Regulation of CD45 Alternative Splicing by Heterogeneous Ribonucleoprotein, hnRNPLL Shalini Oberdoerffer, et al. Science 321, 686 (2008); DOI: 10.1126/science.1157610

The following resources related to this article are available online at www.sciencemag.org (this information is current as of February 1, 2009 ):

Updated information and services, including high-resolution figures, can be found in the online version of this article at: http://www.sciencemag.org/cgi/content/full/321/5889/686

Supporting Online Material can be found at: http://www.sciencemag.org/cgi/content/full/1157610/DC1 A list of selected additional articles on the Science Web sites related to this article can be found at: http://www.sciencemag.org/cgi/content/full/321/5889/686#related-content This article cites 21 articles, 8 of which can be accessed for free: http://www.sciencemag.org/cgi/content/full/321/5889/686#otherarticles on February 1, 2009 This article has been cited by 1 article(s) on the ISI Web of Science.

This article has been cited by 2 articles hosted by HighWire Press; see: http://www.sciencemag.org/cgi/content/full/321/5889/686#otherarticles

This article appears in the following subject collections: Immunology http://www.sciencemag.org/cgi/collection/immunology www.sciencemag.org Information about obtaining reprints of this article or about obtaining permission to reproduce this article in whole or in part can be found at: http://www.sciencemag.org/about/permissions.dtl Downloaded from

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright 2008 by the American Association for the Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS. REPORTS

Fig. 3. Effect of litter quality on decom- 17. H. W. Hunt et al., Biol. Fertil. Soils 3, 57 (1987). poser stoichiometry. (A) Decomposer effi- 18. W. Parton et al., Science 315, 361 (2007). 19. O. N. Krankina, M. E. Harmon, A. V. Griazkin, Can. J. For. ciency, e (left), or rCR (right) as a function r r Res. 29, 20 (1999). of L,0 when B = 0.1. Symbols indicate 20. J. B. Russell, G. M. Cook, Microbiol. Rev. 59, 48 (1995). different litter types as in Fig. 2; ◊ and ♦ 21. P. A. del Giorgio, J. J. Cole, Annu. Rev. Ecol. Syst. 29, 503 refer to a decomposing log (rB =0.122) (1998). 22. J. J. Elser et al., Nature 408, 578 (2000). and the underlying soil (rB =0.135),re- spectively [data elaborated after (16, 30)]. 23. W. J. Mattson, Annu. Rev. Ecol. Syst. 11, 119 (1980). The solid line is a linear least square fit of 24. T. J. Pandian, M. P. Marian, Freshw. Biol. 16, 93 (1986). r r r 25. F. Slansky, P. Feeny, Ecol. Monogr. 47, 209 (1977). the log-transformed CR and L,0 ( CR = 26. J. M. Craine, C. Morrow, N. Fierer, Ecology 88, 2105 0.76 0.45 × rL,0 ; R =0.88;P < 0.0001). The (2007). shaded area shows the effects on e of 27. W. B. McGill, H. W. Hunt, R. G. Woodmansee, J. O. Reuss, in Terrestrial Nitrogen Cycles: Processes, Ecosystem different rB around 0.1 (solid line). The Strategies and Management Impacts, F. E. Clark, dashed curve indicates points where rCR = r T. Rosswall, Eds. (Ecological Bulletins, Stockholm, 1981), L,0: Litter points above this curve need to vol. 33, pp. 49–115. immobilize N; points below release N 28. M. E. Harmon, Long-Term Ecological Research (LTER) since the beginning of decomposition. Intersite Litter Decomposition Experiment (LIDET), Forest (B)Estimatesofe as a function of the ratio Science Data Bank code TD023, Corvallis, OR, 2007, www.fsl.orst.edu/lter/data/abstract.cfm?dbcode=TD023. between food source N:C (rF) and consumer N:C (r )atdifferenttrophiclevels:□,ter- 29. H. L. Gholz, D. A. Wedin, S. M. Smitherman, M. E. Harmon, B W. J. Parton, Glob. Change Biol. 6, 751 (2000). restrial plant residue decomposers (this 30. S. C. Hart, G. E. Nason, D. D. Myrold, D. A. Perry, Ecology study); +, marine bacteria (21); ○,terres- 75, 880 (1994). trial larvae (25); ●, terrestrial insects (23); 31. This research was supported by Department of Energy and ×, aquatic insects (24). The solid line is a (DOE) Forest-Atmosphere Carbon Transfer and Storage linear least square fit of the log-transformed project (FACT-1), NSF DEB 0235425 and 0717191, and e r r e r r 0.60 R DOE PER 64242-0012346. LIDET data sets were provided and F/ B [ =0.43×(F/ B) ; = 0.72; by the Forest Science Data Bank, a partnership between P < 0.0001]. the Department of Forest Science, Oregon State University, and the U.S. Forest Service Pacific Northwest on February 1, 2009 Research Station, Corvallis, Oregon. Significant funding for these data was provided by the NSF Long-Term Ecological Research program (DEB-02-18088). Funding for the CIDET experiment was provided by Climate Change and Ecosystem Processes Network of the Canadian Forest Service and Natural Resources Canada Panel on Energy Research Development. We also thank G. Katul, D. Richter, and two anonymous reviewers for useful comments. establishment report,” Tech. Rep. No. BC-X-378 (Pacific poser N:C ratio is relatively constant, this pattern Supporting Online Material suggests that the decomposer communities are able Forestry Centre, Victoria, Canada, 1998). 13. See supporting materials on Science Online. www.sciencemag.org/cgi/content/full/321/5889/684/DC1

to adapt partially to low-nitrogen substrates (i.e., 14. W. J. Parton, D. S. Schimel, C. V. Cole, D. S. Ojima, Materials and Methods www.sciencemag.org Fig. S1 low rL,0) by decreasing their C-use efficiency and Soil Sci. Soc. Am. J. 51, 1173 (1987). thus the critical N:C of the litter (Fig. 3A). Such a 15. T. R. Moore, J. A. Trofymow, C. E. Prescott, J. Fyles, Table S1 pattern has been observed in aquatic environments B. D. Titus, Ecosystems (New York) 9, 46 (2006). 29 April 2008; accepted 1 July 2008 and at other trophic levels (21–25) and appears to 16. S. C. Hart, Ecology 80, 1385 (1999). 10.1126/science.1159792 be a universal response of decomposers in nutrient- poor conditions (Fig. 3B). Decreasing efficiency results in higher heterotrophic respiration per unit Regulation of CD45 Alternative mass of litter humified or unit nutrient released, Downloaded from suggesting that the soil carbon cycle is likely more open than currently thought. Splicing by Heterogeneous

References and Notes Ribonucleoprotein, hnRNPLL 1. B. Berg, C. A. McClaugherty, Plant Litter: Decomposition, Humus Formation, Carbon Sequestration (Springer, 1 2 1 2 Berlin, 2003). Shalini Oberdoerffer, Luis Ferreira Moita, * Daniel Neems, Rui P. Freitas, * 2,3 1 2. M. J. Swift, O. W. Heal, J. M. Anderson, Decomposition Nir Hacohen, Anjana Rao † in Terrestrial Ecosystems, vol. 5 of Studies in Ecology (Univ. of California Press, Berkeley, 1979). The transition from naïve to activated T cells is marked by alternative splicing of pre-mRNA encoding the 3. S. A. Waksman, J. Agric. Sci. 14, 555 (1924). 4. J. D. Aber, J. M. Melillo, Can. J. Bot. 60, 2263 (1982). transmembrane phosphatase CD45. Using a short hairpin RNA interference screen, we identified 5. G. Seneviratne, Biol. Fertil. Soils 31, 60 (2000). heterogeneous ribonucleoprotein L-like (hnRNPLL) as a critical inducible regulator of CD45 alternative 6. B. Berg, G. Ekbohm, Ecology 64, 63 (1983). splicing. HnRNPLL was up-regulated in stimulated T cells, bound CD45 transcripts, and was both necessary 7. C. C. Cleveland, D. Liptzin, Biogeochemistry 85, 235 (2007). and sufficient for CD45 alternative splicing. Depletion or overexpression of hnRNPLL in B and T cell 8. R. W. Sterner, J. J. Elser, Ecological Stoichiometry: The lines and primary T cells resulted in reciprocal alteration of CD45RA and RO expression. Exon array Biology of Elements from Molecules to the Biosphere (Princeton Univ. Press, Princeton, NJ, 2002). analysis suggested that hnRNPLL acts as a global regulator of alternative splicing in activated T cells. 9. S. Manzoni, A. Porporato, Soil Biol. Biochem. 39, 1542 (2007). Induction of hnRNPLL during hematopoietic cell activation and differentiation may allow cells to rapidly 10. G. I. A˚gren, E. Bosatta, Theoretical Ecosystem Ecology: shift their transcriptomes to favor proliferation and inhibit cell death. Understanding Element Cycles (Cambridge Univ. Press, Cambridge, 1996). 1 2 11. E. Bosatta, H. Staaf, Oikos 39, 143 (1982). t is estimated that greater than 75% of genes genome ( , ). SR (serine-arginine rich) proteins 12. J. A. Trofymow, CIDET, “The Canadian Intersite yield alternative transcripts, contributing to are key positive regulators of alternative splicing Decomposition Experiment (CIDET): Project and site Iconsiderable functional diversity within the that bind enhancer sequences on nascent tran-

686 1 AUGUST 2008 VOL 321 SCIENCE www.sciencemag.org REPORTS scripts and recruit spliceosomal proteins to weak ing, thus forcing a shift of splicing to distal splice alternative splicing that culminates in increased splice sites, thereby facilitating proximal spliceo- sites (4). CD45RO and decreased CD45RA+ transcripts some assembly (3). Heterogeneous nuclear ribo- CD45 is an abundant transmembrane protein within 24 hours, and this can be blocked by nucleoproteins (hnRNPs) also regulate splicing tyrosine phosphatase expressed at the surface of cycloheximide treatment, which suggests the in- by binding negative cis-regulatory elements and T cells, B cells, and other hematopoietic cells (5). volvement of de novo protein synthesis (10). causing exon exclusion from mature mRNA (3). CD45 transcripts undergo extensive alternative To identify trans-acting factors required for SR proteins and hnRNPs function as antagonists splicing in which exons 4, 5, and 6 are variably CD45RO expression in stimulated T cells, we in alternative splicing, with binding of hnRNPs excluded (Fig. 1A) (6). Primary naive T cells and performed an RNA interference (RNAi) screen to silencer sequences inhibiting SR protein bind- B cells express the larger isoforms and are in JSL1 T cells, which basally express both referred to as RA+. In contrast, activated and CD45RO and CD45RA, but in which phorbol 1Department of Pathology, Immune Disease Institute, Harvard 2 memory T cells express the shortest isoform, 12-myristate 13-acetate (PMA) stimulation gen- Medical School, Boston, MA 02115, USA. Center for Immu- CD45RO (5). CD45 initiates signaling through erates a relative increase in the ratio of RO to nology and Inflammatory Diseases, Division of Rheumatology, 10 Allergy, and Immunology, Massachusetts General Hospital, antigen receptors by dephosphorylating the in- RA transcripts through alternative splicing ( ). Charlestown, MA 02129, USA. 3Broad Institute of Harvard and hibitory tyrosine on Src-family kinases (5), but JSL1 cells were infected in duplicate with in- Massachusetts Institute of Technology, Cambridge, MA 02142, attributing specific functions to individual CD45 dividual short hairpin RNAs (shRNAs) from a USA. isoforms has been complicated by the high basal splicing factor–directed library (11, 12), selected *Present address: Instituto de Medicina Molecular, Facul- expression of CD45 in lymphocytes and its with puromycin, stimulated with PMA (40 hours), dade de Medicina, Universidade de Lisboa, 1649-028 Lisboa, Portugal. ability to dephosphorylate both inhibitory and and assayed for CD45RO expression (Fig. 1B). †To whom correspondence should be addressed. E-mail: activation-loop tyrosines of Src-family kinases Knockdown of a single candidate, hnRNPLL [email protected] (7–9). Stimulated T cells show a shift in CD45 (heterogeneous ribonucleoprotein L-like), sub-

Fig. 1. Identification of A B 100 hnRNPLL in an shRNA screen. (A)CD45isoform RO 80 Uninf. Unstim. RBC RAB nomenclature. (B)Sam- 60 Uninf. +PMA ple histogram of cell shRNA- Neg. Regul. 1 3 4 5 6 7 33 surface-CD45RO expres- 40 shRNA- Pos. Regul. sion in uninfected JSL1 ABCRABC on February 1, 2009 cells with or without PMA RB 20

stimulation and JSL1 cells Relative Cell Number 0 transduced with lentivi- 100 101 102 103 104 ral shRNAs. (C) Repre- CD45RO sentative results from a single 96-well plate in C D RT-PCR the screen. All hnRNPLL 10000 3 hnRNPL hairpins decreased PMA- hnRNPLL 2.5 induced CD45RO ex- hnRNPLL 1000 www.sciencemag.org pression. (D)Quantitative Mean MFI 2 CD45RO/RA +/- 1 s.d. RT-PCR for hnRNPLL,

MFI 100 1.5 hnRNPL, CD45RO, and Unstim. CD45RA in JSL1 cells stim- 1 10 ulated with PMA for 12 0.5 and 24 hours. PMA stim- ulation increases hnRNPLL 1 0 Relative mRNA Expression 01224 expression and the ratio of Individual Hairpins Hours Post-Stimulation Downloaded from CD45RO to RA transcripts. (E) Quantitative RT-PCR and anti-hnRNPLL immu- EFRT-PCR Resting PMA-Stimulated noblot confirmed hnRNPLL 1.2 100 100 depletion. (F) Cell-surface 1 expression of CD45 iso- 80 80 0.8 forms assessed in PMA- 60 60 stimulated JSL1 cells stably 0.6 expressing LL-sh1 and 0.4 40 40 LL-sh4. HnRNPLL depletion 0.2 20 20 decreased CD45RO expres- 0 0 0 Uninf. Unstim. sion and increased CD45RA 10 0 10 1 10 2 10 3 10 4 10 0 10 1 10 2 10 3 10 4 Uninf. +PMA expression. Relative mRNA Expression CD45RO LL-sh1 LL-sh4 100 100

80 80 Immunoblot 60 60 Relative Cell Number LL-sh1 LL-sh4 JSL1 Vector shRFP

hnRNPLL 40 40

Rel A 20 20

0 0 100 101 102 103 104 100 101 102 103 104 CD45RA

www.sciencemag.org SCIENCE VOL 321 1 AUGUST 2008 687 REPORTS

stantially reduced PMA-induced CD45RO ex- synthesized factor that promotes alternative splic- ates alternative splicing rather than repressing pression, with three of the five hairpins reducing ing in PMA-stimulated JSL1 cells (10). CD45 transcription. it to a level lower than the unstimulated control In JSL1 cells engineered for stable expression To further examine the influence of hnRNPLL (Fig. 1C). of two of the five lentiviral shRNAs (LL-sh1 and expression, JSL1 cells were transduced with a lenti- Concurrent with the increased ratio of CD45RO LL-sh4), a substantial reduction of hnRNPLL virus encoding bicistronic hnRNPLL and inter- to RA transcripts, PMA stimulation of JSL1 cells protein expression was also observed (Fig. 1E). nal ribosomal entry site–green fluorescent protein increased hnRNPLL transcripts by a factor of Cells expressing these shRNAs showed decreased (IRES-GFP) as a means of tracking hnRNPLL ~2.0 in 24 hours (Fig. 1D), without affecting ex- CD45RO expression, and a concomitant increase expression in these cells. HnRNPLL overexpres- pression of its paralog hnRNPL (heterogeneous in CD45RA expression, under both resting and sion was readily apparent in the total cell pop- ribonucleoprotein L). These data are consistent stimulated conditions (Fig. 1F). This result is ulation (Fig. 2A), even though only ~35% of the with the hypothesis that hnRNPLL is the de novo consistent with the notion that hnRNPLL medi- cells were successfully transduced as judged by

Fig. 2. Overexpression or A B JSL1 GFP+ 100 100 knockdown of hnRNPLL JSL1 Vector LL Immunoblot influences CD45 alter- 80 80 native splicing. (A)Im- hnRNPLL 60 60 hnRNPLL munoblot for hnRNPLL Vector expression in JSL1 cells ret- RelA 40 40 rovirally transduced with 20 20

IRES-GFP hnRNPLL or emp- 0 0 100 101 102 103 104 100 101 102 103 104 ty vector. (B)HnRNPLLover- Relative Cell Number expression in GFP+ JSL1 C CD45RO CD45RA cells increased CD45RO hnRNPLL overexpression: primary cells and decreased CD45RA 100 100 D hnRNPLL knockdown: primary cells 100 100 expression in the absence 80 80 80 80 of stimulation. (C)Increased 60 60 hnRNPLL LL-sh4 CD45RO expression and Vector 60 60 Vector on February 1, 2009 decreased CD45RA expres- 40 40 40 40 sion in GFP+, hnRNPLL- 20 20 20 20 transduced primary human 0 0 + 100 101 102 103 104 100 101 102 103 104 CD4 T cells. (D) Reduced Relative Cell Number 0 0 0 1 2 3 4 0 1 2 3 4 CD45RO expression and in- CD45RO CD45RA Relative Cell Number 10 10 10 10 10 10 10 10 10 10 creased CD45RA expres- CD45RO CD45RA sion in hnRNPLL-depleted E + primary CD4 T cells. (E) 100 100 F RT-PCR CD45 isoform expression in 1.6 80 80 1.4 BJAB and BL41 B cell lines. BL41 1.2 www.sciencemag.org 60 60 (F) Detection of hnRNPLL BJAB 1 BL41 transcripts in BJAB but not 40 40 0.8 0.6 BJAB BL41 cells (value shown 20 20 numerically) by quantitative 0.4 0 0 0.2 RT-PCR. (G) Immunoblot 100 101 102 103 104 100 101 102 103 104 0 8.67E-05 +/-9.51E-05 Relative Cell Number for hnRNPLL expression CD45RO CD45RA Level Relative mRNA hnRNPL hnRNPLL in BL41 cells retrovirally

transduced with IRES-GFP G BL41 Vector LL Immunoblot Downloaded from hnRNPLL or empty vector. hnRNPLL H BL41 GFP+ (H) Increased CD45RO and 100 100 decreased CD45RA expres- + RelA 80 80 sion in GFP hnRNPLL- hnRNPLL transduced BL41 cells. (I) 60 60 Vector Quantitative RT-PCR and 40 40 1.4 immunoblot for hnRNPLL I RT-PCR in knockdown BJAB cells. 1.2 20 20 1 0 0 (J) Decreased CD45RO and 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 0.8 Relative Cell Number increased CD45RA expres- CD45RO CD45RA sion in LL-sh1 and LL-sh4 0.6 BJAB cells. 0.4 J 0.2 BJAB

Relative mRNA Level Relative mRNA 0 100 100 LL- LL- LL- LL- LL- shRFP sh1 sh2 sh3 sh4 sh5 80 80 LLsh-1

60 60 LL-sh4

40 40 Uninfected LL-sh2 LL-sh4 BJAB shRFP Vector Immunoblot hnRNPLL 20 20 0 0 100 101 102 103 104 100 101 102 103 104 RelA Relative Cell Number CD45RO CD45RA

688 1 AUGUST 2008 VOL 321 SCIENCE www.sciencemag.org REPORTS

GFP expression (fig. S1). The entire population of tively express high levels of the largest isoform, CD45 alternative splicing in T cells and also in hnRNPLL-expressing (GFP+) cells up-regulated CD45RABC (B220) and do not express CD45RO B cells. CD45RO, and CD45RA expression was elimi- (e.g., BL41) (Fig. 2E). While screening addi- The paralog of hnRNPLL, hnRNPL, was pre- nated (Fig. 2B). Thus, hnRNPLL expression is tional B cell lines for CD45 isoform expression, viously identified as binding to an exonic splic- both necessary and sufficient for CD45 alternative however, we found that BJAB B cells are 100% ing silencer, ESS1, located in exon 4 of CD45 splicing and CD45RO expression in JSL1 cells. positive for CD45RO and do not express CD45RA+ (14, 15). However, it was not identified in our We next tested the ability of hnRNPLL to in- isoforms (Fig. 2E). HnRNPLL expression di- screen (fig. S2A), and we traced this to an in- fluence CD45 isoform expression in primary hu- rectly correlated with CD45 isoform expression ability to knock down hnRNPL protein expres- man CD4+ T cells. Stimulated naïve CD4+ T cells in these two cell lines, with virtually undetectable sion effectively in JSL1 cells (fig. S2B). In BJAB were infected with lentiviruses encoding IRES-GFP CD45RO and hnRNPLL in BL41 cells (Fig. 2, cells however, L-sh2 substantially diminished hnRNPLL (for overexpression) or LL-sh4 (for E to G) and high levels of hnRNPLL and hnRNPL expression (Fig. 3A, left), with no ob- RNAi-mediated knockdown) and cultured in CD45RO in BJAB cells (Fig. 2, E, F, and I). servable effect on CD45 isoform expression (Fig. interleukin-2–containing medium for 5 to 7 days. Transcript levels of the paralog, hnRNPL, were 3B). Conversely, overexpression of hnRNPL, either As expected, mock-infected primary T cells up- similar in the two cell lines (Fig. 2F). When in JSL1 or BL41 cells, had no effect on CD45 regulated CD45RO and down-regulated CD45RA hnRNPLL was overexpressed in BL41 cells using isoform expression (Fig. 3, C and D). Given the after activation, and this was also seen in hnRNPLL- an IRES-GFP lentiviral expression plasmid (Fig. homology between hnRNPL and hnRNPLL expressing (GFP+) T cells, but to a greater extent 2G and fig. S1), a total population shift was (69% identity at the amino acid level) (16), we (Fig. 2C). In contrast, primary T cells depleted seen, such that infected (GFP+) cells uniform- confirmed that hnRNPLL knockdown did not of hnRNPLL were 100% CD45RO– and CD45RA+ ly expressed CD45RO and lost expression of result in an “off-target” decrease in hnRNPL ex- (Fig. 2D). CD45RA+ isoforms (Fig. 2H). Conversely, de- pression (fig. S2C). Unlike T cells, which readily undergo CD45 pletion of hnRNPLL with LL-sh1 and LL-sh4 We next used RNA-immunoprecipitation (RIP) alternative splicing upon activation, B cells are shRNAs in BJAB cells decreased hnRNPLL tran- (17) followed by quantitative reverse transcription refractory to the CD45RA > RO switch, and PMA script and protein expression (Fig. 2I) and caused polymerase chain reaction (RT-PCR) to test the stimulation of B cell lines does not promote a total population shift of the infected cells to association between hnRNPLL and CD45 tran- exon 4 to 6 exclusion and consequent CD45RO CD45ROlow and CD45RA+ (Fig.2J).These scripts. Because the only commercially available expression (10, 13). Most B cell lines constitu- data confirm an important role for hnRNPLL in antibody directed against hnRNPLL is not suit- on February 1, 2009 Fig. 3. Effect of hnRNPL A B BJAB 1 on CD45 isoform expres- RT-PCR 100 100 sion. (A) Quantitative RT- 0.8 P 80 80 PCR and immunoblot for 0.6 h1 Immunoblot 60 60 L-sh2 hnRNPL in knockdown L-sh3 L-sh2 L-s BJAB shRF Vector 0.4 BJAB cells. (B)Cell-surface hnRNPL 40 40 Vector CD45 isoform expression 0.2 RelA 20 20 was unaltered in L-sh2 0 0 0 0 1 2 3 4 0 1 2 3 4 Relative mRNA Level Relative mRNA BJAB cells. (C)JSL1and L-sh1 L-sh2 L-sh3 L-sh4 shRFP Relative Cell Number 10 10 10 10 10 10 10 10 10 10 (D) BL41 cells retrovirally CD45RO CD45RA www.sciencemag.org transduced with FLAG- CDJSL1 tagged hnRNPL (L-FLAG) BL41 were immunoblotted for 100 100 100 100 hnRNPL and FLAG proteins. 80 80 80 80 HnRNPL overexpression did 60 60 hnRNPL 60 60 hnRNPL notaffectCD45isoformex- 40 40 Vector 40 40 Vector pression. (E) FLAG-tagged 20 20 20 20 hnRNPLL (LL-FLAG) or en- Downloaded from 0 0 0 0 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4

dogenous hnRNPL was Relative Cell Number 10 10 10 10 10 10 10 10 10 10 Relative Cell Number 10 10 10 10 10 10 10 10 10 10 immunoprecipitated from CD45RO CD45RA CD45RO CD45RA stably transduced or un- infected JSL1 cells with Immunoblot antibody to FLAG or anti- JSL1 Vector L-FLAG BL41 Vector L-FLAG Immunoblot body to hnRNPL (anti-L), hnRNPL hnRNPL respectively. Associated FLAG RNA was isolated, and FLAG quantitative RT-PCR was RelA RelA performed to detect CD45 transcripts. Immunoblot E F of LL-FLAG–transduced 7 2.5% input IP: Anti-FLAG RT-PCR JSL1 cells indicated that 6 hnRNPLL was not over- 5 JSL1 expressed relative to en- 4

LL-FLAG (Eluted) Vector LL-FLAG Immunoblot dogenous hnRNPLL. (F) 3 JSL1 LL-FLAG JSL1 Vector LL-FLAG Immunoblot hnRNPLL Anti-FLAG immunoprecip- 2 hnRNPL – itation from LL-FLAG 1 FLAG transduced JSL1 cells 0 RelA followed by immuno- JSL1 LL-FLAG JSL1 JSL1 JSL1 blotting for hnRNPL. Level Relative CD45 Transcript IP: Anti-FLAG Anti-hnRNPL

www.sciencemag.org SCIENCE VOL 321 1 AUGUST 2008 689 REPORTS

able for immunoprecipitation, FLAG-tagged 3F), showing that the two proteins interact with 4A, right) (6). Increased exon retention in re- hnRNPLL or empty vector were stably expressed each other and with CD45 transcripts. sponse to hnRNPLL depletion was confirmed in in JSL1 cells. Cells overexpressing FLAG- Exon array analysis (18) was next used to test peripheral T cells by RT-PCR for CD45 exon 4 hnRNPLL showed only a marginal increase in the possibility that hnRNPLL regulates a broader andCD44exon10,aswellasforSTAT5Aexon5 total hnRNPLL detected with an antibody to program of alternative splicing in activated T cells. (Fig. 4B). There is evidence for the existence of hnRNPLL (Fig. 3E, right). Moreover, FLAG- A total of 132 genes, including CD45 (protein both latter transcripts (fig. S4B) (19). A com- hnRNPLL could not be detected in lysates by tyrosine phosphatase, receptor type C) itself (Fig. plete analysis of alternative exon usage patterns immunoblotting with an antibody to FLAG, al- 4A), showed significant alternative exon usage and their potential functional consequences is though it was detected upon loading high amounts (P < 0.01) in response to hnRNPLL knockdown, being pursued. of anti-FLAG immunoprecipitate (Fig. 3E and but not in response to shGFP infection (table S3). Primary T cells lentivirally infected with LL-sh4 fig. S3C). Nevertheless, anti-FLAG immunopre- Of these, 36 genes showed significant alternative (shRNA directed against hnRNPLL) were con- cipitates from FLAG-hnRNPLL-expressing JSL1 usage of corresponding exons in response to sistently difficult to expand, in contrast to pri- cells showed enrichment of CD45 transcripts activation of cord blood cells (table S4), correlating mary T cells infected with control shRNA directed (exon 4 sequences) by a factor of 20 relative to with an increase in hnRNPLL expression by a against red fluorescent protein (RFP) (Fig. 4C). vector-transduced cells (Fig. 3E). Based on CD45 factor of 5 in the activated cells. Expression of Moreover, these hnRNPLL-depleted T cells re- transcript expression in 5% of the input fraction, CD45 exons 4 and 6 was elevated in naïve cord peatedly showed reduced expression of CD25 com- we calculated that the FLAG-hnRNPLL immuno- blood and hnRNPLL-depleted peripheral T pared with control shRFP-expressing cells (Fig. precipitates bound ~0.25% of all CD45 transcripts cells, relative to their expression in activated 4D). These data are consistent with our finding in the lysates. By comparison, immunoprecipi- cord blood and peripheral T cells, respectively that hnRNPLL targets not only CD45 (Figs. 1 to 3 tates of endogenous hnRNPL from untransduced (Fig. 4A, left). Similarly, expression of CD44 and Fig. 4A) but also STAT5A (Fig. 4B) and di- JSL1 cell lysates bound ~0.09% of all CD45 exons 6 to 14, particularly exons 7 and 10, was verse other proteins involved in cell proliferation, transcripts in the lysates (Fig. 3E) (12). We also increased in conditions lacking hnRNPLL (naïve activation, and cell death (tables S3 and S4). found that endogenous hnRNPL robustly coim- or hnRNPLL-depleted cells) compared with ac- Our data strongly suggest that hnRNPLL is a munoprecipitated with FLAG-hnRNPLL (Fig. tivated cord blood and peripheral T cells (Fig. “master regulator” of alternative splicing in ac-

Fig. 4. HnRNPLL is a glob- A on February 1, 2009 CD45 CD44 al regulator of alternative 9 9 splicing in activated T cells. 8 8 7 7 (A) Exon array analysis of peripheral CD4+ T RNA from activated periph- 6 6 + – 5 5 LL-sh4 eral CD4 Tcells,LL-sh4 4 4 + activated transduced peripheral CD4 3 3 T cells, naïve cord blood 2 2 CD4+ T cells, and activated 1 1 + 0 0 cord blood CD4 T cells. As 10 10 shown for CD45 and CD44 9 9 www.sciencemag.org 8 8 transcripts, low hnRNPLL 7 7 expression (hnRNPLL knock- 6 6 cord blood CD4+ T down or naïve cord blood) 5 5 naive 4 4 activated is associated with increased 3 3 retentionofselectexons. 2 2 (B) Quantitative RT-PCR for 1 Exons 4-6 1 Exons 6-14 Relative Probe Set Expression (Nat. Log) 0 0 retained CD45, CD44, and Individual Probe Sets Individual Probe Sets STAT5A exons in hnRNPLL- Downloaded from depleted peripheral CD4+ T B CD45 Exon 4 CD44 Exon 10 STAT5A Exon 5 cells. (C)HnRNPLL-depletion 16 12 20 + in peripheral CD4 T cells 14 10 12 16 blocks proliferation. (D) 8 10 12 Reduced CD25 expression 8 6 in hnRNPLL-depleted pe- 6 8 + 4 ripheral CD4 T cells. 4 4 2 2 Relative mRNA Level Relative mRNA Level Relative mRNA Level 0 0 0 uninf. shGFP LL-sh4 uninf. shGFP LL-sh4 uninf. shGFP LL-sh4

C D peripheral CD4+ T peripheral CD4+ T shGFP 6 100 LL-sh4 5 1 day post-puro 80 4 6 days post-puro 60 3 40 2 1 20 Relative Cell Counts 0 Relative Cell Numbers 0 LL-sh4 shRFP 100 101 102 103 104 CD25

690 1 AUGUST 2008 VOL 321 SCIENCE www.sciencemag.org REPORTS tivated T cells. HnRNPLL is induced upon T cell may then lead to the assembly of a high-affinity 20. X. M. Ma, S. O. Yoon, C. J. Richardson, K. Julich, J. Blenis, stimulation and as such is likely to correspond to exon repression complex that mediates efficient Cell 133, 303 (2008). 21. P. Valencia, A. P. Dias, R. Reed, Proc. Natl. Acad. the inducible, cycloheximide-sensitive factor postu- CD45RO production as well as global changes in Sci. U.S.A. 105, 3386 (2008). lated to promote the CD45RA > RO transition alternative splicing. 22. S. Guang, A. M. Felthauser, J. E. Mertz, Mol. Cell. Biol. during Tcell activation (10). Depletion of hnRNPLL 25, 6303 (2005). causes an overall shift in patterns of alternative 23. L. H. Hung et al., RNA 14, 284 (2008). References and Notes 24. We thank A. Weiss for the JSL1 cell line, E. Kieff for BJAB splicing, which by and large are in the opposite 1. J. M. Johnson et al., Science 302, 2141 (2003). and BL41 cells, B. North for the lentiviral expression direction from those seen in activated T cells in 2. B. Modrek, C. Lee, Nat. Genet. 30, 13 (2002). plasmid, and the RNAi Consortium at the Broad Institute which hnRNPLL expression is increased. Induction 3. D. L. Black, Annu. Rev. Biochem. 72, 291 (2003). for providing the plasmids to produce the splicing factor of hnRNPLL during the process of T cell activation 4. A. J. Matlin, F. Clark, C. W. Smith, Nat. Rev. Mol. library. We thank C. Moita for assistance in assembling and differentiation may represent a mechanism by Cell Biol. 6, 386 (2005). the library and A. Astier and D. Hafler for aid in 5. M. L. Hermiston, Z. Xu, A. Weiss, Annu. Rev. Immunol. developing the protocol for infecting and silencing genes which the cell can rapidly shift its transcriptome 21, 107 (2003). in primary human T cells. We thank K. Eger and to favor proliferation and inhibit cell death. 6. K. W. Lynch, Nat. Rev. Immunol. 4, 931 (2004). D. Unutmaz for the naïve and activated cord blood There is increasing evidence that splicing and 7. R. J. Salmond, L. McNeill, N. Holmes, D. R. Alexander, samples. We thank J. Burke at Biotique for assistance pre-mRNA processing involve multiprotein com- Int. Immunol. 20, 819 (2008). with the array analysis. This work was supported by NIH 8. N. Holmes, Immunology 117, 145 (2006). grants CA42471, AI40127, and AI44432 to A.R., and plexes in which individual components influence 9. L. McNeill et al., Immunity 27, 425 (2007). U19 AI070352, R21 AI071060, and Defense Advanced distinct aspects of transcript production (20, 21). 10. K. W. Lynch, A. Weiss, Mol. Cell. Biol. 20, 70 (2000). Research Projects Agency grant W81XWH-04-C-0139 to Several proteins, including hnRNPL, PSF (poly- 11. J. Moffat et al., Cell 124, 1283 (2006). N.H. S.O. is supported by postdoctoral training grant pyrimidine tract binding protein–associated splic- 12. Materials and methods are available as supporting T32 HL066987 from the Joint Program in Transfusion material on Science Online. Biology and Medicine, Children’s Hospital, Boston. ing factor), and hnRNPE2, have been shown to 13. I. S. Trowbridge, M. L. Thomas, Annu. Rev. Immunol. 12, bind CD45 transcripts (14, 15). However, none of 85 (1994). Supporting Online Material these factors is induced upon T cell activation 14. C. R. Rothrock, A. E. House, K. W. Lynch, EMBO J. 24, www.sciencemag.org/cgi/content/full/1157610/DC1 (14, 15), and only PSF (splicing factor proline/ 2792 (2005). Materials and Methods 15. A. A. Melton, J. Jackson, J. Wang, K. W. Lynch, Mol. Cell. SOM Text glutamine-rich) was identified in our screen, albeit Biol. 27, 6972 (2007). Figs. S1 to S4 as a weak hit (table S2). Together with its as- 16. I. Shur, D. Ben-Avraham, D. Benayahu, Gene 334, 113 Tables S1 to S4 sociated factors, hnRNPL may influence diverse (2004). References 17. S. A. Tenenbaum, P. J. Lager, C. C. Carson, J. D. Keene, aspects of CD45 pre-mRNA processing, in- on February 1, 2009 Methods 26, 191 (2002). 11 March 2008; accepted 20 June 2008 cluding basal splicing, mRNA stability, and 18. M. J. Moore, P. A. Silver, RNA 14, 197 (2008). Published online 10 July 2008; export (22, 23). After activation, hnRNPLL in- 19. National Center for Biotechnology Information Database; 10.1126/science.1157610 duction and recruitment to CD45 transcripts www.ncbi.nlm.nih.gov/sites/entrez?db=gene Include this information when citing this paper.

predisposition to infection, apparently restricted Pyogenic Bacterial Infections to pyogenic bacterial diseases, particularly IPD, during the first 10 years of life (3). This clinical in Humans with MyD88 Deficiency phenotype is surprising given the central role www.sciencemag.org commonly attributed to both TLRs and IL-1Rs, and the high susceptibility of MyD88-deficient Horst von Bernuth,1,2 Capucine Picard,1,2,3 Zhongbo Jin,4,5 Rungnapa Pankla,4,6 mice to experimental infections with at least 35 Hui Xiao,7 Cheng-Lung Ku,1,2 Maya Chrabieh,1,2 Imen Ben Mustapha,1,2,8 Pegah Ghandil,1,2 pathogens—19 bacteria, seven viruses, five para- Yildiz Camcioglu,9 Júlia Vasconcelos,10 Nicolas Sirvent,11 Margarida Guedes,10 sites, and four fungi (4) (tables S1 to S3). How- Artur Bonito Vitor,12 María José Herrero-Mata,13 Juan Ignacio Aróstegui,14 Carlos Rodrigo,15 ever, IRAK-4–deficient mice have thus far been Laia Alsina,16 Estibaliz Ruiz-Ortiz,13 Manel Juan,14 Claudia Fortuny,16 Jordi Yagüe,14 challenged with only a few pathogens (5). Fibro- Jordi Antón ,16 Mariona Pascal,14 Huey-Hsuan Chang,17 Lucile Janniere,1,2 Yoann Rose,1,2 Downloaded from Ben-Zion Garty,18 Helen Chapel,19 Andrew Issekutz,20 László Maródi,21 22 4 1,2 7 1Human Genetics of Infectious Diseases, INSERM U550, Paris, Carlos Rodriguez-Gallego, Jacques Banchereau, Laurent Abel, Xiaoxia Li, 2 3 4 1,2 1,2,23 France. Paris Descartes University, France. Study Center of Damien Chaussabel, Anne Puel, Jean-Laurent Casanova * Primary Immunodeficiencies, Assistance Publique Hôpitaux de Paris, Necker Hospital, Paris, France. 4Baylor Institute for Im- munology Research, Dallas, TX 75204, USA. 5Baylor University, MyD88 is a key downstream adapter for most Toll-like receptors (TLRs) and interleukin-1 6 7 Waco, TX 76798, USA. Khon Kaen University, Thailand. Cleve- receptors (IL-1Rs). MyD88 deficiency in mice leads to susceptibility to a broad range of land Clinic Foundation, OH 44195, USA. 8Pasteur Institute of pathogens in experimental settings of infection. We describe a distinct situation in a natural Tunis, Tunisia. 9Cerrahpasa Medical School, Istanbul University, setting of human infection. Nine children with autosomal recessive MyD88 deficiency suffered Turkey. 10General Hospital of Santo António, Porto, Portugal. 11 12 from life-threatening, often recurrent pyogenic bacterial infections, including invasive University Hospital Archet 2, Nice, France. Hospital S.João, Porto, Portugal. 13LIRAD–Banco de Sangre y Tejidos, Instituto pneumococcal disease. However, these patients were otherwise healthy, with normal resistance de Investigación Germans Trias i Pujol, Badalona, Barcelona, to other microbes. Their clinical status improved with age, but not due to any cellular Spain. 14Immunology Department, Hospital Clinic, IDIBAPS, leakiness in MyD88 deficiency. The MyD88-dependent TLRs and IL-1Rs are therefore Barcelona, Spain. 15Germans Trias i Pujol Hospital, Barcelona 16 essential for protective immunity to a small number of pyogenic bacteria, but redundant Autonomous University, Spain. Sant Joan de Déu Hospital, Barcelona University, Spain. 17Dendritic Cell Immunobiology, for host defense to most natural infections. Institut Pasteur and INSERM U818, Paris, France. 18Schneider Children’s Medical Center, Petah Tiqva, Israel. 19University of he search for human genetic etiologies disease (IPD) led to the discovery of children Oxford and Oxford Radcliffe Hospital, Oxford, UK. 20Dalhousie 21 of pediatric infectious diseases aims to lacking interleukin-1 receptor–associated kinase University, Halifax, Nova Scotia, Canada. Debrecen Uni- versity, Hungary. 22Gran Canaria Dr Negrin Hospital, Las Palmas Tdecipher the molecular mechanism of 4 (IRAK-4), which is selectively recruited to Toll- de Gran Canaria, Spain. 23Pediatric Hematology-Immunology disease and to reveal the function of immune like receptors (TLRs) and interleukin-1 receptors Unit, Necker Hospital, Paris, France. genes in natura (1). The immunological inves- (IL-1Rs) by MyD88 (2). The patients present *To whom correspondence should be addressed. E-mail: tigation of children with invasive pneumococcal with a life-threatening but narrow and transient [email protected]

www.sciencemag.org SCIENCE VOL 321 1 AUGUST 2008 691