Id2-Mediated Inhibition of E2A Represses Memory CD8 + Differentiation Frederick Masson, Martina Minnich, Moshe Olshansky, Ivan Bilic, Adele M. Mount, Axel Kallies, Terence P. Speed, This information is current as Meinrad Busslinger, Stephen L. Nutt and Gabrielle T. Belz of September 27, 2021. J Immunol 2013; 190:4585-4594; Prepublished online 27 March 2013; doi: 10.4049/jimmunol.1300099

http://www.jimmunol.org/content/190/9/4585 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2013/03/27/jimmunol.130009 Material 9.DC1 http://www.jimmunol.org/ References This article cites 34 articles, 8 of which you can access for free at: http://www.jimmunol.org/content/190/9/4585.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

by guest on September 27, 2021 • No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Author Choice Freely available online through The Journal of Immunology Author Choice option Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2013 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Id2-Mediated Inhibition of E2A Represses Memory CD8+ T Cell Differentiation

Frederick Masson,* Martina Minnich,† Moshe Olshansky,‡ Ivan Bilic,†,1 Adele M. Mount,†,2 Axel Kallies,*,x Terence P. Speed,‡,{ Meinrad Busslinger,† Stephen L. Nutt,*,x and Gabrielle T. Belz*,x

The inhibitor of DNA binding (Id)2 modulates T cell fate decisions, but the molecular mechanism underpinning this regulation is unclear. In this study we show that loss of Id2 cripples effector differentiation and instead programs CD8+ T cells to adopt a memory fate with increased and Tcf7 expression. We demonstrate that Id2 restrains CD8+ T cell memory differentiation by inhibiting E2A-mediated direct activation of Tcf7 and that Id2 expression level mirrors T cell memory recall capacity. As a result of the defective effector differentiation, Id2-deficient CD8+ T cells fail to induce sufficient Tbx21 expression to generate short-lived effector CD8+ T cells. Our findings reveal that the Id2/E2A axis orchestrates T cell differentiation through the Downloaded from induction or repression of downstream transcription factors essential for effector and memory T cell differentiation. The Journal of Immunology, 2013, 190: 4585–4594.

uccessful eradication and protection from reinfection by ory CD8+ T cells poised to rapidly respond to a second encounter

intracellular pathogens such as viruses and bacteria depend with the pathogen. http://www.jimmunol.org/ on the generation of effector and memory CD8+ T cells. Several transcription factors are essential for the specification S + Naive CD8 T cells, on encounter with dendritic cells presenting and differentiation of peripheral T cells following Ag encounter. pathogen Ags, undergo multiple rounds of proliferation and rap- These include the basic helix-loop-helix inhibitor of DNA idly differentiate into short-lived Ag-specific effector T cells with binding (Id)2 and Id3 (1–3), the T-box transcription factors T-bet cytotoxic and cytokine producing capacity. After resolution of the and eomesodermin (Eomes) (4), B –induced matura- infection, effector CD8+ T cell numbers contract significantly, tion 1 (Blimp1) (5–7), Bcl-6 (8), and Tcf7 (also known leaving a small (5–10%) residual population of long-lived mem- as TCF-1) (9). T-bet (encoded by Tbx21) and Blimp1 (encoded by Prdm1) direct the terminal differentiation of effector T cells. by guest on September 27, 2021 *Division of Molecular Immunology, Walter and Eliza Hall Institute of Medical Ablation of either T-bet or Blimp1 results in a complete loss of + low Research, Melbourne, Victoria 3052, Australia; †Research Institute of Molecular KLRG1 IL-7R short-lived effector cells. Furthermore, T-bet ‡ Pathology, Vienna Biocenter, 1030 Vienna, Austria; Division of Bioinformatics, drives effector cell formation in a dose-dependent manner at the Walter and Eliza Hall Institute of Medical Research, Melbourne, Victoria 3052, Australia; xDepartment of Medical Biology, University of Melbourne, Melbourne, expense of the formation of memory cells (10). Maintenance of Victoria 3010, Australia; and {Department of Mathematics and Statistics, University memory cells appears to depend on the induction of Eomes (4). of Melbourne, Melbourne, Victoria 3010, Australia Although our understanding of Eomes regulation is incomplete, 1 Current address: Akron Molecules GmbH, Vienna, Austria. recent evidence suggests that Tcf7, a critical mediator of the Wnt/ 2Current address: CSL Limited, Parkville, VIC, Australia. b-catenin pathway, regulates CD8+ T cell memory by direct Received for publication January 11, 2013. Accepted for publication February 20, binding to the Eomes locus (9). However, the interactions among 2013. these transcription factors, as well as the signals that drive effector This work was supported by grants and fellowships from the National Health and and memory CD8+ T cell fate decisions, are still poorly understood. Medical Research Council of Australia (to G.T.B., S.L.N., and T.P.S.), the Wellcome Trust (to G.T.B.), the Viertel Foundation and the Howard Hughes Medical Institute It is now established that peripheral T cell differentiation is (to G.T.B.), the Australian Research Council (to T.P.S.), and the Australian Research regulated by the activity of Id proteins (11). Four Id proteins (Id1, Council Future Fellowship (to G.T.B., A.K., and S.L.N.). Research of the Busslinger Id2, Id3, Id4), which lack a DNA-binding domain, are capable of Group was supported by Boehringer Ingelheim and the Genome Research in Austria initiative (financed by the Bundesminsterium fu¨r Bildung und Wissenschaft). This binding the E proteins, E2A (possessing two isoforms, E47 and work was also supported by Victorian State Government operational infrastructure E12), HEB, and E2-2. The recent development of reporter mice support and by the Australian Government National Health and Medical Research Council Independent Research Institute Infrastructure Support Scheme. for Id2 and Id3 has identified that these two transcriptional reg- ulators are expressed in a reciprocal manner and regulate distinct Address correspondence and reprint requests to Dr. Gabrielle T. Belz, Division of Molecular Immunology, Walter and Eliza Hall Institute of Medical Research, 1G functions in the differentiation of peripheral T cells (3). For ex- Royal Parade, Parkville, VIC 3052, Australia. E-mail address: [email protected] ample, Id2 is upregulated in effector T cells. In contrast, induction The online version of this article contains supplemental material. of Id3 reflects the emergence of precursors of long-lived memory Abbreviations used in this article: Blimp1, B lymphocyte–induced maturation protein T cells, and repression by Blimp1 limits formation of memory 1; ChIP, immunoprecipitation; DE, differentially expressed; Eomes, eome- CD8+ T cells, thus dispelling the notion that Id2 and Id3 are sodermin; Gzm, granzyme; Id, inhibitor of DNA binding; i.n., intranasal(ly); IRES, internal ribosome entry site; Lm-OVA, Listeria monocytogenes encoding OVA; LN, simply redundant (2, 3). lymph node; MSCV, murine stem cell virus; NP, nucleoprotein; PA, acidic polymer- Id2 has multiple essential functions in the hematopoietic sys- ase; shRNA, short hairpin RNA. tem. It is required for the development of CD103+ and CD8a+ den- This article is distributed under The American Association of Immunologists, Inc., dritic cells, NK cells, a subset of intraepithelial T cells, and lym- Reuse Terms and Conditions for Author Choice articles. phoid tissue inducer cells (12, 13). In Ag-specific CD8+ Tcells,Id2 Copyright Ó 2013 by The American Association of Immunologists, Inc. 0022-1767/13/$16.00 has been proposed to act mainly by regulating their survival during www.jimmunol.org/cgi/doi/10.4049/jimmunol.1300099 4586 Id2 INHIBITS T CELL MEMORY infection (1), but the precise molecular mechanisms downstream BD Pharmingen). Viable cells were analyzed by flow cytometry using of Id2 that determine T cell fate are poorly defined. propidium iodide or Sytox blue exclusion (Invitrogen). Analysis was To understand this pathway in greater depth we generated mice performed on a FACSCanto or a FACS LSRFortessa (BD Biosciences) and data were analyzed using FlowJo flow cytometry analysis software. with a reporter allele encoding GFP under the endogenous Id2 promoter (12) and a conditional allele allowing specific deletion Generation of bone marrow chimeras of Id2 in T cells. This enabled us to examine the cellular and mo- Mixed bone marrow chimeras were established by reconstituting lethally lecular pathway resulting from the loss of Id2 and to explore the irradiated (2 3 0.55 Gy) B6.Ly5.1 or Ly5.1 3 Ly5.2 (F1) recipient mice mechanisms affecting effector and memory T cell fate outcomes with a mixture of Id2fl/flLckCre (3.4 3 106) and wild-type (Ly5.1+, 1.6 3 in an infection setting. We demonstrated that Id2 was essential for 106) T cell–depleted bone marrow cells and allowed to reconstitute for the induction of high levels of Tbx21 and this was required for the 6–8 wk (5, 21). + + generation of short-lived effector CD8 cells. Loss of Id2 in CD8 Microarray analysis T cells impaired effector T cell differentiation and programmed b + + T cells to adopt a memory cell phenotype with increased Eomes D NP366 CD8 T cells were purified by flow cytometric sorting (FAC- SAria flow cytometer; BD Biosciences) from day 9 HKx31-infected PR8- and Tcf7 expression. We also show that induction of Id2 restrains primed Id2gfp/gfp or Id2fl/flLckCre+/Ly5.1+/+ mixed bone marrow chimeras. + CD8 T cell memory differentiation by inhibiting E2A-mediated TotalRNAwaspreparedusingaQiagenRNeasyMiniKit(Qiagen) transactivation of Tcf7 expression and that graded expression of according to the manufacturer’s instructions. DNA microarray analysis of Id2 rather than central or effector memory phenotype correlates expression was performed using Illumina MouseWG-6 v2.0 Ex- with CD8+ T cell memory recall capacity. Overall, we reveal that pression BeadChip together with the manufacturer’s probe annotation (released in November 2008). The Omnibus referenced Id2 is a dose-dependent regulator of T cell differentiation by or- accession nos. for these data are GSE44140 (http://www.ncbi.nlm.nih.gov/ Downloaded from chestrating the induction or repression of downstream transcrip- geo/query/acc.cgi?acc=GSE44140) and GSE44141 (http://www.ncbi.nlm. tion factors critical in effector versus memory differentiation. nih.gov/geo/query/acc.cgi?acc=GSE44141). In vitro T cell culture analysis Materials and Methods CD8+ T cells were enriched from the spleen of Id2fl/flLckCre+ or wild-type Mice 2 mice. CD8a+CD19 T cells were purified either by FACS sorting or by two http://www.jimmunol.org/ Id2gfp/gfp (12), Id2fl/flLckCre (Id2Lck), Tbx212/2 (14), OT-I (15), Flpe (16), rounds of negative selection (MACS; Miltenyi Biotec) to obtain purities + LckCre (17), C57BL/6 (Ly5.2+), B6.SJL-Ptprca Pepcb/BoyJ (B6.Ly5.1, .95%. CD8 T cells were then cultured for 3 d in complete RPMI 1640 + BirA/BirA 3 5 Ly5.1 ), Rosa26 (18), and Ly5.1 3 Ly5.2 (F1) mice were used at medium at 5 10 cells/well in plates coated with anti-CD3 mAb (5 mg/ml) 6–8 wk. Id2fl/fl mice were generated as described (19). All mice were bred together with anti-CD28 mAb (2 mg/ml) and recombinant human IL-2 (100 4 and maintained under specific pathogen-free conditions in accordance with U/ml). Cells were then washed once and cultured (5 3 10 cells/ml) for 2 d, 4 the guidelines of the Walter and Eliza Hall Institute of Medical Research then replated at 6.7 3 10 cells/ml and cultured for an additional 2 d. Animal Ethics Committee. Retroviral constructs Bacterial and viral infections Murine stem cell virus (MSCV)–T-bet–internal ribosome entry site 4 (IRES)–GFP (10) has been previously described and MSCV-Id2-IRES- Mice were anesthetized with methoxyflurane and then inoculated with 10 by guest on September 27, 2021 PFU HKx31 (H3N2) influenza virus. Memory mice were generated by GFP and MSCV-IRES-GFP were a gift from Dr. Manabu Sugai (Kyoto priming with i.p. injection of 107 PFU A/PR/8/34 (PR8) influenza virus University Hospital). Tcfe2a-IRES-GFP and IRES-GFP LMP vectors were (20). At the indicated times, mice were sacrificed, spleens, lymph nodes a gift from Dr. Ross Dickins (Walter and Eliza Hall Institute). (LNs; mediastinal and superficial cervical), and lungs were removed, and single-cell suspension prepared for analysis. For listeria infection, Intracellular flow cytometry staining Listeria monocytogenes recombinant encoding OVA (Lm-OVA) was grown For detection of intracellular granzyme (Gzm) B, T-bet and Eomes staining in brain-heart infusion broth. Bacterial culture samples were grown to mid- was performed directly ex vivo. GzmB was detected using anti-human log phase measured by OD (absorbance at 600 nm) and diluted in PBS for GzmB Ab (GB12; Invitrogen) using a Cytofix/Cytoperm kit (BD Bio- injection. Mice were infected i.v. with 2.5 3 103 rLm-OVA. Injected sciences). Staining for T-bet (4B10; Santa Cruz Biotechnology) and Eomes bacteria numbers were determined by spreading bacterial samples on (Dan11mag; eBioscience) used the Foxp3 kit (eBioscience). brain-heart infusion plates followed by incubation overnight at 37˚C. Chromatin immunoprecipitation-PCR analysis Isolation of CD8+ T cells Chromatin immunoprecipitation (ChIP) was performed with an affinity- + CD8 T cells were enriched from spleen and LNs by generating a single- purified polyclonal rabbit Ab directed against the N-terminal peptide cell suspension and incubating the cells in a mixture of optimally titrated PRRRAANGGLDTQPKKVRKV of the mouse E2A protein. The precip- Abs against Mac-1 (M1/70), Mac-3 (F4/80), Ter-119, GR1 (RB6-8C5), itated DNA was quantified by real-time PCR analysis with SYBR Green, MHC class II (M5/114), and CD4 (GK1.5) for 30 min on ice. The Ab- which was carried out on a MyiQ instrument (Bio-Rad). Primers sequences bound cells were removed using anti-rat IgG Ab-conjugated magnetic are supplied in Supplemental Table I. beads (Dynabeads; Dynal). Mononuclear cells were purified from the lung by generating single-cell Bio-ChIP sequencing suspensions, followed by centrifugation on Histopaque density gradient ∼ 8 Bio/Bio BirA/BirA (1.077 g/ml; Sigma-Aldrich) for 18 min at 2000 rpm at room temperature. Total thymocytes ( 10 ) from Tcfe2a Rosa26 mice were Liver-infiltrating were isolated by centrifugation liver cell used for chromatin precipitation by streptavidin pulldown (Bio-ChIP), as suspension on an 80/40% Percoll (GE Healthcare Biosciences) density recently described in detail (22). gradient for 20 min at 2500 rpm at room temperature. Quantitative RT-PCR Cell surface staining and FACS analysis b + Total RNA was prepared from purified D NP366-specific CD8 T cell Virus-specific CD8+ T cells were detected by staining with PE-coupled populations using an RNeasy Mini Kit (Qiagen). cDNA was synthesized tetrameric H-2b MHC class I complexes loaded with epitopes of influenza from total RNA with oligo(dT) and thermoscript reverse transcriptase b b (Invitrogen). Real-time PCR was performed using the SensiMix SYBR virus nucleoprotein (NP; D NP366–374, H-2D –restricted) or acidic poly- merase (PA; DbPA , H-2Db–restricted) as previously described (20). No-ROX Kit (Bioline). Analyses were done in triplicate and mean nor- 224–233 Hprt To block nonspecific binding of Abs, cells were incubated in rat IgG malized expression was calculated with as the reference gene. (1 mg/ml) together with anti–FcgR mAb (CD16/32, clone 2-4G21) for Retroviral transduction of OT-I TCR transgenic T cells 10 min on ice. Samples were then stained with fluorochrome-labeled Abs against CD8a (53-6.7), CD62L (Mel-14), KLRG1 (2FY), IL-7R (A7R34), Id2fl/flLckCre+ or wild-type CD8+ OT-I T cells were enriched from the Ly5.1 (A20-1.1), Ly5.2 (104), and CD44 (IM7) (all from eBioscience or spleen and cultured in complete RPMI 1640 containing human recombi- The Journal of Immunology 4587 nant human IL-2 (30 U/ml) and 1 mg/ml OVA257–264 peptide for 2 d. Results Retroviral supernatants were then spun for 1 h at 4000 rpm at 4˚C in 12- Loss of Id2 impairs the differentiation of short-lived effector well plates coated with 32 mg/ml RetroNectin. Retrovirus supernatant + was then carefully removed from the plate and OT-I T cells were then CD8 T cells added and cultured with human rIL-2 (30 U/ml) and polybrene (1 mg/ml) To map the expression level and the role of Id2 in different T cell in complete RPMI 1640 for 1 d. For adoptive transfer experiments, OT-I T cells transduced with MSCV-T-bet-IRES-GFP, MSCV-Id2-IRES-GFP, populations, we made use of our previously described reporter or MSCV-IRES-GFP retroviruses were washed and expanded in a 75- allele where an IRES-GFP cassette has been inserted into the 39 cm2 culture flask in the presence of IL-15 (20 ng/ml) for 4 d. GFP+ cells untranslated region of the Id2 gene by homologous recombination were purified by flow cytometric sorting prior to adoptive transfer into (12). To track the induction of Id2, wild-type and Id2 reporter recipient mice that were then infected the following day with Lm-OVA. mice (Id2gfp/gfp) were infected intranasally (i.n.) with influenza A OT-I cells transduced with the short hairpin RNA (shRNA) Tcfe2a- IRES-GFP LMP or IRES-GFP LMP retroviruses were selected in pu- HKx31. Strikingly, Id2-GFP was rapidly upregulated in all b b + romycin at 2 mg/ml for 5–7 d. Live cells were then purified by centri- influenza-specific D NP366- and D PA224-specific CD8 T cells fugation on Histopaque density gradient before quantitative RT-PCR regardless of their phenotype or anatomical location (Fig. 1A, 1B) analysis. To examine how Id2 expression affected the emergence of influenza-specific CD8+ effector and memory T cells, we analyzed Statistical analysis mice that specifically lacked Id2 in T cells through LckCre- Statistical analyses were performed using Prism software. mediated deletion of the entire Id2 coding region (19). To ex- Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 1. Id2-deficient CD8+ T cells do not differentiate into KLRG1+ effector cells. (A) Id2gfp/gfp and wild-type (+/+) mice were infected i.n. with 2 + b + b + + influenza virus HKx31 and analyzed on day 9. (B) Id2-GFP expression in naive (CD44 CD62L ) and D NP366 and D PA224 influenza-specific CD8 b + + Cre+ + T cells from spleen of infected mice. (C) Generation of wild-type and Id2-deficient D NP366-specific CD8 T cells in Ly5.2 Id2Lck :Ly5.1 mixed bone b + marrow chimeric mice 10 d after i.n. influenza infection. Data are representative of 11 mice. (D) Frequency of D NP366-specific CD8 T cells within the Id2fl/flLckCre+ (Ly5.2+) or the wild-type (Ly5.1+) splenic CD8+ T cell compartment from chimeric mice 10 d after HKx31 infection (left panel). Data are the + b + means 6 SEM from three independent experiments (n = 11). Right panel, Relative contribution of Id2-deficient CD8 T cells to the D NP366-CD8 T cell b + + fl/fl Cre+ populations in the indicated tissues (LNs, pooled superficial cervical and mediastinal LNs) expressed as log2 ratio of D NP366 CD8 T cells in Id2 Lck and wild-type compartments. Data were normalized for the proportion of each genotype in the total CD8+ T cells from the spleen and are pooled from two b + + independent experiments (n = 8). (E) Bone marrow chimeric mice were infected with HKx31 and D NP366 influenza-specific CD8 T cells in spleen were analyzed for the indicated markers within the wild-type (Ly5.22) and Id2-deficient (Id2fl/flLckCre+, Ly5.2+) compartments. Data are representative of at least + b + three independent experiments. (F) Frequency of KLRG1 D NP366 CD8 T cells following primary HKx31 influenza infection. Data show the means 6 SEM of KLRG1+ influenza virus-specific cells within the CD8+ T cell compartment in indicated tissues. For (E) and (F), data are pooled from three in- dependent experiments (n = 8). (G and H)OT-IId2fl/flLckCre+ (Ly5.2) and OT-I wild-type (Ly5.1 3 Ly5.2) CD8+ T cells, mixed in a 1:1 ratio, were transferred into wild-type Ly5.1+ mice that were infected i.v. with Lm-OVA 24 h later. (G) Flow cytometric analysis of wild-type (Ly5.1+Ly5.2+ F1) or Id2fl/flLckCre+ (Ly5.2+) OT-I CD8+ T cell population isolated from the spleen 9 d after infection with Lm-OVA. Right panel, Bar graph shows the mean percentage 6 SEM of KLRG1+ wild-type or Id2fl/flLckCre+ OT-I CD8+ T cell populations in the spleen. (H) Bar graph shows the mean 6 SEM of the ratio of wild-type OT-I to Id2fl/flLckCre+ OT-I T cells in total OT-I cells and in memory precursors cells (IL-7RhighKLRG12). Data are pooled from two independent experiments (n = 10). Statistically significant differences were determined using a two-tailed paired Student t test (A–G) or a Mann–Whitney nonparametric test (H). *p , 0.05, **p , 0.01, ***p , 0.001. 4588 Id2 INHIBITS T CELL MEMORY clude secondary effects due to differing resolution of the infection ciency (Fig. 1H). This demonstrated that in the absence of Id2, the in intact mice, Ly5.2+Id2Lck (fl/fl; LckCre+) and Ly5.1+Id2 (+/+) strong inflammatory stimulus of Lm-OVA resulted in the selective mixed bone marrow chimeras were generated and infected with failure of the short-lived effector (IL-7RlowKLRG1+) CD8+ T cell influenza virus (HKx31). This revealed that the proportions of subset to differentiate and/or survive. In contrast, memory pre- virus-specific CD8+ T cells in Id2Lck and wild-type compartments cursor (IL-7RhighKLRG12) CD8+ T cells were unaffected show- were similar in spleen and LNs (Fig. 1C, 1D). However, within ing that Id2 loss does not globally impair virus-specific CD8+ Lck b the lung and liver, the relative frequency of Id2 D NP366- and T cell survival. b + D PA224-specific CD8 T cells was reduced, suggesting a defect in the ability to migrate to and/or survive in nonlymphoid tissues Id2 expression restricts memory recall capacity (Fig. 1D). KLRG1+IL-7R2 short-lived effectors were selectively Id2 deficiency significantly affected effector T cell differentiation, absent from the Id2-deficient CD8+ T cell compartment in ac- but it was not clear whether normal memory was formed when Id2 cordance with earlier observations (Fig. 1E, 1F, Supplemental Fig. was lacking. To understand the effect of Id2 loss on memory, we 1) (3). The unaltered frequency of influenza-specific CD8+ T cells quantitated the proportion of influenza-specific memory CD8+ in Id2-deficient and wild-type compartments within the spleen and T cells within the wild-type and Id2Lck CD8+ T cell compartments LNs was surprising given a previous report that suggested Id2 was of mixed BM chimeras infected 9 wk previously with HKx31 virus critical to control effector CD8+ T cell survival (1). In this earlier (Fig. 2A, 2B). Neither the frequency of virus-specific memory study, Id2-deficient OVA-specific TCR transgenic CD8+ T cells CD8+ T cells (Fig. 2A, 2B) nor the proportion of central memory (OT-I) responding to systemic infection with rLm-OVA were T cells (F. Masson and G.T. Belz, unpublished observations) was highly susceptible to apoptosis mediated by the proapoptotic significantly different from control cells. To accurately assess the Downloaded from molecule Bim. To better understand the factors that underpin the recall potential of Id2-deficient memory T cells, CD8+ T cells were discrepancies between the two studies, we analyzed the quantity enriched from the spleens of HKx31-infected mixed chimeric mice and the quality of the response of Id2Lck (Ly5.2+) and wild-type and adoptively transferred into naive recipients prior to influenza + + + (Ly5.1 Ly5.2 F1) OT-I T cells transferred into Ly5.1 recipients infection (Fig. 2C). Surprisingly, 10 d after infection, we found that following i.v. Lm-OVA infection. Similar to previous results, Id2-deficient CD8+ T cells had significantly increased in proportion Lck Id2 OT-I T cells were outcompeted by wild-type OT-I T cells at compared with wild-type cells (Fig. 2D), indicating that Id2 ex- http://www.jimmunol.org/ 9 d after Lm-OVA infection (Fig. 1G, 1H) (1). However, closer pression restricts memory recall potential of CD8+ Tcells. analysis revealed that following Lm-OVA infection most (∼80%) To understand the role of Id2 in the memory T cell compart- wild-type OT-I T cells formed short-lived effectors whereas ment in more detail, we assessed the expression of Id2-GFP in Ag- memory precursors represented a smaller proportion of the overall specific CD8+ T cells during the memory phase of influenza virus population. This short-lived effector population was absent in the infection. Both effector memory and central memory T cells Id2Lck OT-I CD8+ T cell compartment (Fig. 1G), but notably the expressed a lower level of Id2-GFP than did acutely activated memory precursor compartment was not affected by Id2 defi- effector cells (Fig. 3A). Remarkably, approximately half of the by guest on September 27, 2021

FIGURE 2. Id2-deficient memory CD8+ T cells exhibit increased recall capacity. (A and B) Mixed bone marrow chimeras reconstituted with Id2fl/flLckCre+ (Ly5.2+) and wild-type (Ly5.1+) bone marrow were analyzed 9 wk after i.n. HKx31 infection. (A) Representative flow cytometric profiles and (B) cu- b + fl/fl Cre+ + + + mulative analysis of the persistence of D NP366-specific CD8 T cells within the Id2 Lck (Ly5.2 ) and the wild-type (Ly5.1 ) CD8 T cell com- partments. Individual mice in (B) are shown by an open circle (wild-type) or filled square (Id2fl/flLckCre+). Data are pooled from three independent experiments (n = 13). Statistically significant differences were determined using a paired two-tailed Student t test. (C and D) Analysis of recall responses of Id2-deficient (Ly5.2+) and wild-type (Ly5.1+) virus-specific CD8+ T cells. (C) Schematic diagram showing the experimental approach for analyses of fl/fl Cre+ + b + memory T cell recall responses. (D) Relative contribution of Id2 Lck CD8 T cells to the D NP366-specific CD8 T cell population in spleen, LN, and b + + lung. Data were normalized for the proportion of each genotype represented in the input D NP366 CD8 T cell population and are expressed as log2 of the fl/fl Cre+ + b + + b + ratio of Id2 Lck CD8 D NP366 /wild-type CD8 D NP366 T cells. Data are pooled from 14 to 18 animals from three independent experiments. Statistically significant differences of the test group to 0 were determined using an unpaired two-tailed Student t test. **p , 0.01, ***p , 0.001. The Journal of Immunology 4589 Downloaded from FIGURE 3. Inverse correlation between Id2-GFP expression and memory CD8+ T cell recall capacity. (A) Id2gfp/gfp mice were infected with HKx31 influenza virus and analyzed 9–10 d (acute) or 12–17 wk (memory) after infection. GFP expression is shown at various stages of virus specific CD8+ T cell differentiation (naive, CD62L+CD442; effector [acute time point]/effector memory [memory time point], CD62L2CD44+; central memory, CD62L+CD44+). Data are pooled from two independent experiments (n = 7) and show the mean Id2-GFP mean fluorescence intensity (MFI) 6 SEM. (B) HKx31-infected Id2gfp/gfp and wild-type mice were analyzed at 17 wk for their expression of Id2-GFP expression in naive (CD442CD62L+), central memory (Tcm; CD44+ + + 2 b b + gfp/gfp CD62L ), and effector memory (Tem; CD44 CD62L )inDNP366- and D PA224-specific CD8 T cells. (C and D) Id2 mice were primed i.p. with + PR8 virus, and 4 wk later mice were challenged i.n. with HKx31 influenza virus. Seven to 8 mo after challenge, CD8 T cells were analyzed. (C) Schematic http://www.jimmunol.org/ diagram outlining the isolation and analysis of influenza-specific memory CD8+ T cells expressing different levels of Id2-GFP. (D) Enumeration of the low int b + + b + + + 2 recall capacity of Id2-GFP and Id2-GFP D NP366 CD8 T cells following HKx31 infection as described in (C). D NP366 CD8 (Ly5.2 , Ly5.1 ) isolated from the indicated organs were enumerated by flow cytometry. Data show the means 6 SEM of virus-specific CD8+ T cells in the spleen, lungs, and mediastinal LNs pooled from three experiments (n = 10). Statistically significant differences were determined using a two-way ANOVA. **p , 0.01. influenza-specific central memory CD8+ T cells had a reduced Id2 To extend this analysis, we examined how loss of Id2 affected the expression comparable with naive CD8+ T cells, suggesting that induction of the effector transcriptional program. Remarkably, the low Id2 expression level is linked to the reported higher recall expression of the cytolytic molecules. GzmA and GzmB were

b + by guest on September 27, 2021 capacity of central memory T cells (Fig. 3B). To directly test significantly reduced in Id2-deficient D NP366-specific CD8 whether Id2 expression determines the recall potential of CD8+ T cells (Fig. 4A, 4B). Id2-deficient cells also displayed an altered memory T cells, influenza-specific effector memory and central expression of integrins and chemokine receptors such as CD49a, memory T cells from Id2gfp/gfp mice were partitioned into cells CD103, CX3CR1, and CCR7 likely to impair T cell migration expressing intermediate (Id2-GFPint) or low (Id2-GFPlow) levels of (22–24) (Fig. 4A). This is consistent with our observation that Id2- Id2-GFP (Fig. 3C). Following adoptive transfer of each memory deficient CD8+ T cell frequency is reduced within nonlymphoid T cell subset into congenically marked (Ly5.1+) recipients, anal- tissues (Fig. 1D). Overall, our data show that Id2-deficient virus- ysis of mice infected with HKx31 revealed that regardless of specific CD8+ T cells exhibit an impaired transcriptional program b + whether transferred D NP366-specific CD8 T cells were derived of effector differentiation and consequently fail to appropriately from the effector or central memory T cell compartment, the differentiate into short-lived effector CD8+ T cells. highest proliferative potential was maintained within Id2-GFPlow Having established that Id2 was important for memory CD8+ populations (Fig. 3D). These data suggest that Id2 limits the recall T cells, we speculated that distinct levels of Id2 were deterministic potential of memory CD8+ T cells in a dose-dependent manner. in the transcriptional program of Ag-specific CD8+ T cells. To test b + + this hypothesis, we subjected D NP366-specific effector CD8 Id2-deficient CD8 T cells adopt a gene expression signature T cells purified according to their differential expression of Id2- characteristic of memory precursor cells GFP (Id2-GFPint and Id2-GFPhigh) to microarray analysis and To gain insight into how Id2 regulates the lineage choice between compared their gene expression profiles to the 233 DE + b effector and memory CD8 T cells, we compared the gene ex- identified by comparing Id2-deficient and wild-type D NP366- Lck b + + pression profile of Id2 and wild-type D NP366-specific CD8 specific CD8 T cells (Supplemental Fig. 2). This analysis T cells purified from spleen of Id2Lck:Ly5.1 mixed BM chimeras revealed that most of the DE genes were strongly dependent on 9 d after HKx31 influenza virus infection by microarray analysis. Id2 expression levels. Indeed, 76% of DE genes found to be up- To ensure that meaningful comparisons could be made between regulated in Id2-deficient cells were also upregulated in Id2- wild-type and Id2Lck cells, KLRG1+ cells were excluded during GFPint cells compared with Id2-GFPhigh cells. Similarly, 83% of FACS sorting. Two hundred thirty-three differentially expressed the DE genes observed to be downregulated in absence of Id2 (DE) genes were identified (Fold change $ 2 and p value # 0.05) were downmodulated in Id2-GFPint cells (Supplemental Fig. 2B). (Supplemental Table II). Id2-deficient CD8+ T cells expressed Furthermore, expression of Eomes, Id3,andTcf7 mRNA was significantly higher levels of key transcriptional regulators im- higher on Id2-GFPint cells compared with Id2-GFPhigh T cells, portant for memory T cell differentiation including Id3, Tcf7 and whereas GzmB expression was significantly reduced in Id2-GFPint Eomes (Fig. 4). In contrast, the expression of T-bet and Blimp1, cells T cells (Supplemental Fig. 2B, 2C). However, despite an two important regulators of effector T cell differentiation, were increased expression of genes involved in memory formation, Id2- markedly reduced in absence of Id2 (Fig. 4B, 4C). GFPint virus-specific CD8+ T cells did not exhibit increased long- 4590 Id2 INHIBITS T CELL MEMORY Downloaded from

FIGURE 4. Id2-deficient influenza-specific CD8+ T cells express memory signature genes. (A) Whole genome analyses of wild-type and Id2fl/flLckCre+ b + D NP366-specific CD8 T cells isolated from spleen of PR8-primed/HKx31-infected mice. Graphs show the mean fold difference of 37 (of 233) DE genes fl/fl Cre+ b + between Id2 Lck and wild-type D NP366-specific CD8 T cells. Genes were selected based on their role in T cell function, migration, and differ- entiation. Data are pooled from three biologically independent samples. (B) Flow cytometric analysis of T-bet and GzmB was compared for naive CD442 + fl/fl Cre+ + + b + CD8 T cells (dotted black line) and Id2 Lck (Ly5.2 ; solid red line) and wild-type (Ly5.1 ; solid blue line) D NP366-specific CD8 T cells from PR8- primed mixed bone marrow chimeras 9 d after HKx31 infection. Bar graph shows means 6 SEM of the mean fluorescence intensity (MFI) of staining for http://www.jimmunol.org/ b + + 2 fl/fl Cre+ + intracellular protein expression within wild-type D NP366 KLRG1 , KLRG1 ,orId2 Lck CD8 T cells from PR8-primed HKx31 infected mice. Data are pooled from 16 to 20 chimeric mice from three independent experiments. (C) Quantitative analysis of mRNA expression by RT-PCR within wild-type b + + 2 fl/fl Cre+ + 2 + + D NP366 KLRG1 , KLRG1 ,orId2 Lck CD8 T cells from PR8-primed HKx31-infected mice (day 9) and naive splenic CD44 CD62L CD8 T cells. Data show the means 6 SEM of gene expression relative to that found in naive CD8+ T cells expression from three independent samples. Sta- tistically significant differences were determined using a paired (B) or an unpaired (C) two-tailed Student t test. *p , 0.05, **p , 0.01, ***p , 0.001. term survival compared with Id2-GFPhigh virus-specific CD8+ To confirm the involvement of E2A, we then wanted to deter- T cells (Supplemental Fig. 2D, 2E). This indicated that the sur- mine the proportion of genes having E2A occupancy at promoter/ by guest on September 27, 2021 vival of memory cells is not dictated by the level of Id2 expres- enhancer distance among the list of 233 DE genes in the absence sion, consistent with our previous results (Fig. 2B). of Id2. We took advantage of Tcfe2aBio/Bio mice (I. Bilic and Overall, our data demonstrate that the transcriptional program M. Busslinger, unpublished observations), which express an E2A- of CD8+ T cell differentiation is exquisitely sensitive to the con- Bio protein containing a C-terminal biotin acceptor sequence that is centration of Id2. efficiently biotinylated in vivo by coexpression of the Escherichia coli biotin ligase BirA from the Rosa26BirA allele. We then deter- Id2 controls effector differentiation by inhibiting E2A mined the genome-wide pattern of E2A binding in total thymocytes Id2 is a key negative regulator of the E protein transcription factor from Tcfe2aBio/Bio mice, in which Id2 expression is low (27) and family but it is less clear which specific E protein Id2 might act hence allows high E2A binding to its target genes. Using a strin- on during T cell differentiation. We hypothesized that Id2 controls gent p value of ,10210 for peak calling, we detected 4337 E2A- CD8+ T cell differentiation by limiting the transcriptional activity binding regions, which defined 2541 E2A target genes in total of E2A. To investigate this question, we transduced wild-type and thymocytes. We next used these E2A target genes to cross- Id2Lck OT-I T cells with retroviruses encoding either a shRNA reference our microarray data comparing Id2-deficient and wild- against Tcfe2a (encoding E2A) or a control hairpin and then an- type virus-specific CD8+ T cells. Strikingly, the results from this alyzed the expression of key target genes identified as DE in our analysis showed that among our list of genes found to be upreg- previous microarray analysis. The Tcfe2a shRNA induced an 80% ulated in absence of Id2, the proportion of those with an E2A- reduction of Tcfe2a expression, and this reduction was not com- binding site is significantly higher than among all genes of the pensated by an increased of the other E protein family members genome, consistent with the transactivating function of E2A (Fig. Tcf12 (encoding HEB) and Tcf4 (encoding E2.2) expression (Fig. 5B). In contrast, there is no significant difference of enrichment of 5A). Remarkably, silencing of Tcfe2a in Id2Lck OT-I T cells and, to genes with E2A occupancy among the genes found to be down- a lesser extent, in wild-type OT-I T cells resulted in a decrease in regulated in absence of Id2 compared with the whole genome, the expression of several genes important for the development suggesting that they are not directly regulated by E2A (Fig. 5B). and/or the maintenance of memory T cells, such as Tcf7, Id3, and Consistent with the Tcfe2a silencing data (Fig. 5A), our Bio-ChIP Socs3 (2, 3, 9, 25) (Fig. 5A). In contrast, Id2Lck OT-I T cells sequencing analysis identified several E2A-binding sites at the loci of transduced with the Tcf2e2a shRNA exhibited an increased ex- key genes involved in memory T cell differentiation or function such pression of genes encoding the effector molecules GzmB and as Tcf7, Id3, Socs3,orCcr7, whereas no binding sites were detected GzmK compared with cells transduced with the control shRNA at the loci of genes associated with an effector phenotype such as retrovirus (Fig. 5A). Of note, the expression of Tbx21 or Prdm1 Tbx21, Gzmk,andGzmb (F. Masson and G.T. Belz, unpublished was not affected by Tcfe2a knockdown in vitro, suggesting that observations). Overall, our results suggest that Id2 restrains memory they are not direct targets of E2A. T cell development by inhibiting E2A transcriptional activity. The Journal of Immunology 4591

FIGURE 5. Id2 controls effector and memory CD8+ T cell differentiation by inhibiting E2A transcriptional activity. (A) Wild-type and Id2fl/flLckCre+ OT-I cells were transduced with shRNA-tcfe2a-GFP and shRNA-control-GFP retroviruses. GFP+ cells (95% positive) were subjected to quantitative RT- PCR for the indicated genes. Graphs show the means 6 SEM of the experimental triplicates of the expression relative to Hprt and are representative of two independent experiments. (B) Bio-ChIP sequencing data of total thymocytes from Tcfe2aBio/Bio Rosa26BirA/BirA mice. Bar graph shows the proportion of Downloaded from genes with E2A-binding sites in the whole genome or among the 233 DE (upregulated or downregulated) genes in Id2-deficient virus-specific CD8+ T cells. Statistically significant differences were determined using a Fisher exact test. ***p , 0.001.

E2A regulates Tcf7 expression in peripheral CD8+ T cells To test whether downregulation of T-bet expression in Id2- + + deficient CD8 T cells was directly responsible for the loss of http://www.jimmunol.org/ Tcf7, a critical regulator of memory CD8 T cell differentiation, + Lck + persistence, and recall potential (9), was significantly upregulated effector CD8 T cells, we transduced Id2 OT-I CD8 T cells in Id2-deficient CD8+ T cells (Fig. 4C) and correlated with the dose with retroviruses encoding T-bet or Id2 or with a control retro- virus. In vitro–activated transduced Id2Lck and wild-type OT-I of Id2 (Fig. 4D). Tcf7 expression was also downregulated in Id2- + deficient T cells in which E2A was silenced (Fig. 5A). Our Bio- T cells were adoptively transferred into Ly5.1 recipient mice. Mice were infected with Lm-OVA 24 h later. After 8 d, the for- ChIP sequencing analysis in total thymocytes identified several + E2A-binding sites at the Tcf7 locus (Fig. 6A). However, because mation of effector CD8 T cells from transduced OT-I cells was Id2 is not expressed in double-positive thymocytes (27) whereas it is expressed at low level in naive CD8+ T cells (Fig. 1B) and at a high level in effector CD8+ T cells (Fig 1B), the binding of E2A to by guest on September 27, 2021 the Tcf7 locus is expected to be inversely correlated to the level of Id2 expression in these different T cell subsets. Therefore, we then investigated whether Tcf7 was also a direct target of E2A in pe- ripheral CD8+ T cells. Using conventional ChIP-quantitative PCR analysis, we tested whether E2A could bind to these sequences also in naive and in vitro–activated CD8+ T cells. In naive CD8+ T cells, E2A no longer bound to site 2 and interacted weakly with sites 1, 3, and 5 in contrast to thymocytes (Fig. 6B). However, E2A efficiently bound to site 4 (located 33 kb upstream of the Tcf7 transcription starting start), suggesting that E2A activates the Tcf7 gene through this upstream enhancer in naive CD8+ T cells. Critically, the in- teraction of E2A with site 4 and the weaker binding sites was lost following activation and concurrent induction of Id2 expression (Fig. 6B), suggesting that E2A directly regulated Tcf7 expression in a manner that was inhibited in the presence of Id2. + Id2-dependent control of T-bet expression is required for the FIGURE 6. E2A binding to the Tcf7 locus of peripheral CD8 T cells is A differentiation of short-lived effector cells lost following effector differentiation in vitro. ( ) Identification of E2A- binding regions at the Tcf7 locus by ChIP sequencing of total thymocytes. T-bet is a key transcriptional regulator of effector differentiation Arrows indicate the sites tested for E2A binding in (B). (B) ChIP analysis that is downregulated in the absence of Id2 (Fig. 4B). Because T- of E2A binding in naive and in vitro activated effector CD8+ T cells. Input bet promotes short-lived effector cell differentiation in a dose- and precipitated DNA were quantified by real-time PCR with primer pairs dependent manner (10), we speculated that the reduction of T- amplifying the indicated E2A-binding regions of the Tcf7 locus [shown in bet expression observed in Id2Lck T cells might be responsible for (A)] and a geneless control region of mouse 1 (Supplemental the observed defect in short-lived effector CD8+ T cell formation. Table II). The relative enrichment of E2A binding was determined by dividing the percentage of precipitated DNA at the E2A-binding site In line with this hypothesis, analysis of mixed bone marrow chi- (ChIP/input) by the percentage of precipitated DNA at the geneless control + Tbx21+/2 + meras created using wild-type (Ly5.1 )and (Ly5.2 ) region of mouse chromosome 1 (ChIP/input). The relative enrichment of heterozygous bone marrow revealed that a 2-fold reduction of E2A binding is shown as means 6 SEM of three independent ChIP Lck T-bet (Fig. 7A), similar to that observed in Id2 T cells (Fig. 4B), experiments. No enrichment is indicated by a dashed line. Statistically was sufficient to impair the generation of short-lived effector significant differences were determined using an unpaired two-tailed T cells after influenza infection (Fig. 7B). Student t test. *p , 0.05, **p , 0.01, ***p , 0.001. 4592 Id2 INHIBITS T CELL MEMORY Downloaded from

FIGURE 7. Enforced expression of Tbx21 rescues differentiation of Id2-deficient short-lived effector cells. (A and B) PR8-primed mixed bone marrow chimeras reconstituted with a mix of Tbx21 heterozygote (+/2, Ly5.2) and wild-type (+/+, Ly5.1) bone marrow cells were analyzed 9 d after HKx31 i.n. infection. (A) Histogram (left panel) and mean fluorescence intensitiy (MFI) (right panel) of expression of T-bet in naive (CD442, black line) or activated DbNP-specific CD8+ T cells gated on wild-type (Ly5.1+, blue line) or Tbx21+/2 (Ly5.2+, red line) CD8+ T cell compartments from mixed bone marrow chimeras. Data show the means 6 SEM (n = 11 animals pooled from three independent experiments). (B) Representative contour plots showing IL-7R and http://www.jimmunol.org/ b + KLRG1 expression in D NP366-specific CD8 T cells. Statistically significant differences were determined using a paired two-tailed Student t test. (C–F) Wild-type (Ly5.1+) and Id2fl/flLckCre+ (Ly5.2+) OT-I cells were transduced with control-GFP, Id2-GFP, or T-bet-GFP retroviruses. One week after activation, GFP+ cells were transferred into recipient mice that were then infected with Lm-OVA. Eight days after infection, mice were sacrificed and OT-I T cell response was analyzed in the spleen. (C) Flow cytometric analysis of KLRG1 and GFP expression on transduced GFP+Id2fl/flLckCre+ (Ly5.2+) OT-I T cells or wild-type (Ly5.1+) OT-I T cells 8 d after infection with Lm-OVA. Data are pooled from five to six animals for each condition from two independent experiments. (D) Analysis of the frequency of KLRG1+ cells within the GFP+Id2fl/flLckCre+ or wild-type OT-I T cells that had been transduced with the indicated vectors. Data show the mean percentage 6 SEM for spleen with at least five mice for each group pooled from two independent experiments. (E) Bar graph shows mean 6 SEM of the MFI of staining for intracellular GzmB expression within GFP+Id2fl/flLckCre+ or wild-type OT-I T cells that had been transduced with the indicated vectors. Data are pooled from five to six mice from two independent experiments. (F) Bar graph shows the frequency of OT-I

Ly5.2+CD8+ T cells transduced with the indicated vectors among the total CD8+ T cell population from the spleen 8 d after Lm-OVA infection and is by guest on September 27, 2021 representative of two experiments with three mice per group in each experiments. Statistically significant differences were determined using an unpaired two-tailed Student t test. *p , 0.05, **p , 0.01, ***p , 0.001. (G) Quantitative analysis of mRNA expression by RT-PCR within GFP+ wild-type or Id2fl/fl LckCre+ OT-I T cells that had been transduced with the indicated vectors. Bar graph represents the means 6 SEM of the experimental triplicates of the expression relative to Hprt. Data are representative of two independent experiments. evaluated (Fig. 7C). Re-expression of Id2 into Id2-deficient OT-I overall survival of virus-specific CD8+ T cell during influenza in- cells rescued the development of KLRG1+ effector cells. Trans- fection and this has allowed us to dissociate the differentiation duction of Id2-deficient OT-I T cells with the T-bet retrovirus defect from the survival defect reported in the former studies. Our resulted in the rescue of T-bet expression to a level similar to Id2- data support an alternate model in which Id2 can act as a rheostat deficient CD8+ T cells transduced with the Id2 retrovirus to directly regulate the differentiation of naive CD8+ T cells into (F. Masson and G.T. Belz, unpublished observations). Strikingly, effector or memory T cells. We demonstrated that Id2 limits Id2-deficient OT-I cells overexpressing T-bet were also able to memory CD8+ T cell formation and recall potential in a dose- form KLRG1+ effector T cells in similar frequency (Fig. 7C, 7D). dependent manner by inhibiting E2A transcriptional activity. In Additionally, T-bet overexpression partly rescued GzmB expres- particular, Id2 inhibits E2A-mediated transactivation of Tcf7. Fi- sion in vivo and Prdm1 and GzmK mRNA expression after in vitro nally, we showed that Id2 is required for the optimal induction of culture in the presence of IL-2 (Fig. 7E, 7G). T-bet overexpressing T-bet and thereby for the differentiation of KLRG1+ short-lived OT-I cells, however, did not expand as well as OT-I cells trans- effector cells. duced with Id2 or the control vector (Fig. 7F), consistent with Id2 deficiency resulted in a profound alteration of the tran- the observation that T-bet overexpression increase activation- scriptional program that drives the differentiation of virus-specific induced cell death in CD8+ T cells (28). Collectively, our data CD8+ T cells. Id2-deficient cells exhibit an increased expression indicate that Id2 is required to induce sufficient Tbx21 expression in multiple genes involved in memory T cell differentiation and to generate short-lived effector CD8+ T cells. function (Tcf7, Id3, Socs3, Eomes, Il2, Ccr7) paralleled by a concurrent decrease in the expression of genes associated with Discussion an effector phenotype (Tbx21, Prdm1, Il2ra, Grzma, Gzmb, Gzmk, Id2 is a key regulator of CD8+ T cell–mediated immunity, but there Cd103, Cx3cr1, Cd49a). Remarkably, the comparison of our is a limited understanding of the molecular mechanisms by which microarray data analysis with E2A CHIP sequencing data revealed Id2 exerts its effects influencing effector and memory T cell fate that there was a significant enrichment for genes with E2A- decisions. Early studies suggested that the main action of Id2 is to binding sites among the genes upregulated in absence of Id2 ensure the survival of effector CD8+ T cells following activation compared with the DE genes that were downregulated. This (1). In contrast, our study shows the absence of a major defect in the suggests that Id2 represses the memory program of CD8+ T cell The Journal of Immunology 4593 differentiation by inhibiting the E2A-mediated transactivation of regulated in absence of Id2. In particular, the expression of T-bet, several key genes involved in the process of memory differenti- a key transcriptional regulator of effector T cell differentiation, was ation such as Tcf7. also affected. Our results suggest that T-bet expression is indirectly Tcf7 is a key mediator of the Wnt/b-catenin signaling pathway regulated by E2A because no E protein binding sites were detected essential for the generation and the persistence of memory T cells at its locus using in silico bioinformatical analysis, or ChIP se- through its control of Eomes expression, as well as for the recall quencing in whole thymocytes (F. Masson and G.T. Belz, unpub- potential of memory CD8+ T cells (9). Tcf7 is highly expressed in lished observations), or in the A12 T cell line (34). It is therefore naive CD8+ T cells and is downregulated in effector CD8+ T cells. likely that the decrease in the expression of this key transcriptional We showed that both Tcf7 and Eomes were expressed at higher regulator is a consequence of the altered differentiation of Id2- levels in Id2-deficient effector CD8+ T cells compared with wild- deficient CD8+ T cells toward the memory phenotype. Surpris- type effector T cells. Using E2A ChIP analyses, we discovered ingly, E2A silencing did not affect Tbx21 expression in vitro (Fig. that E2A directly binds to the Tcf7 locus and this binding is lost 5A), suggesting that either the level of E2A silencing was insuf- following activation concurrent with the induction of Id2. Finally, ficient to induce a significant change of T-bet expression or that silencing of Tcf2ea in Id2-deficient CD8+ T cells resulted in the other E protein family members HEB and E2-2 may compensate downregulation of Tcf7 expression. Taken together, these data for the loss of E2A in the regulation of T-bet expression. Further suggest that a major pathway by which Id2 restrains the formation analyses showed that decreased expression of T-bet was re- of memory precursor cells is by inhibiting the E2A-dependent sponsible for the loss of KLRG1+ cells in Id2-deficient CD8+ activation of Tcf7 in effector CD8+ T cells. T cells. In support of this finding, haploinsufficiency of Tbx21 (a

Recently it has been proposed that the repression of Id3 by decrease in T-bet expression comparable to the downregulation Downloaded from Blimp1 promoted the development of effector T cells at the induced by Id2 ablation) was sufficient to block short-lived CD8+ expense of memory cell differentiation (2). In our studies, in the effector T cell differentiation (Fig. 7B) (10), and enforced ex- absence of Id2, Id3 was strongly upregulated (Figs. 2D, 5A) pression of T-bet in Id2-deficient CD8+ T cells rescued short-lived whereas silencing of Tcfe2a in Id2-deficient CD8+ T cells reversed effector T cell differentiation (Fig. 7C, 7D). this upregulation of Id3 in Id2-deficient CD8+ T cells in vitro In conclusion, our study has identified Id2 as a critical dosage-

without affecting Blimp1 expression (Fig. 5A). From our E2A dependent regulator of E protein transcriptional activity, in par- http://www.jimmunol.org/ ChIP sequencing data, we identified several binding sites at the ticular E2A, which determines whether a naive CD8+ T cell will Id3 locus (F. Masson and G.T. Belz, unpublished observations), commit to effector or memory differentiation. We propose a model confirming that Id3 is a direct target of E2A. It has also been in which Id2 promotes effector differentiation by restraining the reported that Id3-deficient CD8+ T cells did not show any re- E2A-mediated transactivation of several key genes essential for duction in memory precursor cell frequency (or increase in the memory T cell development and function. frequency of short-lived effector T cells) (2). Thus, repression of Id3 does not appear to influence the lineage choice decision between Acknowledgments effector and memory cells at the acute time point of an infection, We thank Annie Xin, Wei Shi, Malou Zuidscherwoude, Philippe Bouillet, by guest on September 27, 2021 although it does affect the maintenance of memory cells in the long Marc Pellegrini, Lynn Corcoran, Ross Dickins, Manabu Sugai, Simon Pres- term. From our studies, we propose that Id3 expression is directly ton, and Mary Camilleri for provision of reagents, advice, and technical influenced by Id2 expression dynamics, which in turn regulates assistance, and Markus Jaritz for bioinformatic analysis. the E2A-dependent transcriptional activation of the Id3 locus. Thus, Id2 rather than Id3 expression level delineates the effector Disclosures and memory phenotype during the acute phase of an infection. The authors have no financial conflicts of interest. Id2 and Id3 are both known to interact with the same basic helix- loop-helix partners (11, 29), but it is not clear whether they bind different E proteins with a similar affinity. Indeed, Id2 and Id3 References have clearly different functions in effector and memory cells. This 1.Cannarile,M.A.,N.A.Lind,R.Rivera,A.D.Sheridan,K.A.Camfield, B. B. Wu, K. P. Cheung, Z. Ding, and A. W. Goldrath. 2006. Transcriptional is exemplified by the loss of long-term memory cells in absence of + regulator Id2 mediates CD8 T cell immunity. Nat. Immunol. 7: 1317– Id3 as opposed to the loss of short-lived effector cells observed in 1325. the absence of Id2 (1, 3). We have shown in this study that re- 2. Ji, Y., Z. Pos, M. Rao, C. A. Klebanoff, Z. Yu, M. Sukumar, R. N. Reger, D. C. Palmer, Z. A. Borman, P. Muranski, et al. 2011. Repression of the DNA- duction or loss of Id2 expression in endogenous virus-specific binding inhibitor Id3 by Blimp-1 limits the formation of memory CD8+ T cells. + memory CD8 T cells leads to enhanced recall responses. In Nat. Immunol. 12: 1230–1237. contrast, Id3 overexpression improved the magnitude of the recall 3. Yang, C. Y., J. A. Best, J. Knell, E. Yang, A. D. Sheridan, A. K. Jesionek, + H. S. Li, R. R. Rivera, K. C. Lind, L. M. D’Cruz, et al. 2011. The transcriptional response of memory CD8 T cells (2). In this latter setting, it is regulators Id2 and Id3 control the formation of distinct memory CD8+ Tcell not clear whether this represents the physiological action of Id3. subsets. Nat. Immunol. 12: 1221–1229. Given the contrasting roles of Id3 and Id2 in T cell differentiation, 4. Intlekofer, A. M., N. Takemoto, E. J. Wherry, S. A. Longworth, J. T. Northrup, V. R. Palanivel, A. C. Mullen, C. R. Gasink, S. M. Kaech, J. D. Miller, et al. it will be crucial to determine with which E protein partners each 2005. Effector and memory CD8+ T cell fate coupled by T-bet and eomeso- Id protein interacts when expressed at physiological levels. This dermin. Nat. Immunol. 6: 1236–1244. 5. Kallies, A., A. Xin, G. T. Belz, and S. L. Nutt. 2009. Blimp-1 transcription factor partner could be E2A or another member of the E protein family is required for the differentiation of effector CD8+ T cells and memory (HEB and E2-2). Indeed, HEB has been recently reported to responses. Immunity 31: 283–295. collaborate with E2A in the generation of memory cells (30). 6. Shin, H., S. D. Blackburn, A. M. Intlekofer, C. Kao, J. M. Angelosanto, S. L. Reiner, and E. J. Wherry. 2009. A role for the transcriptional However, it may also be another protein unrelated to the E protein Blimp-1 in CD8+ T cell exhaustion during chronic viral infection. Immunity 31: family. Indeed, even though E proteins have been described as the 309–320. main heterodimerizing partners of Id proteins, other proteins such 7. Rutishauser, R. L., G. A. Martins, S. Kalachikov, A. Chandele, I. A. Parish, E. Meffre, J. Jacob, K. Calame, and S. M. Kaech. 2009. Transcriptional repressor as the myogenic factors PU.1 and Rb have been described to in- Blimp-1 promotes CD8+ T cell terminal differentiation and represses the ac- teract with Id2 (31–33). quisition of central memory T cell properties. Immunity 31: 296–308. 8. Ichii, H., A. Sakamoto, M. Hatano, S. Okada, H. Toyama, S. Taki, M. Arima, Our microarray data analysis has shown that many genes in- Y. Kuroda, and T. Tokuhisa. 2002. Role for Bcl-6 in the generation and main- volved in effector differentiation and/or functions were down- tenance of memory CD8+ T cells. Nat. Immunol. 3: 558–563. 4594 Id2 INHIBITS T CELL MEMORY

9. Zhou, X., S. Yu, D. M. Zhao, J. T. Harty, V. P. Badovinac, and H. H. Xue. 2010. 22. Ebert, A., S. McManus, H. Tagoh, J. Medvedovic, G. Salvagiotto, Differentiation and persistence of memory CD8+ T cells depend on T cell factor M. Novatchkova, I. Tamir, A. Sommer, M. Jaritz, and M. Busslinger. 2011. The 1. Immunity 33: 229–240. distal VH gene cluster of the Igh locus contains distinct regulatory elements with 10. Joshi, N. S., W. Cui, A. Chandele, H. K. Lee, D. R. Urso, J. Hagman, L. Gapin, Pax5 transcription factor-dependent activity in pro-B cells. Immunity 34: 175– and S. M. Kaech. 2007. Inflammation directs memory precursor and short-lived 187. effector CD8+ T cell fates via the graded expression of T-bet transcription factor. 23. Chapman, T. J., and D. J. Topham. 2010. Identification of a unique population of Immunity 27: 281–295. tissue-memory CD4+ T cells in the airways after influenza infection that is de- 11. Engel, I., and C. Murre. 2001. The function of E- and Id proteins in lymphocyte pendent on the integrin VLA-1. J. Immunol. 184: 3841–3849. development. Nat. Rev. Immunol. 1: 193–199. 24. Masson, F., T. Calzascia, W. Di Berardino-Besson, N. de Tribolet, P. Y. Dietrich, 12. Jackson, J. T., Y. Hu, R. Liu, F. Masson, A. D’Amico, S. Carotta, A. Xin, and P. R. Walker. 2007. Brain microenvironment promotes the final functional M. J. Camilleri, A. M. Mount, A. Kallies, et al. 2011. Id2 expression delineates maturation of tumor-specific effector CD8+ T cells. J. Immunol. 179: 845–853. differential checkpoints in the genetic program of CD8a+ and CD103+ dendritic 25. Mionnet, C., V. Buatois, A. Kanda, V. Milcent, S. Fleury, D. Lair, M. Langelot, cell lineages. EMBO J. 30: 2690–2704. Y. Lacoeuille, E. Hessel, R. Coffman, et al. 2010. CX3CR1 is required for airway 13. Rankin, L., and G. T. Belz. 2011. Diverse roles of inhibitor of differentiation 2 in inflammation by promoting T helper cell survival and maintenance in inflamed adaptive immunity. Clin. Dev. Immunol. 2011: 281569. lung. Nat. Med. 16: 1305–1312. 14. Finotto, S., M. F. Neurath, J. N. Glickman, S. Qin, H. A. Lehr, F. H. Green, 26. Cui, W., Y. Liu, J. S. Weinstein, J. Craft, and S. M. Kaech. 2011. An interleukin- K. Ackerman, K. Haley, P. R. Galle, S. J. Szabo, et al. 2002. Development of 21-interleukin-10-STAT3 pathway is critical for functional maturation of spontaneous airway changes consistent with human asthma in mice lacking T- memory CD8+ T cells. Immunity 35: 792–805. bet. Science 295: 336–338. 27. Jones-Mason, M. E., X. Zhao, D. Kappes, A. Lasorella, A. Iavarone, and 15. Hogquist, K. A., S. C. Jameson, W. R. Heath, J. L. Howard, M. J. Bevan, and Y. Zhuang. 2012. E protein transcription factors are required for the development F. R. Carbone. 1994. T cell antagonist peptides induce positive selec- of CD4+ lineage T cells. Immunity 36: 348–361. tion. Cell 76: 17–27. 28. Yeo, C. J., and D. T. Fearon. 2011. T-bet-mediated differentiation of the activated 16. Farley, F. W., P. Soriano, L. S. Steffen, and S. M. Dymecki. 2000. Widespread CD8+ T cell. Eur. J. Immunol. 41: 60–66. recombinase expression using FLPeR (flipper) mice. Genesis 28: 106–110. 29. Kee, B. L. 2009. E and ID proteins branch out. Nat. Rev. Immunol. 9: 175–184. 17. Hennet, T., F. K. Hagen, L. A. Tabak, and J. D. Marth. 1995. T-cell-specific 30. D’Cruz, L. M., K. C. Lind, B. B. Wu, J. K. Fujimoto, and A. W. Goldrath. 2012. deletion of a polypeptide N-acetylgalactosaminyl-transferase gene by site- Loss of E protein transcription factors E2A and HEB delays memory-precursor Downloaded from directed recombination. Proc. Natl. Acad. Sci. USA 92: 12070–12074. formation during the CD8+ T-cell immune response. Eur. J. Immunol. 42: 2031– 18. Driegen, S., R. Ferreira, A. van Zon, J. Strouboulis, M. Jaegle, F. Grosveld, 2041. S. Philipsen, and D. Meijer. 2005. A generic tool for biotinylation of tagged 31. Langlands, K., X. Yin, G. Anand, and E. V. Prochownik. 1997. Differential proteins in transgenic mice. Transgenic Res. 14: 477–482. interactions of Id proteins with basic-helix-loop-helix transcription factors. J. 19. Seillet, C., J. T. Jackson, K. A. Markey, H. J. Brady, G. R. Hill, K. P. Macdonald, Biol. Chem. 272: 19785–19793. S. L. Nutt, and G. T. Belz. 2013. CD8a+ DCs can be induced in the absence of 32. Lasorella, A., M. Noseda, M. Beyna, Y. Yokota, and A. Iavarone. 2000. Id2 is transcription factors Id2, Nfil3, and Batf3. Blood. 121: 1574–1583. a target and mediates signalling by oncoproteins.

20. Belz, G. T., W. Xie, J. D. Altman, and P. C. Doherty. 2000. A previously un- Nature 407: 592–598. http://www.jimmunol.org/ recognized H-2Db-restricted peptide prominent in the primary influenza A virus- 33. Ji, M., H. Li, H. C. Suh, K. D. Klarmann, Y. Yokota, and J. R. Keller. 2008. Id2 specific CD8+ T-cell response is much less apparent following secondary chal- intrinsically regulates lymphoid and erythroid development via interaction with lenge. J. Virol. 74: 3486–3493. different target proteins. Blood 112: 1068–1077. 21. Belz, G. T., N. S. Wilson, C. M. Smith, A. M. Mount, F. R. Carbone, and 34. Lin, Y. C., S. Jhunjhunwala, C. Benner, S. Heinz, E. Welinder, R. Mansson, W. R. Heath. 2006. Bone marrow-derived cells expand memory CD8+ Tcells M. Sigvardsson, J. Hagman, C. A. Espinoza, J. Dutkowski, et al. 2010. A global in response to viral infections of the lung and skin. Eur. J. Immunol. 36: 327– network of transcription factors, involving E2A, EBF1 and Foxo1, that orches- 335. trates fate. Nat. Immunol. 11: 635–643. by guest on September 27, 2021 

SUPPLEMENTAL TABLE 1. Location of the E2A binding sites in the Tcf7 locus and PCR primer pairs used for the CHIP-PCR assay.

Primer Location Forward Primer 5’ - 3’ Reverse Primer 5’ - 3’ number chromosome 11

Mouse  CATAGATGAAGCTGCCAC GTGGGCAAGGACAAAGCAT chromosome ATAGGT TA 1 gene-less region

Tcf7-1 52,088,862 - GGCAGAGGATGTACCTTC AGGAAAACCCTTCTGCCTCT 52,089,974 CA

Tcf7-2 52,106,572 - TGGGAGAGGCACTTAACC CTCCCCAGCTGTGCTAACTC 52,106,676 AC   

Tcf7-3 52,123,159 - GGGTAGGATGTTTTGCCTC GTCAACCCTTTCCCAAGGAC 52,123,270 A

Tcf7-4 52,127,629 - TTGTTGCCCTATCTCGTGG ACCCAGTGGAACAGATCAG 52,127,715 T G

Tcf7-5 52,132,030 - TCAGCTGGCTTTGATGTCA CTGACTTCCAGCTGCTTGC 52,132,144 C

1 SUPPLEMENTAL TABLE 2. Comparison of influenza-specific Id2-deficient and wild-type CD8+ T cells (Day 9)

TargetID Id2fl/fl LckCre+_vs_WT_log2FC Idfl/fl LckCre+_vs_WT_adj.P.Val Cx3cr1 -4.307198124 2.59E-05 Olfm1 -2.893651346 0.000171381 Nhsl2 -3.776050467 0.000171381 Slc43a1 2.830456142 0.000200171 Id3 3.238373539 0.000200171 Itgad 2.059043161 0.000200171 Ctse -3.83586328 0.000200171 Jdp2 1.889618795 0.000229604 Tbc1d2 -2.137858351 0.000250215 Gzma -1.816270463 0.000250215 Slc16a5 3.601919268 0.000250215 S1pr5 -3.559780394 0.000250215 Ptprv -2.315043994 0.000250215 Zeb2 -2.657630565 0.000250215 Ifitm1 -3.442856289 0.000250215 Slamf6 1.905011515 0.000250215 Kbtbd11 2.507004366 0.000337226 Adam8 -2.519356239 0.000526992 Rnf130 -1.943570467 0.000551141 Itga1 -2.073639471 0.000551141 Klre1 -3.33609698 0.000577898 Lat2 -2.040063279 0.000577898 Ifitm2 -2.997152175 0.000577898 Kdm5b 1.459107874 0.000577898 Sytl2 -1.410183371 0.000577898 Cxcr5 3.600483522 0.000577898 Fam129b -2.253106117 0.000635368 Cish -2.441457218 0.000635368 Slc48a1 -1.423712954 0.000656109 Srgap3 -2.62284122 0.000660052 Il6ra 1.843268927 0.000727721 Nrbp2 -2.225376322 0.000727721 AI646023 -2.174948302 0.000727721 Fam101b 2.115110318 0.000727721 Msc -2.661327119 0.000727721 Tppp3 -4.098068284 0.000759941 Trib3 1.601718228 0.000759941 Ccl9 -2.138842385 0.000759941 Osbpl3 -1.274793035 0.000817171 Enho -2.12295514 0.000817891 Speer6-ps1 1.81373482 0.000827119 Rab3d -1.439028971 0.000827796 Il18rap -1.364319515 0.000831177 AA467197 -1.640017858 0.000831177 Aqp9 -2.642219908 0.000831177 Qpct -1.536815964 0.000913809 Ephx1 1.597779658 0.000959075 Palm -1.499724701 0.00097787 Muc1 -1.520509396 0.00097787 Spo11 1.942905629 0.000978241 Mybl2 -1.356496896 0.000994364 Adk 1.570495843 0.00099795 Fam20a -1.732974506 0.00099795 Dennd3 -1.156424685 0.00099795 Arl5c 1.202873274 0.00099795 Prkar2b -1.496948019 0.00099795 Dapk2 -1.193386977 0.001031276 Sema7a 1.677485247 0.001081481 2600006K01Rik -1.667983568 0.001081481 Trim36 -1.379026095 0.001089729 Fosl2 -1.353419861 0.001089729 Miat -2.928894074 0.001089729 Egln3 1.584812607 0.001089729 Reep5 -1.376897344 0.001089729 Stard10 -1.413512164 0.001089729 Cbfa2t3 2.491233939 0.001089729 Thada 1.240441784 0.001092087 Gpr97 -1.993022923 0.001092087 Rasl12 -1.345753211 0.001092087 Lamc1 -1.281573062 0.001092087 4930583H14Rik 1.978876981 0.001191267 Ret -1.944482797 0.001224629 Btla 1.378988248 0.001224629 Tspan17 -2.145799477 0.001257546 Matk -1.701626634 0.001359099 Vim -1.406648652 0.001359099 Pygl -1.488976138 0.001359099 Nsg2 1.855716477 0.001490384 Car5b -1.391039075 0.001501447 Specc1 -1.339324282 0.001501447 Hmgn1 1.174639662 0.001527652 Cercam -1.885098765 0.001527652 Osr2 -2.611359387 0.001672534 Tec 1.894181592 0.001672534 Susd1 1.394868963 0.001672534 Il2 1.725939166 0.00167617 Swap70 -1.753289215 0.001740502 Cd163l1 1.367616791 0.001740502 E230008N13Rik -1.015466289 0.001824566 Capg -1.237897821 0.001838739 Nccrp1 -1.259290082 0.001838739 Zfp710 1.067165902 0.001841028 Rasl11a -1.871562068 0.001851288 H2-DMb1 1.529048343 0.001851288 Klrg1 -2.079452546 0.001851288 Kcnj8 -1.741281888 0.001851288 Speer3 2.407909532 0.001930671 Lgals3 -1.123933102 0.001969884 Prdx4 -1.190139824 0.001985523 Lpar6 1.017924816 0.001986263 Klrk1 -1.774211492 0.001986263 Chn2 -1.152908639 0.001986263 Cd9 1.841073877 0.001986263 Actn1 1.280130083 0.002123404 Cd244 -2.284245396 0.002313969 Mgst3 -1.134739001 0.002313969 Pdk1 1.117201311 0.002313969 Trib2 1.822416438 0.002313969 Tns1 1.244495094 0.002318172 Pitpnm2 1.468964177 0.002361626 Satb1 -1.146410901 0.002571857 Mboat1 1.096106186 0.002571857 Cables1 1.214878088 0.002571857 Zbtb32 1.069477705 0.002571857 Ezh2 1.140064082 0.002643349 Ogfrl1 -1.080585244 0.002746686 Xdh -1.548590627 0.002746686 H2-Aa 1.886967712 0.002746686 Fxyd4 -1.80824761 0.002766812 Pacsin1 2.426924895 0.002847082 Smarcal1 1.048357335 0.002904941 Pou2af1 1.202535172 0.003152806 Ccr2 -1.485454044 0.003152806 Gcnt2 -1.385460638 0.003159062 Csf1 -1.378772049 0.003261457 Egr3 1.893891499 0.003299652 Sema4f -1.210887796 0.003501272 Fbxo27 -1.399013677 0.003501272 Fn3k -1.518536762 0.003585367 Acot11 -1.44806258 0.003774862 Clip4 -1.692297792 0.003774862 Gngt2 1.175306274 0.003794309 Ly6d 1.589977176 0.003794309 Ccr7 1.934954441 0.004280628 Rcbtb2 1.246158075 0.004294228 Gpr155 1.051063152 0.004403078 Cd74 2.528755503 0.004403078 2900041M22Rik -1.150887213 0.004424674 Klrd1 -1.027143783 0.005138551 Pde2a 1.040253423 0.005845908 H2-Ab1 2.372323643 0.005845908 Tbxa2r 1.496381954 0.005845908 Aifm2 -1.265388964 0.005845908 H2-Ob 1.606529091 0.005889683 Pou6f1 1.017446815 0.006172894 Efcab6 -2.477376053 0.006184888 Impdh1 -1.2675844 0.006278779 Ptpn7 2.366275171 0.006404892 Crybb1 -1.507739355 0.006404892 Prkd3 -1.024043196 0.006404892 Cib2 -1.620347798 0.006418822 Etv3 1.069770872 0.006709955 Stx1a -1.121491323 0.00672653 Wdr95 -1.192447144 0.006823175 Speer1-ps1 1.929004531 0.006823175 Gzmm 1.493491158 0.006823175 Dkkl1 -1.122882782 0.006823175 Rap1gap2 -1.090404749 0.006823175 Klrb1c -1.968168706 0.006823175 Plekhg3 -1.119152568 0.006844902 Id2 -4.34582788 0.007140115 Actn2 -1.162119005 0.00723893 Tktl1 -3.168934273 0.007354061 Dpysl3 -1.075626724 0.007419853 Tnfsf9 -1.073907222 0.007647225 Dok4 -1.309744126 0.007997761 Ddr1 1.375567678 0.00828561 Prmt2 -1.002476482 0.00830816 Cd200 1.183114991 0.00830816 Soat2 -1.272161742 0.00830816 Rabgap1l 1.117885816 0.008477364 Arsb -1.337787321 0.008745999 St6gal1 2.039643589 0.009129732 Car2 1.330666651 0.009614196 Rap2a -1.067689368 0.009614196 Cd99l2 -1.094451575 0.009910214 Aim2 1.12939882 0.010631522 Tcrg-V2 -1.759624146 0.010865624 Tcf7 1.42633519 0.011026026 Fam71b -1.073989871 0.011026026 Irf8 1.221474974 0.011413957 Ttyh3 1.030103035 0.0120366 Cox6a2 1.301197079 0.012324793 Wdr12 1.162318556 0.012814567 Klri2 -1.1219968 0.012825214 Tmem205 -1.031070016 0.01308249 Fcrla 1.146742397 0.01308249 Ctsh 1.744067943 0.01308249 Fcgr2b -1.738575566 0.01319292 Gpr114 -1.323710564 0.013260236 Lmna -1.338215571 0.01327089 Acpp 1.047676792 0.013469797 Art2b 1.370845577 0.013958504 Rnase6 1.265771955 0.014078615 Hist1h2ak -1.115743272 0.014078615 Suox -1.322539868 0.014148937 Il2ra -1.014885784 0.014403604 Camk2n1 -1.115317744 0.014705139 Zfp365 1.349324701 0.015024633 Tnfrsf26 1.056677469 0.015181065 Bambi-ps1 1.141447682 0.016133424 Depdc1b -1.130371809 0.016572134 Egr2 1.63579718 0.016626177 Mfge8 -1.005967424 0.017553081 Ypel1 -1.25115551 0.017894211 2010001M09Rik 1.697094274 0.018072533 Smtn -1.321598303 0.018161471 Dnajc27 1.022196714 0.019095901 Coro2a -1.222971635 0.019644764 H2-Eb1 2.134277664 0.020866616 Fgl2 -1.739676062 0.020866616 Dennd2d 1.160965408 0.023118426 Eomes 1.453185507 0.023247073 Gpx8 -1.005372513 0.023247073 Klra3 -1.042887431 0.023247073 Cpm 1.079043754 0.023247073 Pls1 1.239540056 0.024696645 Tmem159 -1.150223847 0.027015481 Il10 2.719036379 0.028023666 1700001O22Rik 1.092499159 0.029974498 Socs3 1.575801485 0.031871985 Cd79b 1.002888376 0.033916651 4933402N22Rik 1.839827431 0.035976116 Chst2 1.034562657 0.0395583 Anxa1 -1.801667419 0.039821035 Tyrobp 1.096890933 0.040598208 Il6st 1.015090574 0.04060524 Itgae -1.581748027 0.040803663 Adcy6 1.043773041 0.041885998 Klrc1 -1.984871569 0.043506586 Mical2 1.143532003 0.045201322 Gramd1a -1.201317446 0.045899911 Prss2 1.299925256 0.047521206

List of differentially expressed genes in Id2-deficient virus-specific CD8+ T cells following fl/fl Cre+ b influenza virus infection. Microarray analysis of wild-type and Id2 Lck D NP366-specific CD8+ T cells isolated from PR8-primed/HKx31-infected mice spleen and analysed by whole- genome mRNA expression microarray. Genes are considered differentially expressed if they had an adjusted p-value ≤ 0.05 and a fold change ≥ 2. Fold change (log2) between fl/fl Cre+ b + Id2 Lck D NP366-specific CD8 T cells and their wild-type counterpart and adjusted p- values are shown. 

Figure S1

Supplemental Figure 1. Id2-deficient CD8+ T cells do not differentiate into KLRG1+ effector cells following secondary infection with influenza virus. (A) PR8-primed Ly5.2+Id2LckCre+:Ly5.1+ mixed bone marrow chimeric mice were infected with b + + HKx31 and D NP366 influenza-specific CD8 T cells in spleen and were analysed for the indicated markers within the wild-type (Ly5.2-) and Id2-deficient (Id2fl/flLckCre+, Ly5.2+) compartments. Data are representative of at least 3 independent experiments. + b + (B) Frequency of KLRG1 D NP366 CD8 T cells following secondary (PR8/HKx31 infection) influenza infection. Data show the mean ± SEM of KLRG1+ influenza virus-specific cells within the CD8+ T cell compartment in indicated tissues. Data are pooled from three independent experiments (n=16).



1

Figure S2

Supplemental Figure 2. Id2 dose-dependent transcriptional control of effector CD8+ b + + T cell differentiation. (A) Segregation of D NP366 CD8 T cells based on Id2-GFP expression. PR8-primed Id2gfp/gfp mice were challenged after one month with influenza i.n. HKx31 virus and analysed on day 9 for the expression of Id2-GFP in b + + + b + D NP366 CD44 CD8 T cells (contour plot, left panel). D NP366-specific CD8 T cells (pre-sort), were then separated into virus-specific CD8+ T cells that expressed intermediate (GFPint) or high levels of GFP (GFPhigh). Data show representative int high b + profiles of post-sorted Id2-GFP and Id2-GFP D NP366-specific CD8 T cells. (B) int high b + Purified Id2-GFP and Id2-GFP D NP366-specific CD8 T cells were analysed by mRNA whole genome microarray (Illumina). Scatter plot shows on the x-axis the log2 fl/fl Cre+ b fold change of DE expressed genes in Id2 Lck vs. wild-type (WT) D NP366- + specific CD8 T cells and the y-axis shows the corresponding fold change (log2) in int high b + Id2-GFP vs. GFP DNP366-specific CD8 T cell populations. The position of select genes is indicated. The linear regression line is shown in blue. (C) Differential + b + expression of GzmB in Id2-GFP expressing CD8 T cells. D NP366-specific CD8 T cells segregated on Id2-GFP expression were co-cultured with naïve Ly5.1+ splenocytes for 4 h in presence of protein transport inhibitor with or without the

NP366-374 peptide. Flow cytometric histograms show intracellular staining for GzmB in + + int high b + Ly5.1 naïve, or Ly5.2 Id2-GFP and Id2-GFP D NP366-specific CD8 T cells. Data show the MFI ± SEM of 13 mice pooled from three independent experiments. 4 int high b + (D, E) 3  10 sorted Id2-GFP and Id2-GFP D NP366-specific CD8 T cells were adoptively transferred i.v. into non infected Ly5.1 recipient mice. One month after high int b transfer, the total number (D) and Id2-GFP and Id2-GFP MFI (E) of D NP366- specific Ly5.2+ CD8+ T cells were enumerated from pooled LNs and spleen. Results are representative of two independent experiments. Individual mice are shown by an open square (GFPint) or black circle (GFPhigh). Statistically significant differences were determined using a 2-tailed unpaired Student’s t-test. NS, not significant.