cbs.dtu.dkrservicesr.
We characterized by single-cell RT-PCR the expression of IL-15 mRNA transcripts in cultured microglia derived from mouse brain. Cytoplasmic RNA was harvested from 93 electrophysiologically and morphologically identified microglial cells and subjected to RT-PCR using IL-15 specific primers. All investigated microglial cells displayed the characteristic inwardly rectifying membrane currents activated by a series of de- and hyperpolarizing voltage steps Žranging from y160 to q20 mV in 10 mV increments. starting at the holding potential of y70 mV ŽFig. 1..
Fig. 2. Characterization of the mouse IL-15 gene at the exon 4rexon 5 boundary. Nucleotide sequences of the exons 4 and 5 of the IL-15 gene are shown X in capital letters with the corresponding amino acid sequences below. The novel sequence of exon 5A located directly at the 5 end of exon 5 is shown in bold. In small, italic letters the intervening intron 4 sequence Ž914 bp. is presented. The splice donor site and the two alternatively used splice acceptor sites of exon 5r5A are shaded. The splice junction is indicated by vertical bars. Use of the intron 4rexon 5A splice acceptor site leads to transcripts with an additional insert of 136 nt introducing 5 in-frame stop codons Žunderlined. and a new translation initiation site Žarrowhead..
M. Prinz et al.r Molecular Brain Research 63 (1998) 155–162
3.1. The IL-15 gene is alternatiÕely spliced in mouse microglia Using IL-15 specific primers, cDNA fragments could be amplified from 65% of the cytoplasmic samples Ž60 out of 93.. In 95% Ž n s 57. of these samples, cDNA fragments had the predicted length of 160 bp ŽFig. 1A, lane 1.. However, in a small percentage Ž5%, n s 3., a band at about 300 bp was detected ŽFig. 1A, lanes 2 and 3.. In two out of these three samples, the ; 300 bp fragment was the only amplified fragment, one cell expressed both fragments, 160 and ; 300 bp. The low abundance of this
159
fragment is also substantiated by RT-PCR of RNA derived from microglial culture extracts and mouse brain homogenates. In both cDNA preparations, the ; 300 bp is only a minor band ŽFig. 1A, Br, MG.. PCR sequencing of the smaller 160 bp fragment revealed the expected IL-15 mRNA derived sequence. Sequencing of the ; 300 bp band revealed a sequence also identical to the predicted IL-15 cDNA fragment, except for an inserted stretch of additional 136 bp giving rise to a 296 bp fragment ŽFig. 2.. The inserted sequence contains five stop codons in frame and a new ATG translation initiation site. A sequence comparison showed that the insert is the
Fig. 3. Characterization of the alternatively spliced exon 5A of the mouse IL-15 gene. ŽA. Comparison of mouse exon 5A with human exon A. The sequences of mouse and human exon A are boxed, stop codons are underlined, the ATG start site is depicted in bold. ŽB. Alignment of the N-terminal regions of all known IL-15 polypeptides. So far, only mouse and human IL-15 peptides with shortened leader sequence are known. Amino acids common to all peptides are indicated by asterisks. The first two amino acids of the mature IL-15 are in bold. The arrowhead indicates the predicted cleavage site.
160
M. Prinz et al.r Molecular Brain Research 63 (1998) 155–162
mouse homologue of the alternatively spliced exon A known from the human IL-15 gene ŽFig. 1CFigs. 2 and 3A. w22x. 3.2. Exon 5 r 5A possesses an internal splice acceptor site Using the same primer pair as for RT-PCR, a ; 1200 bp fragment was amplified and sequenced from mouse genomic DNA. The 136 bp sequence could be located on this fragment ŽFig. 2.. At its 5X end, it is separated from exon 4 by an intron of 914 bp Žfor exon designation see Ref. w1x and Fig. 1C.. However, at its 3X end, it is directly linked to the 5X end of exon 5 without any intervening sequences. This suggests, that exon 5 is 136 bp longer than described w1x and that it contains a yet unidentified internal splice acceptor site. The internal splice site is used for splicing of the predominant shorter IL-15 transcript while it is omitted for the longer transcript described here. The 136 bp sequence, now designated exon 5A instead of exon A, is surrounded at its 5X end as well as at its 3X end by splice acceptor junctions which are in good agreement with reported consensus sequences w5x. This further substantiates it to be the 5X part of a complex exon with an internal splice acceptor site w4x. The human exon A, however, is not part of such a complex exon. It has a conventional intron sequence at its 3X end w19,26x. 3.3. Exon 5A leads to a shorter leader sequence of IL-15 by use of a new translation–initiation site Mouse exon 5A is 19 bp longer than its human counterpart ŽFig. 3A.. The extent of homology is only 44% while it is 78% for the other regions of the coding sequence. However, the same structural features are conserved. Human exon A contains also stop codons Žthree TAA. at its 5X end and a new translation initiation site further downstream giving rise to a new leader sequence of 21 aa ŽFig. 3A,B.. Use of the new initiation site leads to an IL-15 protein with a shorter leader sequence Ž26 instead of 48 aa. ŽFig. 3B. while the mature IL-15 protein would not be affected by this splice event. Both leader sequences will probably be cleaved off at the same position w24x. Sequence comparison of the start site regions in both mouse and human in exon 5A revealed a high degree of homology with the translation favourable Kozak consensus sequence w17x ŽFig. 4A.. In contrast, the conventional IL-15 start site did not show such a high similarity. The mouse exon 5A start site contains the particularly favourable purines A at y3 and G at q4 ŽFig. 4A.. In addition, further computer-assisted analysis of the start site context in both mouse IL-15 transcript sequences predicted only the exon 5A-start site as a putative translation initiation site ŽNetStart w27x.. In mouse as well as in human, the translation initiation sites of the alternatively spliced mRNA variants, in contrast to the ‘classical’ IL-15 transcripts, are embedded in
Fig. 4. Translation initiation site within the IL-15 exon A. ŽA. The translation initiation sites present in mouse and in human exon A display a more favourable Kozak consensus sequence than the used sites in exon A-lacking transcripts. Nucleotides of mouse and human IL-15 exon A initiation sites identical to the Kozak consensus sequence are shaded in gray. The start site in mouse exon A contains the particular favourable purines A at y3 and G at q4. ŽB. Secondary RNA-stem loop structures are predicted for the translation initiation site in exon A of mouse Žleft. as well as for human Žright.. For these structures free energies of y7.8 kcalrmol and y7.1 kcalrmol, respectively, are calculated. The ATG codon for initiation is presented in bold.
an RNA-stem loop structure ŽFig. 4B. as predicted by the RNA folding program mfold w16,32,33x. Both stem loops have comparable free energies of y7.8 kcalrmol and y7.1 kcalrmol, respectively. It has to be shown whether these stem–loops are, indeed, targets for translational control factors.
4. Discussion Mouse microglial cells express the IL-15 and IL-15 receptor signaling system which shares many properties with that of IL-2 w14x. In the present study, we further investigated the expression of IL-15 mRNA in mouse brain and microglial cell cultures. Using Žsingle-cell. RTPCR and subsequent PCR sequencing, we identified and characterized the alternatively spliced exon A of the mouse IL-15 gene as the yet unidentified 5X part of exon 5. Therefore, mouse exon A should be designated exon 5A and 5r5A will be used below to describe the complete
M. Prinz et al.r Molecular Brain Research 63 (1998) 155–162
exon 5. The mouse IL-15 gene encodes two different transcripts due to the use of two different splice acceptor sites of exon 5r5A. In cultured mouse microglia as well as in mouse brain and spleen tissue, the transcript IL-15A Žcontaining exon 5A. constitutes only a minor fraction of total IL-15 mRNA, probably less than a few percent. This, however, does not automatically imply that the encoded functional polypeptide represents a similarly minor fraction of the IL-15 protein finally synthesized and released by the cells. As outlined below, there are two levels at which the signal could be amplified: Ži. at the level of translation and Žii. at the level of protein processing and secretion. 4.1. Regulation of IL-15 translation The translational efficacy of the ‘classical’ IL-15 was shown to be negatively influenced by the presence of several translation initiation start sites followed by in-frame stop codons at the 5X ends of IL-15 mRNAs w1,30x. As demonstrated in the present work, insertion of exon 5A in the alternatively spliced variant of IL-15 mRNA adds in-frame stop codons and a new start site. Moreover, this new start site is found to be embedded in a favourable Kozak consensus sequence suggesting its preferred use for translation w18,27x. In such a favourable context, start sites can be one order of magnitude more efficient in translation, thereby even compensating for low abundance of the respective mRNA species w17x. In addition, translation can be modulated by factors binding to the RNA stem–loop structure predicted for the start site of exon 5A. Such regulatory proteins have been described for the translational control of several genes involved in iron metabolism like ferritin or the transferrin receptor w7x. RNA stem–loop binding proteins also activate translation of the human androgen receptor w23x and have been shown to also regulate cytokine expression. A stem–loop of 23 bp, for example, is essential for activated translation of tumor necrosis factor b w6x. The free energy of most of these secondary structures ranges from y7 to y9 kcalrmol, a value comparable to that calculated for the RNA stem–loop of exon 5A in the present work. Thus, the Kozak consensus sequence could facilitate the translation of the alternatively spliced IL-15 while binding of yet unidentified factors binding to the RNA stem cell loop structure could tightly control IL-15 protein synthesis. Therefore, we would like to speculate that IL-15A transcripts are the preferred template for IL-15 synthesis. 4.2. Signal peptide and secretion of IL-15 Besides the translational control, IL-15 secretion will be regulated also at the posttranslational level. IL-15 contains an unusually long signal peptide which apparently interferes with an efficient secretion as experimental manipulation of the leader sequence was shown to have a dramatic
161
effect on IL-15 release w3x. Different signal peptides encoded by the two different IL-15 mRNA isoforms could differentially modulate the processing and secretion of IL-15 protein. Thus far, such studies were only performed in cell lines and could not support a differential secretion of the two human isoforms w21,26x. To make the matters more complex, both posttranslational polypeptide variants, either containing the short or the long signal peptide, were synthesized but not secreted and thus remained intracellularly. In contrast, exchange of the IL-15 leader sequence with that of a conventional transmembrane glycoprotein, such as CD33, led to a substantial release of mature IL-15 w26x. Microglia comprises more than one population of cells which differ in respect to their function or state of activation. The small population of IL-15A expressing cells found in microglial culture, therefore, may represent a distinct population of activated microglia and its functional difference could reside in different levels of IL-15 release. To which extent the existence of two splice variants in total mouse brain mRNA preparations relates to differential IL-15 protein synthesis and release by microglia or other cells is not known. Taken together, mouse microglia express two distinct mRNAs for IL-15 which result from the alternative use of an internal splice site within exon 5r5A. Use of the exon 5A sequence may finally affect the translational efficacy and the rate of IL-15 protein release. Acknowledgements The authors wish to thank Susan A. Lyons for critical comments on the manuscript. The excellent technical assistance of Sibylle Just and Gerda Muller is gratefully ac¨ knowledged. This work was supported by grants from the Federal Ministry for Education and Research of Germany ŽBMBF. and the German Research Foundation ŽDFG, SFB 507rB1, HK.. References w1x D.M. Anderson, L. Johnson, M.B. Glaccum, N.G. Copeland, D.J. Gilbert, N.A. Jenkins, V. Valentine, M.N. Kirstein, D.N. Shapiro, S.W. Morris, K.H. Grabstein, D. Cosman, Chromosomal assignment and genomic structure of IL 15, Genomics 25 Ž1995. 701–706. w2x D.M. Anderson, S. Kumaki, M. Ahdieh, J. Bertles, M. Tometsko, A. Loomis, J. Giri, N.G. Copeland, D.J. Gilbert, N.A. Jenkins, V. Valentine, D.N. Shapiro, S.W. Morris, L.S. Park, D. Cosman, Functional characterization of the human interleukin-15 receptor alpha chain and close linkage of IL15RA and IL2RA genes, J. Biol. Chem. 270 Ž1995. 29862–29869. w3x R.N. Bamford, A.P. Battiata, J.D. Burton, H. Sharma, T.A. Waldmann, Interleukin ŽIL. 15rIL-T production by the adult T-cell leukemia cell line HuT-102 is associated with a human T-cell lymphotrophic virus type I region rIL-15 fusion message that lacks many upstream AUGs that normally attenuates IL-15 mRNA translation, Proc. Natl. Acad. Sci. U.S.A. 93 Ž1996. 2897–2902.
162
M. Prinz et al.r Molecular Brain Research 63 (1998) 155–162
w4x R.E. Breitbart, A. Andreadis, G.B. Nadal, Alternative splicing: a ubiquitous mechanism for the generation of multiple protein isoforms from single genes, Annu. Rev. Biochem. 56 Ž1987. 467–495. w5x S. Brunak, J. Engelbrecht, S. Knudsen, Prediction of human mRNA donor and acceptor sites from the DNA sequence, J. Mol. Biol. 220 Ž1991. 49–65. w6x L. Buonaguro, F.M. Buonaguro, G. Giraldo, B. Ensoli, The human immunodeficiency virus type 1 Tat protein transactivates tumor necrosis factor beta gene expression through a TAR-like structure, J. Virol. 68 Ž1994. 2677–2682. w7x J. Butt, H.Y. Kim, J.P. Basilion, S. Cohen, K. Iwai, C.C. Philpott, S. Altschul, R.D. Klausner, T.A. Rouault, Differences in the RNA binding sites of iron regulatory proteins and potential target diversity, Proc. Natl. Acad. Sci. U.S.A. 93 Ž1996. 4345–4349. w8x A. Canals, L.C. Gasbarre, P.C. Boyd, S. Almeria, D.S. Zarlenga, Cloning and expression of bovine interleukin-15: analysis and modulation of transcription by exogenous stimulation, J. Interferon. Cytokine Res. 17 Ž1997. 473–480. w9x A. Canals, D.R. Grimm, L.C. Gasbarre, J.K. Lunney, D.S. Zarlenga, Molecular cloning of cDNA encoding porcine interleukin-15, Gene 195 Ž1997. 337–339. w10x P. Chomczynski, A reagent for the single-step simultaneous isolation of RNA, DNA and proteins from cell and tissue samples, Biotechniques 15 Ž1993. 532–534, 536–537. w11x D. Giulian, T.J. Baker, Characterization of ameboid microglia isolated from developing mammalian brain, J. Neurosci. 6 Ž1986. 2163–2178. w12x K.H. Grabstein, J. Eisenman, K. Shanebeck, C. Rauch, S. Srinivasan, V. Fung, C. Beers, J. Richardson, M.A. Schoenborn, M. Ahdieh, L. Johnson, M.R. Alderson, J.D. Watson, D.M. Anderson, J.G. Giri, Cloning of a T cell growth factor that interacts with the beta chain of the interleukin-2 receptor, Science 264 Ž1994. 965–968. w13x O.P. Hamill, A. Marty, E. Neher, B. Sakmann, F.J. Sigworth, Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches, Pflugers Arch. 391 ¨ Ž1981. 85–100. w14x U.K. Hanisch, S.A. Lyons, M. Prinz, C. Nolte, J.R. Weber, H. Kettenmann, F. Kirchhoff, Mouse brain microglia express interleukin-15 and its multimeric receptor complex functionally coupled to janus kinase activity, J. Biol. Chem. 272 Ž1997. 28853–28860. w15x U.K. Hanisch, R. Quirion, Interleukin-2 as a neuroregulatory cytokine., Brain Res. Brain Res. Rev. 21 Ž1995. 246–284. w16x J.A. Jaeger, D.H. Turner, M. Zuker, Improved predictions of secondary structures for RNA, Proc. Natl. Acad. Sci. U.S.A. 86 Ž1989. 7706–7710. w17x M. Kozak, The scanning model for translation: an update, J. Cell Biol. 108 Ž1989. 229–241. w18x M. Kozak, Regulation of translation in eukaryotic systems, Annu. Rev. Cell Biol. 8 Ž1992. 197–225.
w19x H. Krause, B. Jandrig, C. Wernicke, P.S. Bulfone, T. Pohl, T. Diamantstein, Genomic structure and chromosomal localization of the human interleukin 15 gene ŽIL-15., Cytokine 8 Ž1996. 667–674. w20x G.W. Kreutzberg, Microglia: a sensor for pathological events in the CNS, Trends Neurosci. 19 Ž1996. 312–318. w21x R. Meazza, A. Gaggero, F. Neglia, S. Basso, S. Sforzini, R. Pereno, B. Azzarone, S. Ferrini, Expression of two interleukin-15 mRNA isoforms in human tumors does not correlate with secretion: role of different signal peptides, Eur. J. Immunol. 27 Ž1997. 1049–1054. w22x R. Meazza, S. Verdiani, R. Biassoni, M. Coppolecchia, A. Gaggero, A.M. Orengo, M.P. Colombo, B. Azzarone, S. Ferrini, Identification of a novel interleukin-15 ŽIL-15. transcript isoform generated by alternative splicing in human small cell lung cancer cell lines, Brain Res. 12 Ž1996. 2187–2192. w23x A. Mizokami, C. Chang, Induction of translation by the 5X-untranslated region of human androgen receptor mRNA, J. Biol. Chem. 269 Ž1994. 25655–25659. w24x H. Nielsen, J. Engelbrecht, S. Brunak, G. von Heyne, ¨ Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites, Protein Eng. 10 Ž1997. 1–6. w25x W. Norenberg, P.J. Gebicke Haerter, P. Illes, Inflammatory stimuli ¨ induce a new Kq outward current in cultured rat microglia, Neurosci. Lett. 147 Ž1992. 171–174. w26x A. Onu, T. Pohl, H. Krause, S. Bulfone Paus, Regulation of IL-15 secretion via the leader peptide of two IL-15 isoforms, J. Immunol. 158 Ž1997. 255–262. w27x A.G. Pedersen, H. Nielsen, Neural network prediction of translation initiation sites in eukaryotes: perspectives for EST and genome analysis, Proceedings of the Fifth International Conference on Intelligent Systems for Molecular Biology ŽISMB 5., 1997, 226–233. w28x D.K. Pettit, T.P. Bonnert, J. Eisenman, S. Srinivasan, R. Paxton, C. Beers, D. Lynch, B. Miller, J. Yost, K.H. Grabstein, W.R. Gombotz, Structure–function studies of interleukin 15 using site-specific mutagenesis, polyethylene glycol conjugation, and homology modeling, J. Biol. Chem. 272 Ž1997. 2312–2318. w29x H.C. Reinecker, R.P. MacDermott, S. Mirau, A. Dignass, D.K. Podolsky, Intestinal epithelial cells both express and respond to interleukin 15, Gastroenterology 111 Ž1996. 1706–1713. w30x Y. Tagaya, R.N. Bamford, A.P. DeFilippis, T.A. Waldmann, IL-15: a pleiotropic cytokine with diverse receptorrsignaling pathways whose expression is controlled at multiple levels, Immunity 4 Ž1996. 329–336. w31x F. Villinger, S.S. Brar, A. Mayne, N. Chikkala, A.A. Ansari, Comparative sequence analysis of cytokine genes from human and nonhuman primates, J. Immunol. 155 Ž1995. 3946–3954. w32x M. Zuker, Computer prediction of RNA structure, Methods Enzymol. 180 Ž1989. 262–288. w33x M. Zuker, On finding all suboptimal foldings of an RNA molecule, Science 244 Ž1989. 48–52.