© 2013 Nature America, Inc. All rights reserved. A.W.G. ( School of Medicine, New York, New York, USA. 1 or activators transcriptional certain that, of confirmation Intraits. cellular down stream targets that enable and enforce stage-specific key with factors transcription of network a by orchestrated to be thought memory—is and effector effec differentiated tor, terminally state—naive, differentiation Each reinfection. from host the protect to and memory long-term form to ability the with few select CD8 understood. well by expression gene and networks tran scriptional on environments proinflammatory unique such of effect the CD8 of potential memory and differentiation that environments and the of effector inflammatory expansion T population the kinetics cell influence infection-specific creates gens patho by different induced response immune innate the expansion, differentiation of stage each reflect that expression gene by That is progression accompanied changes production. in cytokine rapid for capacity the and function cytolytic with cells proliferating active, to metabolically cells effector poor quiescent, from transition CD8 infection, After organs. lymphoid peripheral populate CD8 naive resting, of repertoire diverse a state, uninfected the In tumors. and pathogens intracellular of clearance the in roles CD8 antigens. or viral by bacterial CD8 in initiated programs transcriptional the track and identify to sought we Here preparation data and collection sample of methods standardized and gene-expression highly by system immune mouse the of cells defining in networks regulatory comprehensively and carefully of goal the with partnership biologists computational a and is immunologists between Project (ImmGen) Genome Immunological The T cell immunity. viral model pathogens. Our results provide insights into the regulation transcriptional that influence memory formation and CD8 core signatures transcriptional were regulated similarly, whether polyclonal or transgenic, and whether responding to bacterial or inflammatory milieu and T cell precursor frequencies influenced the of differentiation CD8 memory CD8 expression of key gene signatures can be used to predict the potential memory-precursor of CD8 infection to establish signatures gene-expression and identify putative regulators. Notably,transcriptional we found that the data of unparalleled breadth from the Genome Immunological Project, we analyzed the CD8 After infection, many factors coordinate the population expansion and of differentiation CD8 Michael L Dustin J Adam Best response to infection and memory T cell formation Transcriptional insights into the CD8 nature immunology nature Received 9 August 2012; accepted 21 December 2012; published online 10 February 2013; Division Division of Biological Sciences, University of California San Diego, La Jolla, California, USA. After pathogen clearance, most CD8 most clearance, pathogen After [email protected] + cells ultimately expressed a small subset of genes shared by natural killer T and 1 , , David A Blair + T cells during the response to to response the during cells T 2 , , Ananda W Goldrath

). ). aDV A NCE ONLINE PUBLIC ONLINE NCE + 2 cytotoxic T cells have important important have T cells cytotoxic , , Jamie Knell 3 + A A list of members and affiliations appears at the end of the paper. should Correspondence be addressed to T cells die, which leaves a leaves which die, cells T + T cells T 1 in vivo in & The Genome Immunological Project Consortium A 1 TION + , , Edward Yang T cells is not not is cells T 6 , 7 2– . However, . activation activation

repressors repressors 5 . During During . + + T T cells cells T

cells cells 1 - - - - .

1 over the course of infection, we transferred congenic naive OT-I T OT-I naive congenic transferred we infection, of course the over CD8 pathogen-reactive of profile molecular a Toestablish CD8 in patterns expression regulated Temporally RESULTS populations. cell memory and effector of tion insights into that processes the transcriptional govern the differentia us to profile CD8 allowed analyses These expression. in differences with genes or sors ImmGen Consortium to predict transcriptional activators and repres the of analyses network-reconstruction used and genes coregulated of clusters unknown previously identified have we data, these From settings. infection of various context the in cells, memory long-term CD8 of responses antigen-specific the through occur that changes tional We system. have immune made a systematic and the temporally resolved analysis of of transcrip cells of types different of hundreds for data transcriptomic comparable directly of inventory unmatched its Program offered a unique to opportunity address this given question, populations is lymphocyte not The cytolytic ImmGen well described. how CD8 system; between relationship ever, the transcriptional immune the of cells other with factors transcription transcriptomics. as Such factors are more by identified efficiently unbiased methods such CD8 affect that factors Tcf7 by expression gene of CD8 regulators essential as established well are , , Viveka Mayya CD8 , , + + Eomes T cells, from early time points of activation to the analysis of of analysis the to activation of points time early from cells, T T cells during infection, including those encoded by by encoded those including infection, during cells T 2 + doi:10.1038/ni.253 Skirball Skirball Institute of Biomolecular Medicine, New York University T cells are known to share certain functional abilities and and abilities functional certain share to known are cells T , , Id2 , , + Id3 T cells with potential and memory differing obtain 2 , , Andrew Doedens and and + T cell differentiation are yet to be described. are described. to yet be differentiation T cell 6 + Prdm1 + effector and memory populations, + T cell + T cell throughout transcriptome effector and memory T cells. Using g + , yet it is likely that many additional additional many that likely is it yet , d effector cells. Long-lived T cells. Although distinct 1 3 ,

e c r u o s e r + T and cells other + T cells

+ T cells cells T Tbx21

+

 - - - - ,

© 2013 Nature America, Inc. All rights reserved. related to metabolism (key), presented relative to all genes with that tag. Data are representative of three experiments with a compilation of two (48 h (48 two of a compilation with experiments three of representative are Data tag. that with genes all to relative presented (key), metabolism to related ( genes. correlated of 15% top the showing (right), 100 day at only upregulated genes of or (left) infection of 100 day and 45 day at only upregulated genes of (top) exemplar artificial an to fit expression gene mean of coefficients ( cluster. each in interest of genes plots, above probes; of number indicate corners right bottom in numbers by expression ( 120. over of value expression mean and 0.5 than less of variation of coefficient a set, data the in anywhere twofold over of expression in a change for filtered Lm-OVA, with infection after points time ( axis). (horizontal infection during points time various at cells OT-I naive in expression their to OT-Irelative cells exposed infection- in (Down) downregulated or (Up) upregulated genes CD8 of formation memory and Figure 1 OT-I1 d later,CD8 transferred lowing alternative approach: we first infected mice with Lm-OVA and, response after activation, before the expansion phase, we used the fol changes in gene expression that during the occur stages earliest of the for H2-K cific frequency, albeit higher than the endogenous repertoire of T precursor cells spe low relatively a represented which immunization, 10 before × 5 transferred we later, and 6 days on lection col For analysis. for cells responding of recovery adequate allowed Note 1 ( pipelines ity-control and qual by data-generation profiling the ImmGen gene-expression 8, 10, 15, 45 and 100 of infection and sorted the cells to high purity for CD8 splenic We antigen. collected gen-associated OVA-expressing H-2K ecule mol complex histocompatibility major the by presented 257–264) acids (OVA; amino ovalbumin of fragment a recognizes that (TCR) receptor antigen cell T a of expression transgenic have (which cells e c r u o s e r  ( spleens pooled from sorted samples independent points) time other (all three or 100) day and b c a

) Hierarchical clustering analysis of OT-I cells sorted at various 2 Genes (×10 ) Expression (fold) Expression (fold) 10 15 20 N Gzmb, Il2ra,Il2 response: Initial cytokineoreffector

N 0 5 Time (h) Time (h) Cxcr3, Cx3cr1,Itgam Zeb2, Klrg1,S1pr5, and memory: Short-term effecto

12 12 of OT-Inumber still that ). Weminimum cells the transferred

Gene-expression profiles associated with the activation 24 24 48 48 12 b Ctla4, Ifng, 6 6 Time (h) ) into C57BL/6J mice, which we then immunized with with immunized then we which mice, C57BL/6J into ) K Time (d) Time (d) 8

8 b -means clustering analysis, filtered as in in as filtered analysis, clustering -means 24

10 10 –OVA peptide Id2,

VI 15 15 Listeria monocytogenes I c r 45 45 48 28 53 ) Ten clusters with the most dynamic dynamic most the with ) Ten clusters 100 100 Supplementary Fig. 1a Fig. Supplementary N 8 6 N Egr2, Cd69,Pkm2 division: Preparation forcel Time (h) Time (h) Bcl2, Tcf7,Il7r,Foxo3 Memory precursor: 12 12 24 24

+ 48 48 T cells. ( T cells. Time (d) 8 Myc, Tnf,Id3, 6 10

6 , 9 + Time (d) Time (d)

8 8 of the . To understanding better gain 10 10 isolated and then mice into the cells 15

15 VI 15 II I 577 12 45 45 l a 100 100 45 3 Down Up ) Quantification of ) Quantification (Lm-OVA) as a model patho 100

N Klf3, S1pr4,S1pr1 memory: Naive orlateeffector N Time (h) Time (h) Cdc4 Myb, Hist1h3a,Cdk1, Cell cycle&division: 12 12 24 24

48 48 5 and and

6 Cxcr6, Klf2, 6 b Time (d) Time (d)

8 8 +

10 10 6, on days T cells 3 VIII

b donor cells 1 d 1 cells donor Supplementary Supplementary

15 15 III Naiv but with a change in expression of over 1.4-fold. Each line represents a single probe; probe; a single represents line Each 1.4-fold. over of expression in a change with but

21 29

45 45

100 9 100 1

e

N N Sell, Nsg2,Slfn5,Cnr Naive andlatememory memory: Short-term effectoror Time (h) Time (h) 12 12 24 24 12 e Genes per GO tag (%) 48 48 Prdm1, Hif2 100 6 6 20 40 60 80 Time (d) Time (d) ------0 8 8

10 10 Time (h)

15 15 IV IX altered expression were downregulated than were upregulated ( u infection; of h 48 within peaked cells OT-I naive versus cells OT-I activated. were cells transferred the of all that indicated ( as expected upregulated, uniformly of the expression whereas sion CD69 and cells, in CD44 was memory by reexpres followed activation, with downregulated was and CD27 of and/or memory.activation markers We found that CD62L phenotypic CD127, expression of expression for cytometry flow by cells T ( subset memory the into differentiation and contraction rapid more 10 × (5 frequency at a precursor lower to that transferred of cells was similar of markers associated with activation and differentiation by these cells cells were rapidly recruited into the immune response. The expression OT-I transferred that so established to become infection the allowed 10 × (1 cells precursor of frequency greater a included approach This transfer. after 2 and 1 0.5, days on cells the 45 45 24 Supplementary Fig. 2 Fig. Supplementary I 82 84 n e 100 100 2 ) Quantification of genes in each cluster a given gene ontology (GO) tag tag (GO) ontology gene a given cluster each in genes of ) Quantification The number of genes with different expression in infection-exposed a expectedly, at later time points, a greater proportion of genes with with of genes proportion a greater points, at time later expectedly, :

N Cd44, Klre1,Il12rb Tbx21, Prf1,Bhlhe40, Late effectorormemory

Time (h) N Time (h) Snor memory: Early effector,late 12 12 3 48 II 24 24 with consistent were differences any and cells), transferred a aDV 48 48

6 6 Ly6a, Rpl, Time (d) 8 Time (d) 8 III A 10 10 6 n NCE ONLINE PUBLIC ONLINE NCE

15 15 V X ≥ 225 2

3 per sample). sample). 3 per 45 45 10 92 100 0 IV : d ). We analyzed the transferred OT-I CD8 OT-I transferred the analyzed We ). 8 N d ) Heat map (bottom) of the correlation correlation the of (bottom) map ) Heat –3. Expression (fold) V 6 Time (d) Cluste 0 8 10 10

r 15 Time (d) VI A

45 Fig. 1b Supplementary TION 3.

0 100 Bcl2 Cnrip Prss1 Edarad Dmrta1 Aqp9 15 VII

1 2 nature immunology nature d 6 transferred cells) and and cells) transferred –3. Expression (fold) N Fatty-acid metabolis Respiratory chain Glycolysi 0 VIII

6 45 Time (d) 8

s 10 3.

IX 15 0 100 45 ), which ), which 100 Fig. 1 Dock Myo3b Cdh1 Yes1 Xcl1 Unc5a Atn m X 1 a 9 ), + -

© 2013 Nature America, Inc. All rights reserved. specifically downregulated at the later time points, which was unex was which points, time later at the downregulated specifically Fig. Supplementary ( PCR quantitative by abundance mRNA specific’.We confirmed ‘late-memory as seven and specific’ ‘memory as genes six 100; day or 45 day R until expressed not (genes cluster CD8 each of memory plar late and memory of T function the in specifically recognition. antigen after h 12 as as early was initiated priming with that occurred function effector cells dendritic peptide–pulsed OVA with immunization interferon- after h 24 as early as an detected be can that signal mice reporter in shown been has it Although interferon- B, as such granzyme molecules effector of expression of upregulation rapid the was unexpected More technique. the of validity the of confirmation survival and proliferation to linked been have the 12 interleukin receptor which and (IL-12R), Krüppel-like-factors, ( known roles in memory formation, such as T-bet ( points. time memory into high remained X cluster in expression gene whereas time; over slowly fell then and rose then cells, naive in low was IX cluster in expression gene 6; day after recovered expression and response, the in early cells T naive in than infection to responding cells in lower was VIII cluster in sion were present in cluster VII, including cells memory-precursor with associated Genes as and CX3CR1. S1PR5 such receptors, migratory as well as Id2, and CD11b KLRG1, as ally with associated effector and CD8 effector-memory Cluster VI genes included that phenotypic markers encode tradition population. memory the in as well as cells activated early in and naive expression moderate had and RNAs nuclear small and proteins ing includ response, immune the to have in roles suppressing suspected as such genes with expression in the highest naive and CD8 memory with burst. IV the Cluster proliferative coincident included processes III encoded molecules mostly related to the cell cycle and proliferation, as such TCR activation, to RNA processing, but this cluster also included genes downstream of related or RNA-processing molecules molecules encoded proportion large a (577); genes most the had II Cluster molecules. effector early This cluster had the fewest genes and notably included genes encoding cells. naive in that than higher remained but immediately decreased genes with expression that was upregulated 12 h then after activation, designations ( cluster, each groups of that genes gene-ontology shared we identified patterns. dynamic most the with clusters of expression (by ences in expression into unbiased groups according to kinetic patterns 1 Table Supplementary ( examined genes 25,194 of total the from set data the in two any in twofold least at of expression in difference a had CD8 To the of T course activation. cell response the to infection over with Lm-OVA, initial expression we gene selected with 7,195 genesin that changes associated tem visualize a expression required gene memory of to pering transition the that suggested which nature immunology nature given pected, the quiescent state of cells. memory As anticipated, Prdm1 2

el, e okd o tasrps ht orltd ih n exem an with correlated that transcripts for looked we cells, > 0.85). Very few genes met those strict criteria; we identified identified we criteria; strict those met genes few Very 0.85). > To identify unique genes encoding molecules potentially involved involved potentially molecules encoding genes Tounique identify Included in clusters IX and X were many genes whose products have with associated probably processes biological the determine To Cnr2 ), in addition to inflammatory receptors such as S1PR1 and and S1PR1 as such receptors inflammatory to addition in ), Sell and (which encodes CD62L), and genes encoding molecules molecules encoding genes and CD62L), encodes (which Slfn5 Fig. Fig. 1 K . The genes of cluster V mostly encoded ribosomal . The genes of ribosomal cluster V encoded mostly -means clustering). We further investigated the ten c and

3 ). According to those criteria, no genes were were genes no criteria, those to According ).

Egr2 ). We next parsed those probes with differ with probes those parsed Wenext ). γ Supplementary TableSupplementary 2 aDV and IL-2, encoded by genes of cluster I. I. cluster of genes by encoded IL-2, and and A 1 NCE ONLINE PUBLIC ONLINE NCE 1 , our data suggested that the gain of gain the that suggested data our , Cd69. Il7r As expected, genes in cluster genes As expected, , , Bcl2 and 1 0 ; this provided early early provided this ; Tbx21 ). ). Cluster I included Tcf7 A TION + . . Gene expres ) and Blimp-1 subsets, such Fig. 1 Fig. Fig.

+ samples samples

T cells, 1 d b and and and and Bcl2 + + γ ------

effector-cell effector-cell marker KLRG1 (ref. calcium-dependent homophilic adhesion molecule and a ligand of the fied formation memory and/or maintenance, to with our knowledge. Notably, we associated identi before never genes, late-memory-specific and of memory-specific of our identification accuracy in the fidence populations effector-memory emergence of long-lived memory cells from the heterogeneousthe pool with of consistent survival, and formation memory to linked been subset; memory the in higher was expression involved in fatty-acid biosynthesis, whereas the latter two clusters clusters two latter the whereas biosynthesis, fatty-acid in involved resented rapidly dividing cells and included genes encoding molecules tributed among clusters II, III, V and VIII; the former dis two were clusters rep metabolism fatty-acid and chain respiratory the to related molecules encoding genes whereas II, cluster in exclusively were almost molecules glycolysis-related encoding Genes pathways. bolic to annotations examine the expression of gene-ontology genes encoding molecules in meta specific used we interest, broad of topic a is of cells. late-memory identifier surface unique a as serve also may Expression niches. microenvironmental distinct access may and interactions adhesive This that suggested the oldest cells memory may be capable of unique regulators of gene expression gene of regulators potential identify to engineering reverse network cutting-edge used we system, immune the of cells of types all includes which Project, of clusters ten CD8 the across observed regulation gene of patterns ent differ the of basis the investigate to which with platform a provided Project in from ImmGen the data set the of The breadth information response T cell the of regulators Predicted formation memory proper for necessary is metabolism fatty-acid to glycolysis from switch a that observation the supported fatty-acid oxidation ( those encoding molecules involved in the transport of acetyl-CoA and and cells as memory as well cells in naive maintained genes included or repressors from all of the fine modules that showed enrichment enrichment showed that modules fine the of all from repressors or activators predicted We pooled type. cell particular a in module lar for as a a particu regulator its activity indicates that type cell a given in weight’ cluster, in a each with each ‘regulatory identified genes the of ‘regulators’ transcriptional predicted were module fine each Associated with X. cluster in genes for enrichment significant most the CD8 activated and cells NKT cells, (NK) killer ral For example, fine module 99, which includes genes expressed in natu ( cluster each for of genes enrichment showed significant modules fine that several the to P lower or 0.05 of rate false-discovery Benjamini-Hochberg a of application (with Project ImmGen of the modules fine in I–X ters clus of genes for enrichment significant statistically any identify to CD8 each (comparing two for groups test hypergeometric a used we modules, those in were ters the from data ImmGen Project. Toall where determine our activated CD8 including analysis by ‘leveraged’ sets, gene those of profile expression the of basis the on module each for ­regulators algorithm the Ontogenet used then ref. modules, of modules, (explanation Project ImmGen the entire of the set from data identified genes coregulated of groups of consists M -value table of all ten clusters throughout all fine modules). We found o As the metabolic regulation of the differentiation of memory T d Cdh1 + s R T cells identified. As part of the global analysis of As the analysis ImmGen of part the global T identified. cells e g as a gene.late-memory-specific Its product, E-cadherin, is a s / h m t t o Supplementary Fig. Supplementary p d : / u / w l e w s . Fig. Fig. 1 w h t . m i m l ). Each ‘module’ of the ImmGen Project Project ImmGen the of ‘module’ Each ). m e and g e 1 1 n 8 2 + (metadata, (metadata, . . That our enhanced con observation . T cell cluster to each fine module) module) fine each to cluster cell T o Supplementary TableSupplementary 2 1 r 3 g

) ) and the integrin CD103 (ref. 4a 1 / 8 M to predict likely transcriptional transcriptional to likely predict and and o d s R Supplementary Table3 Supplementary h e t g t p s / : / m e c r u o s e r Bcl2 / w o + w d T cells, showed showed cells, T expression has has expression u w l . e i 15– ). ). These data m s + . T cell clus h m 1 t 7 m . g e l n . We ).

. cells cells o 1 1 r 4 g 8 ). ). ). ).  ------/ ;

© 2013 Nature America, Inc. All rights reserved. sion by the CD8 the by sion upregulated only very late during infection when we evaluated expres IL-7R by expressed Genes or slowly in decreased the phase memory ( sustained was then and h) (48 infection of peak atthe increased sion genes from clusters VI, IX expres and genes whose included X; these Fig. 5a were ( cluster in III of which half almost and mitosis, ulation; these included genes encoding molecules involved in cell cycle the IL-7R for enrichment the majority of (86%) genes expression with in higher Lm- with OVA infection ( of context the in genes those of expression the virus choriomenengitis lymphocytic with infection IL-7R of expression gene the comparing data microarray published used we paradigm, cells’ effector ‘memory-precursor called been have and whereas cells memory long-lived become to cells, on go that effector short-lived as IL-7R identified been have cells predict T cell fate early in the immune response response. cell CD8 of the regulators of additional identification for the promise that loss of associated with naive and long-term memory populations; it is known genes many include which VIII, IV,and VII clusters in genes the of subset this of generation impaired in results regulators these of each of loss the that known is it cells; effector-memory short-term in expressed positively regulate genes in clusters VI, IX and X, which include genes the products of Fig. ( proteins STAT encoding genes many Erg2 roles in CD8 in roles have to known regulators transcriptional encoding genes many fied h by atthe meta-analysis indicated (as expression gene of regulators as role predicted their of the of regulators expression activity-weighted the of basis the on type cell each in genes modules’ of a generate of the expression prediction the Weweight. to regulatory model a used linear predicted the of order in them ranked and CD8 from genes for population. cell that of program regulatory the to contribute not 0 do of a ‘weight’ with genes which for weight, predicted of order in and biology T cell to ( algorithm Ontogenet the of use the through predicted (key) T cells of regulators selected encoding genes for Consortium, ImmGen the by identified genes coregulated of modules fine of context CD8 activated in genes for Enrichment activation. T cell of regulation transcriptional predict 2 Figure e c r u o s e r  Supplementary Fig. 4 Fig. Supplementary

t m The expression patterns of CD127 (IL-7R) and KLRG1 can be used to +

l , , 4 ) in T cells, CD8 ) T in cells, T cell clusters (identified in in (identified clusters T cell ). For example, on the basis of these results, we predicted that that predicted we results, these of basis the on example, For ). hi Egr3 Supplementary Supplementary Fig. 5 19 , c KLRG1

, ). ). The IL-7R 2 1 Coregulated genes can be used to used be can genes Coregulated lo 0 8 1 . . Conversely, the we identified CD8 based on enriched modules modules enriched on based , , 8 Tcf7 . We assigned colors on the basis basis on the . Wecolors assigned Prdm1 + T cell activation and differentiation, including including differentiation, and activation cell T lo impairs memory formation. Thus, this strategy holds Thus, formation. strategy this impairs memory Id2 + CD8 T cell population than in the IL-7R + population as a whole, which supported the idea idea the supported which whole, a as population , , , , Tbx21 hi Bcl11b lo ), ‘curated’ by relevance relevance by ‘curated’ ), + + and IL-7R and effector cell population showed enrichment for + T cells or activated T cells ( T cells or activated T cells T cell clusters (I–X) (I–X) clusters cell T h T cell populations include a subset of cells cells of subset a include populations cell T t t p and hi , , : / cells near the peak of the response were response the of peak the near cells ). ). Cells near the peak of infection showed / Tcf7 w Prdm1 w Fig. Fig. lo , , w 1 CD8 9 Bcl6 . . . To put our data in context of that i m 1 ) in the the ) in in various combinations would m , , + g Foxo1 effector cells responding to to responding cells effector Fig. Fig. Tcf7 e n Supplementary Fig. Supplementary . o as encoding a as regulator encoding r 2 , , g 21– and and / Foxo3 M 2 Fig. Fig. o hi 3 VIII VI . . IL-7R IX VI IV III V X d II CD8 I Supplementary Supplementary 1 I Supplementary Supplementary s 9 , , R and analyzed analyzed and 2 Id2 0.199 to0.00044) Smyd1 Tbx2 Smad3, Prdm1 Rxra, Tcfe3,Stat6,Irf5,Id2,Smad2,Klf6,Tbx21,Eomes, Smad4, Klf6, Nfatc2, Pias1, Tbx21, Gata3,Nfatc3,Lef1,Tox,Nfatc1,Bcl11b,Tcf7, Zeb2 Klf12, Relb Nab1, Klf6,Irf5,Tcfe3,Rxra,Id2,Ciita, Smad7, Stat2 Cnbp, Irf1, Ets2,Hif1a Nfatc3, Nfatc1, Ets1, Irf7,Irf9,Bcl11b,Elk4, Hat1, Whsc1 Uhrf1, Dnmt1,Foxm1,E2f8,Suv39h1,Hmgb2, E2f6, Nkrf,Myc,Tfdp2 Churc1, Phf5a,Cbx5,Hdac6,Taf9,Mybl2,E2f1,Elof1, Runx2 Tbx2 e ). We). identi g + s T cell pop lo , , / 1 1 Smad7

(regulator weights:0.117to0.00182) m

, Nab1,Tcfe3,Irf5,Id2,Runx2,Foxp1,Eomes, , Id2,Relb, KLRG1

Tcf3 Smyd3 Foxo , Bhlhe41,Stat4, (regulator weights:0.199to0.00468)

Nab1 Eome o Tbx21

d , 5a Rora 3 , Tcf7,Tox,Id2,Rxra,Foxo4,Foxo3,Tbx21, and and u , Tcf4 , Stat6,Ets1,Bcl11b,Irf7,Lef1,Gata3,Tcfe3, (regulator weights:0.0130to0.00794) , Myc,Elof1,Hat1,Zeb1,Rarb,Hdac2,Atoh1, s (regulator weights:0.135to0.00058) l +

,

Egr (regulator weights:0.0758to0.00097) e (regulator weights:0.199to0.0262) (regulator weights:0.129to0.0101) c T , hi s ). ).

Rora

- - - - Smyd1 . , Predicted activators 2 (regulator weights:0.0953to0.00294)

,

Nab2 , Klf12, (regulator weights0.131to0.0743)

em eoy oprmn ( compartment long- the memory seeded term that precursors represented truly cells these that cells from day 6 of infection; and Id3 and infection; of 6 day from cells from day 6 of infection; KLRG1 KLRG1 Id2-wild-type versus Id2-deficient populations: CD8 of comparisons three for clusters expression ten our of context the in profiles gene-expression evaluated we populations, of particular potential memory known with and were correlated dict in gene expression early in the immune response could be used to pre to used potential be predict memory-precursor can expression Id3 abundant and population, memory cells KLRG1 effector accumulate to failure the in results and cells effector of survival the impairs expression Id2 of loss example, For by CD8 to alter shown Id2 been Id3 and has family E-protein the of factors transcription of Inhibition cells. precursor memory long-term or effector-memory short-lived into ( We next used the CD8 CD8 during clusters gene Core CD8 of subsets those of patterns gene-expression the control may lators of products the that suggested data These cells. effector-memory late short-term and memory effector, short-term by expressed genes include which X, and VI clusters in genes of regulators positive predicted encode Zeb2 example, for functions; potential their about information preliminary provided expression of KLRG1 context in the regulators ( KLRG1 in than subset and PCR showed quantitative that of the abundance involvement, predicted for potential a Confirming strategy. this by CD8 of differentiation the in unappreciated ously antigen exposure. after soon very have occurred that response, which suggested some priming of gene expression may the of 45 day until re-emerge to not h, 48 by off’ ‘turned genes these Zeb2 , Fig. Fig. Supplementary Fig. 4c Fig. Supplementary

Runx2 ,

We also identified many regulators whose involvement was previ was involvement whose many regulators We identified also Foxo1 Lef1 , Runx3 Smyd1 Tcf19 + were expressed ‘preferentially’ in the KLRG1 the in ‘preferentially’ expressed were 1 , Tox,Trim14, effector cells. effector , Smad3,Stat4,Prdm1, ) ) to correlate changes in gene expression during differentiation ,

Zbtb7b (regulatorweights: , Eomes,Nab1,Stat4

mRNA was greater in the KLRG1 the in greater was mRNA Pias3 aDV , 2

Stat2 , 4

Eome A . Conversely, Id3 is needed to sustain the long-lived long-lived the sustain to needed is Id3 Conversely, .

Tcf7 Tcf19 Rora NCE ONLINE PUBLIC ONLINE NCE ,

Tcfe3, , Gata3, s , E2f7, , Rora

Rara, and and + lo , T cell cluster gene above signatures identified , effector cells sorted from the same response response same the from sorted cells effector ). Comparison of the expression of those those of expression the of Comparison ). Pax5, Hdac7 Maf, Etv3,Hic1,Mycn,Runx1,Pax5, Hdac7, Ets1 Maf, weights: –0.00042to–0.0231) Tcfeb, Klf8,Tcfe2a,Pax5,Bcl6,Foxp4,Ctbp2 Stat2, Tcf20,Nfkb1,Nfatc1,Relb,Creg1,Tcf4,Insm1,Zeb1 (regulator weights:–0.00077to–0.10023) Prmt5, Chd7,Nfix,Klf3,E2f2,Tcf4,Irf5,Mef2c,Lmo2 –0.00039 to–0.0826) Ebf1, Maf,Gata3,Mllt3,Bcl11b Taf2, Pax5,Hic1,Mycn,Nfkb1,Pparg,Relb,Cebpb,Zfhx3 Runx3, Pbrm1 Foxd4, Tbx6,Mycl1,Ppard,Cited2,Cbx2, (regulator weights:–0.00007to–0.0735) Irf8, Runx2, Mybl1,Klf12,Lmo2,Egr3,Scmh1,Rorc,Smad3, –0.00006 to–0.0807) Klf12, Ciita,Nfe2l1,Chd7,Foxj2 weights: –0.00086to–0.0215) Pou2af1, Pax5,Mycn,Ebf,Pparg,Cebpb Arnt Insm1, Scmh1, Zeb2 l Notch3

, Tcf7,Foxj2 Scmh1 and other previously unknown regu unknown previously other and lo upeetr Fig. Supplementary IL-7R + , Tbx21,Tcfe2a,Stat6,Ctbp2,Cux1,Tcf4,Irf5 , Zbtb7b,Pax5,Zeb1,Hmgn3,Satb1,Foxp4

(regulator weights:–0.00028to–0.0641) T cell differentiation T cell (regulator weights:–0.00028to–0.0854) (regulator weights:–0.00011to–0.0188)

2 A Klf 4 TION hi (regulator weights:–0.00032to–0.0493) . . To how determine differences 3 Predicted repressors hi , Runx1, versus Id3 versus versus KLRG1 + T cell differentiation T cell

nature immunology nature Cited hi CD8

(regulator weights: (regulator weights: Rora 2 , Foxd4,Nfkbia,Foxp1 lo hi Nab + KLRG1 subset and both both and subset

T cell response response cell T + , , 5b effector T cell cell T effector Zeb2 2 (regulator hi , Prdm1,Srebf2 E2f4 hi , CD8 Activate CD8 T IL-7R d lo short-lived short-lived (regulator cells . ay of Many ). ,Klf7, OT-I cells cells OT-I , , + + + Rora

Tox lo T cells T cells

effector effector

IL-7R d lo OT-I , , 24– and and , Ets1

, , , 2 , 6 lo - - - .

© 2013 Nature America, Inc. All rights reserved. tively). This analysis also provided several unexpected observations. observations. unexpected several provided also analysis This tively). KLRG1 (Id2-wild-type, populations counterpart had VII cluster in KLRG1 expression in higher those Id2-deficient, of 72–81% and IV cluster in genes the of precursor) in Id2-deficient, KLRG1 linked to cluster IV (naive and late memory) and cluster VII (memory III, IV and VII ( II, clusters for comparison each of plots generated we comparisons, KLRG1 Id2-deficient, the in expressed ‘preferentially’ were memory, late and effector early naive, with associated genes included which VII, IV,and V clusters KLRG1 Id2-wild-type, the in expressed ‘preferentially’ were genes, associated effector-memory and many of IX VI, effector included and the III, in X, clusters which analysis CD8 key the of This many for skewing 0.5). shows of independence of hypothesis null the (versus onefold over of expression in change a with genes of fraction the by in any of the comparisons ( potential effector versus potential memory with cells toward biased were cluster each from genes We whether listed. assessed population each comparison, cells with greater memory potential were in the first cent protein (GFP) reporter of Id3 on day 5 of infection fluores green a of expression their of basis the on sorted OT-Icells, nature immunology nature encoding genes which for II, cluster was us to interest particular Of c KLRG1 b a

To further demonstrate the relative gene expression in the three three the in expression gene relative the demonstrate Tofurther P (log ) P (log ) P (log ) KLRG1 10 10 10 10 10 10 10 Id2-KO vsId2-WT –4 –3 –2 –1 –5 –4 –3 –2 –1 –3 –2 –1 2 3 4 5 0 0 0 IL-7R lo 2 14 0.2 Id3 KLRG Id3 Id2-WT vsKLRG 99 53 202 Expression (fold) Expression (fold) Expression (fold) 0.1 0.1 10 lo hi 2 Id2-KO vsId3 10 1 hi 3 1 1 II 1 10 1 KLRG1 hi lo Id2-KO 4 Fig. Fig. 3 85 10 273 317 Id3 191 1 10 10 0.3 0.5 0* 5 I I hi 5 lo 5 29 c lo 0.2 94 90 0.86* 0.73* 113 ). ). We found considerable enrichment for genes 0.49 Expression (fold) Expression (fold) Expression (fold) 0.1 0.1 and Id3 and

I Id2-W aDV hi Fig. Fig. 3 0.11 0.18 0.15 and Id3 and 1 1 III III T KLRG1 1 Id2-KO vsId2-W A NCE ONLINE PUBLIC ONLINE NCE 17 48 hi Id3 b 0.97* 0.97* 10 10 0.9* populations. ); ); here, we determined ‘enrichment’ IV 16 20 14 hi lo lo vsId3 5 vsKLRG1 2 84 and Id3 lo 0.2 1 1 4 0.92* 0.78* populations, whereas genes in in genes whereas populations, Expression (fold) Expression (fold) 0.6* Expression (fold) 0.1 0.1 V V Id3 lo + T cell gene clusters: genes genes clusters: gene cell T T hi lo 0.14* 0.15* 0.35* hi 1 1 IV and Id3 1 hi I cells ( 14 0 0.72* 0.79* 0.81* Frequency ofenrichment 10 10 VII hi 3 3 3 1 5 1 5 2 5 72 24 5 and Id3 and A Fig. 3 hi 1.0 0.2 TION 0.19* 0.47 Expression (fold) 0.6* Expression (fold) Expression (fold) cells than their 0.1 0.1 VIII 4 5 1 2 Id3 4 0.5 ( lo b Fig. Fig. 3

0.03* 0.14* 0.45 VII lo ); ); 90–97% 1 1 IX 1 , respec , 0 3 0 0.12* 0.38 10 a 10 0.2* 5 5 5 X ). ). In 4 9 8 5 - - differently by Id2-deficient versus Id2-wild-type cells and by KLRG1 expressed were division and proliferation with associated molecules ( spleens pooled ( genotype per mice three with experiments of independent three are representative Data of in region. genes that the number indicate and corners left right in bottom Numbers for comparison. each genes cluster-specific showing KLRG1 type KLRG1 wild-type versus ( * potential. effector indicates of 0.0 (blue) frequency and a potential, memory of indicates 1.0 (red) a frequency comparison; for that enrichment of with cells proportion in map indicate Numbers or Id3 middle), KLRG1 top), KLRG1 (both cells wild-type versus Id2-deficient comparisons: VSV-OVAwith for the following of points) time infection pre-peak (both Lm-OVA with and Id3 infection ( in ( cells each. percent VSV-OVA.with day 6 indicate of infection in quadrants Numbers and Id3 (Id2-WT) mice and Id2-wild-type ( cells. 3 Figure ( infection of 45 at day expression in difference a such had fewer even transgenic in cells versus tetramer-positive OVAtwofold cells atover peptide–specific day 8, of and expression in difference a had genes of above ( OT-I the gene identified T across clusters cells, that with cells (tetramer-positive) OVA peptide–specific polyclonal of response transcriptional the compare to data profiling resultant ( infection of 45 and 8 days CD8 antigen-specific endogenous, identify to OVApeptide with loaded H-2K of tetramers We used clusters. gene core of expression the in changes to led frequency and repertoire TCR in differences if mine with infection effector of and memory cells, including the induction of T-bet differentiation expression during the in differences to lead environments, endogenous polyclonal CD8 of frequency precursor lower and of affinity range broader the that known is it However, markers. congenic of expression and/or frequency precursor high artificially by tored and donor cells with transgenic expression of TCRs that can be moni pathogens model of use made have advances immunological Many signatures gene core alter not does clonality TCR thought. previously than heterogeneous more be actually may homogeneous, the populationbe to assumed to containreported memory precursorsbeen has it although Furthermore, cells. deficient of analyses in earlier not identified defects II and III, in of cluster genes in or on activation based differences division the regulation temporal the in responses context of in deficiency Id2immune or Id3 the may have previously of unappreciated phases earliest the that suggested ( expansion of peak the at even counterparts, IV cluster and (short-term division) effector memory) than did and their more short-lived cycle effector (cell III cluster in genes of sion had lower expres potential precursor toward of memory biased cells and Id3 versus KLRG1 cells sion at effector or memory time points. The few genes found to be be to found genes few The points. time memory or effector at sion expres gene alter greatly not did repertoire TCR and precursors cell c 1 Fig. Fig. ) ‘Volcano plots’ of the comparison of Id2-deficient KLRG1 of Id2-deficient ) of ‘Volcano the comparison plots’ + +

4 T cells from nontransgenic mice infected with Lm-OVA on on Lm-OVA with infected mice nontransgenic from cells T T cell responses, as well as pathogen-specific inflammatory inflammatory pathogen-specific as well as responses, cell T c a a ) in wild-type and Id2-deficient samples collected on day 6 of collected samples and Id2-deficient ) in wild-type ) Flow cytometry of CD8 ) cytometry Flow lo

), which suggested that the frequency of antigen-specific T antigen-specific of frequency the that suggested which ), Regulation of core gene-expression modules by memory precursor precursor by memory modules of gene-expression core Regulation cells. Notably, all three comparisons indicated that the subset a ) or three experiments with three independent samples from samples independent three with ) or experiments three hi lo cells (middle), and Id3 (middle), cells IL-7R b hi ) Heat map of the frequency of enrichment of clusters of clusters of ) enrichment map Heat of the frequency b L. monocytogenes L. versus Id3 versus , c ; ; hi n cells versus KLRG1 versus cells lo

≥ cells but in cells expression Id3 had equivalent lo 3 per sample). cells (top), wild-type KLRG1 wild-type (top), cells lo cells (both CD44 (both cells hi + and Id3 Supplementary Fig. 6 Fig. Supplementary T cells from Id2-deficient mice (Id2-KO) (Id2-KO) mice T Id2-deficient from cells 27– hi cells versus Id3 versus cells 2 9 lo . We therefore sought to deter to sought We . therefore hi hi samples collected on day 5 collected samples and Id3 IL-7R + KLRG1 lo cells (both wild-type; wild-type; (both cells Fig. 3a Fig. lo CD8 e c r u o s e r P lo < ( 0.05 lo lo cells versus wild- versus cells IL-7R cells (bottom), (bottom), cells + ). We used the the used We ). T on cells, , c lo ). These data data These ). lo cells cells Fig. Fig. lo ; bottom). ; bottom). χ IL-7R 2 test). test). 1 hi ). ). Few hi cells ; hi  b - - - -

© 2013 Nature America, Inc. All rights reserved. h imn rsos tios tef o ah ahgnc het by threat cytokine pathogenic inflammatory the including each cues, molecular to to responding itself tailors response immune The infections different during signatures gene Similar signatures. gene core in divergence substantial a represent context this in formation memory in differences suggest to used differences phenotypic of number small the that conclusion the supported and formation memory during programming’ tional ‘transcrip similar reflected responses monoclonal and endogenous that showed data our expression, gene of assessment genome-wide byThus, in two populations. the similarly were regulated genes most linked to memory potential or formation (clusters VI–X) showed that clusters of examination however, Notably, limiting. be may signals such which for cells, precursor high-frequency by encountered tion competi the than signals for and antigen activation competition less cell T population drove more proliferation of those cells, antigen-specific perhaps because of low-frequency, a for event priming the that involved molecules encoding in cell cycle and division, this suggested ( populations memory and effector ratio tetramer-positive polyclonal higher by expression uniformly of but small a had III and II clusters in genes that apparent also was it but conditions, both in expression gene of for OT-I a we and cells, ‘evolution’similar tetramer-positive observed 45 day at that to 8 day at expression of ratio the comparing By cells. naive in than cells OT-I in respectively, expression, higher and 3.9-fold 2.4- 3.3-, had and cells naive in than cells tetramer-positive in and Klra9 Klra9 and The only notable exceptions were modest enrichment for RNAs. nuclear small and proteins exclusively ribosomal almost TCRs, encoding were genes types cell two the in differently ­regulated group. per mice three with experiments ** (mean expression in change the indicate key in values (key); clusters of colors OT-I for that CD8 against plotted in as infection after 45 day 8 versus day on ( twofold. of expression in difference a indicate lines diagonal blue (key); clusters of colors match (genes) by identified genes for Lm-OVA, with infection of 45 OT-I day (OT-I)8 or day on versus cells H-2K CD8 4 Figure e c r u o s e r  milieu, antigen load, innate signaling and requirements for CD4

b a P Expression (Tet+) < 0.00001 ( < 0.00001 10 10 10 + b effector and memory T cells. ( T cells. memory and effector 2 3 4 –OVA tetramer–positive antigen-specific (endogenous) cells (Tet cells (endogenous) antigen-specific –OVA tetramer–positive

y Expression (fold) i

− +

Common gene-expression patterns of transgenic and endogenous endogenous and transgenic of patterns gene-expression Common Tet : day 8 vs day 45 0. had 1.9-, 1.8- and 2.4-fold higher expression, respectively, respectively, expression, higher 2.4-fold and 1.8- 1.9-, had x transcripts in OT-I cells (data not shown); shown); not (data cells OT-I in transcripts 2 1 5 i ) ) Expression (OT-I) ± 10 0. Fig. 4 Fig. s.e.m.; diagonal line indicates indicates line diagonal s.e.m.; 2 1 2 t -test). Data are representative of two independent independent two of representative are Data -test). Day OT-I: day8vs45 K Expression (fold) -means clustering analysis ( analysis clustering -means b 8 10 ). As clusters II and III included many genes genes many included III and II clusters As ). 3 10 + 4 T cells; colors in plots (genes) match match (genes) plots in colors T cells; VIII 5 a IX VI III b II ) Difference in gene expression of expression gene in ) Difference ) Comparison of gene expression expression gene of ) Comparison a VIII VI IX VI IV for tetramer-positive cells, cells, tetramer-positive for III X V II I I y = Fig. Fig. –0.014 –0.054 –0.093 x 0.132 0.025 0.270 0.273 0.428 0.240 0.278 . . * D ay 45 1 P ); colors in plots plots in colors );

< 0.001 and and < 0.001 ± ± ± ± ± ± ± ± ± ± 0.04 0.07 0.01 0.01 0.08 0.015** 0.025 0.034** 0.013** 0.11 8 5 0 5 0 9 Klra3 Klra3 * 8 , 3 0 13 do not not do , , + , , T Klra8 Klra8 Klra8 VI + IV

X V cell ) I I - - analysis of variance (ANOVA); of results variance the for simplicity, analysis we grouped events. inflammatory pathogen-specific to ‘blind’ and focused suggested that many aspects of the CD8 ( I–X by clusters identified signatures core gene same VSV-OVALm or with infection after response the of contraction and OVA.induction The VSV-OVA with Lm- with infection after infection after as effectively transcripts repressed during the effector were period not reinduced as ( genes of blocks same the of ’ similar, ‘shut-off and induction generally sequential with were course time each over patterns expression CD8 cell–derived agent. We OT-I either with recipient mice collected the then infected OVA). We transferred OT-I cells as above described and subsequently (VSV- ovalbumin recombinant expressing virus stomatitis vesicular used we above, analyzed Lm-OVAmodel the with Tocontrast tion. same antigen in the context of a bacterial infection versus a viral infec notable differences in the were transcriptional there if response determine to to stimulation Wesought cells. by memory the and effector as the to recruited response, the of kinetics cells and their activation, their fatedifferentiation T of number the affect can which of all help, T cell clusters during their activation, as a clue to differentiation differentiation to clue a as activation, their CD8 during our clusters from cell genes T for enrichment significant statistically any gene- and cell NKT cell, conserved B of pairs pared share system immune com we Project, ImmGen the from Using data modules. the ­expression of types cell Many types cell related functionally of signatures Shared expression. of amplitude the in differences ( populations CD8 in the conserved were of clusters our ten programs gene-expression core the infection, viral and bacterial by ( other mal indication of transcripts uniquely affected in one condition or the mini was VSV-OVA Lm-OVA there with with or quantitative, were monocytogenes L. to responding to VSV of those than responding of populations T cell Fig. 7d ( magnitude in differences subtle to due largely was infection with Lm-OVA ( interferons I type to responsive VSV-OVA with ( infection after than Lm-OVA with extent greater infection a after to upregulated were but infections both in ulated upreg were genes IL-12-responsive the of most that in magnitude, ( tion had moderately higher expression during infection with Lm-OVA (PD-1; whole a as population the by expression different and KLRG1) and (CTLA-4 cells of subset a by differences expression reflected results at of the differences the several those protein level by flow cytometry; atconserved the effector and memory phases ( be to tended VSV-OVA; with differences infection these during sion expres PD1) had higher molecule the costimulatory encodes (which Lm-OVA, with whereas infection during effectively and a (with change in of expression over by twofold) ( analysis this transcripts distinct few a We identified respectively. phase, memory and phase effector the into 45–100 days and 5–10 days for obtained Fig. Supplementary Fig. 7b Fig. Supplementary Supplementary Fig. 7a Fig. Supplementary To better delineate differences between the responses, we used used we responses, the between differences delineate better To that apparent was it responses, overall the plotted we When Supplementary Table 4 Supplementary

5 a Supplementary Table4 Supplementary ). ). These data may yield insight into the more rapid contraction Fig. 5 Fig. ). Some distinctions were detectable; for example, a group of a group for example, detectable; were distinctions Some ). aDV d Fig. 5 Fig. ). Genes known to be IL-12 dependent to known be IL-12 ). Genes A NCE ONLINE PUBLIC ONLINE NCE + . However, whereas the differences after infection infection after differences the whereas However, . T cells and profiled them as described above. as described them and T profiled cells e ), and the few differences tended to represent represent to tended differences few the and ), - OVA were generally superimposable for the the for superimposable generally OVA were ). We found a very similar pattern for genes genes for pattern similar We). very a found Supplementary Supplementary Fig. 7c ). However, this was mainly a difference in in difference a mainly was this However, ). ). For example, For example, ). ). Despite the distinct milieu elicited elicited milieu distinct the Despite ). 3 1 , with moderate skewing toward toward skewing moderate with , γ A δ TION T cell populations to identify identify to populations cell T + T cell responses were antigen

Klrg1 nature immunology nature + Fig. Fig. 5 effector and memory and memory effector ); ); however, again this was induced more induced was c Ctla4 Supplementary Supplementary ). ). We confirmed 3 Fig. 5 Fig. 1 during infec during and and b ), which ), which Fig. Pdcd1

5 + c ------

© 2013 Nature America, Inc. All rights reserved. cytomegalovirus; thy, thymus; spl, spleen; T spleen; spl, thy, thymus; cytomegalovirus; * memory. activated; act, zone; marginal MZ, follicular; blue, foll, (>0.75); center; germinal frequency GC, higher red, (<0.25). (25%): frequency lower correlation of frequency lowest and highest the with comparisons cluster indicate (key) blue and Red CD8 for enrichment of 6 Figure germinal in presence for bias significant a had proliferation) and division in involved molecules encoding (mostly III and II clusters in genes whereas clusters, effector-like or cor memory any with little relation very showed cells B follicular with cells B zone marginal and center germinal of comparisons approach, this of validity the of confirmation In onefold. over of expression in change a with genes ( lineages other or ( mice three with experiments in as VSV-OVA (presented with infection after comparison that ( populations. CTLA-4 percent indicate lines bracketed above numbers (MFI); intensity fluorescent median indicate top ( interest. of genes infection-specific indicate VSV-OVA; labels with infection after upregulated genes blue, Lm-OVA; with infection after upregulated genes red, axis); (vertical points time ( in as lines (blue twofold over of expression in difference with genes of number indicate corners right bottom and left top in numbers ( plots). (above points time memory or effector matching at VSV-OVA or (VSV) (LM) Lm-OVA with infection during cells memory pooled in expression their to relative cells effector pooled in twofold than more 5 Figure nature immunology nature group. per mice three least a c c ) Change in gene expression after infection with Lm-OVA versus VSV-OVA at pooled effector time points (horizontal axis) versus that at pooled memory memory pooled at that versus axis) (horizontal points time effector VSV-OVA pooled at versus Lm-OVA with infection after expression gene in ) Change a Time (d Expression (fold) 1.0 ucinl ahas osre i atvtd CD8 activated in conserved pathways functional (memory cells) 0. 2 5 Correlation )

Car2 iv i 0 Regulation of genes associated with activation state is independent of infection. ( infection. of independent is state activation with associated genes of Regulation Genes induced in CD8 in induced Genes 0.50 0. P Early &lateme 2 Clusters < 0.05 ( 0.05 < 8 6 e Ctla4 ) Comparison of gene expression by pooled effector cells versus pooled memory cells after infection with Lm-OVA (horizontal axis) versus versus axis) (horizontal Lm-OVA with infection after cells memory pooled versus cells effector pooled by expression gene of ) Comparison Early mem Late me vi Fig. Pdcd1 i 0 Expression (fold Lm-OVA (ef ii v χ fector cells VIII

1 VII 2 6 IX VI IV + 0 III m m X V test). ( test). II T cell gene clusters ( clusters gene cell T a I ). We defined ‘enrichment’ as the fraction of of fraction the as ‘enrichment’ Wedefined ).

15 0.34* 0.22* 0.21* 0.24* 0.99* 0.89* 0.62 0.45 0.59 0.57 0.38 0.38 0.5 aDV GC vs foll B cell v iii a ) 45 b ) – Gpr114 iii 0.68* 0.31* 0.83* 0.62* 0.64* 0.48 0.47 0.62 0.51 0.61 0.43 0.46 ) Principle-component analysis of various cells (labels in plot and key) for genes defined in in defined genes for key) and plot in (labels cells various of analysis Principle-component ) c + 0.5 A 100 , memory T cells correlate with gene expression by NKT cells and activated activated and cells NKT by expression gene with correlate cells T memory

MZ vs. foll B cell e NCE ONLINE PUBLIC ONLINE NCE Aqp9 Klrg1 ) or four mice ( mice four ) or d Satb1 0.88* 0.33* 0.98* 0.03* 0.15* 0.87* 0.15* 0.82* 0.58 0.86 0.92 0.8* ) Flow cytometry of OT-Iof CD8 cytometry ) Flow 1* 5 5 0 NKT CD4 Klra3 ix vi 0.75* 0.74* 0.39* 0.23* 0.94* 0.13* 0.18* 0.88* 0.41* 0.82* 0.6 0.8 0.7 VSV-OV γδ NKT CD4 – d Lm-OV Fig. 1 Fig. spleen vs CD4 , , VSV-OV 8 6 0.21* 0.89* 0.77* 0.25* 0.14* 0.86* 0.44 0.38 0.68 0.86 0.88 0.92 γδ

1* + d T cell; IEL, intraepithelial lymphocyte. Data are pooled from three independent experiments with at at with experiments independent three from pooled are Data lymphocyte. intraepithelial IEL, cell; T Act

spleen vs CD4 A A ) per group. ) per c γδ Relative expression ) or memory-specific genes ( genes memory-specific or ) 0.97* 0.64* 0.92* 0.12* 0.62* 0.66 0.56 0.54 0.1* 0.2* 2 – + A

Events (% of max) –3. 1* CTLA-4 15 1 1 Act vs 10 A b γδ 0 MFI 5,477 MFI 1,612 TION γδ 2 45 2 – VSV-OVA; or Lm-OVA with infection after point time each at expression of comparison ) Direct 2 + + 10

+ 3

vs 3. 1. 100 10 T cells and and cells T 4 5 3.

γδ 4 Fig. 4 Fig. 10

2 + 0 5 PD- + cells at day 6 of Lm-OVA infection or day 5 of VSV-OVA infection; numbers at numbers VSV-OVA5 of infection; day or infection Lm-OVA 6 of day at cells b b 1 ). * ). MFI 193 MFI 443 Expression

P (Lm-OVA) - Expression (VSV-OVA) 10 10 10 < 0.001 and ** and < 0.001

b PC 2 2 3 4 18 KLRG1 all NKT cells to demonstrate a strong bias with our memory-specific our memory-specific a with bias strong to demonstrate cells NKT all expected have we might Thus, antigen. cognate their to respond idly expression of markers activation such as CD44, and the ability to rap interferon- as such cytokines of production homeostasis, rapid and survival for IL-15 on dependence a share cells T notably traits, NKT cells and ‘memory-like’ have to known cells of subsets certain in correlations strong found we Conversely, population. activated an cells, B center Expansion 10 T γδ Fig. 1 Fig. 9. MFI 903 MFI 875 2 CD8 17 10 8

Thy 3 + IEL orNK 10 (day6) 69 18 NKT Thy d 4 69 ) in populations of B cells, NKT cells and and cells NKT cells, B of populations in ) 149 e P a CD < 0.00001 ( < 0.00001 ) Heat map of all genes upregulated or downregulated downregulated or upregulated genes all of map ) Heat

CD Expression (fold) Peak 8 (VSV-OVA) + 8 NKT 0. me + naive 2 5 γδ m 149 72 2 NK 72 0. + 16 , V , 2 Contractio PC Lm-OVA versusVSV-OV + γ t cells (left) or KLRG1 or (left) cells 2 -test). Data are from three independent independent three from are Data -test). 1 + CD8 CD8

T Expression (fold γδ γδ γ γδ δ T cell; cell; T 65 + T cells. ( cells. T Spl n (Lm-OVA) 16 + T NKT cells. cells and CD8 (day8) NKT (day10or15) 65 16 Day 15 γ δ a 2 ) ) Heat map of the frequency frequency the of map Heat ) − , V , 54 A 16 Figure 1 Figure 54 γ 2 32 + cells (right) in gated gated in (right) cells − 5 Memor

γδ e c r u o s e r γδ T cell; mem, mem, cell; T Fig. 4 Fig. γ VIII a T cells. cells. T VI , cytolytic ability, cytolytic , IX VI IV III . MCMV, mouse MCMV,. mouse X V II y I I 74 32 IE IE T Act T T T NK NK NK NK NK NK NK OT OT OT OT OT OT OT NK me NK MCMVday NK MCMVday NK 74 γδ γδ γδ γδ a –0.249 –0.303 –0.162 –0.207 L L -I day10 -I day45 -I day15 -I day10 -I day -I day -I naiv ). ). T T T T T T T 25 0.014 0.247 0.048 0.270 0.155 0.089 Late memor sp th T ac V V 4 4 4 4 44 44 44 γδ γ γ +

y

2 2 t – + – + l m memory memory live sp live sp + – + + – 1. 1. 1. e 8 6 ± ± ± ± ± ± ± ± ± ± l l 1 1 1 0.033* 0.078* 0.01 0.010 0.063 0.01 0.089 0.093 0.009* 0.031 r r – + – 0

th th th 57 y 25 y y y 1 1* 7 1 57 *  * -

© 2013 Nature America, Inc. All rights reserved. tor CX3CR1. The identification of IL-12R of identification The CX3CR1. tor proin the IL-12R molecule flammatory to addition in expressed, also were family receptor S1P in T cells expressed when proliferation curb subfamily 5 Schlafen the of members Similarly, cells. T memory and rophages as ( sets gene memory-specific and IV cluster in association neuronal strong with genes of prevalence the was result unexpected One infection. during kinetics expression characteristic with genes additional many tified CD8 activated in pathways biological the present knowledge of gene expression and several known describe fit clusters These stage. each during operate that processes biological tify, through evidence of coordinated regulation, previously unknown gene expression with progressive stages of T cell activation and to iden T CD8 of regulators putative identify and activation cell T of nisms mecha transcriptional the into insight valuable provide data these of relevance transgenic experimental models biological widely used in immunology. the Together of confirmation unavailable previously to those of their endogenous polyclonal counterparts, which provided similar very events differentiation underwent TCR the of expression CD8 of populations monoclonal that found we Furthermore, contexts. infection different substantially despite pathogens, bacterial versus for viral tailored necessarily not is tiation tional program governing and effector CD8 memory transcrip the that concluded we infections, viral and bacterial from response. By comparing the response of CD8 allowed us to visualize core transcriptional changes during the immune we identified clusters of genes with similar expression patterns, which set, data Usingthis memory. late to activation after early from states CD8 of range a covering infection, of course the better understand the complexities of gene-expression changes during Through the use of high-resolution microarray analyses, we sought to DISCUSSION lymphocytes. intraepithelial and cells NK including cells, CD8 CD8 with characteristics, functional as well as profile, transcriptional their of and cells NKT that suggested data these Together analyzed. populations memory the of any to than lations NK were cells more to similar NK cell and CD8 antigen-experienced that indicated which particular, in populations CD8 the with group not did population cell NK ‘memory’ and lymphocytes CD8 effector as as T well cells memory ( cells T memory and subsets innate among different the variation to visualize analysis resting with compared sets, gene memory late V other comparisons, which showed bias in only to a few contrast clusters, In activated responses. immune adaptive and innate between gap T cific’ genes was splenic CD4 spe ‘memory the of expression toward bias significant showed that ( above generated sets gene e c r u o s e r  IL-12 is known to drive CD8

γ cell responses, which offers a offers to resource community.the which responses, cell cells ( cells The ten clusters in our analysis allowed us to correlate changes in in changes correlate to us allowed analysis our in clusters ten The To address those further,similarities we used principal-component 2 Cnr2 + + and V and effector T cells may be more similar to that of innate effector effector innate of that to similar more be may cells T effector , encode molecules with immunosuppressive effects in mac in effects immunosuppressive with molecules encode , Fig. 3 2 + Nsg2 and might have a role in maintaining homeostasis in naive memory cells, whereas the gene-expression signature of of signature gene-expression the whereas cells, memory γ

2 6 − a

γ , , ). This supported the idea that NKT cells bridge the the bridge cells NKT that idea the supported This ). δ Cnr2 T cells showed highly significant bias for early and for early bias significant highly showed T cells Fig. 6 Fig. , , Cnrip1 b ). We found that NKT cells mapped near near mapped cells NKT that found We ). γ δ − + T cells, whereas NK cells, intraepithelial intraepithelial NK cells, whereas T cells, Fig. 1 Fig. NKT cells, compared with resting CD4 T cells toward a terminally β and and + and chemokine recep the prosurvival T cells mapped together. Notably, the Prss12 d ). However, the only population population only the However, ). + T cells but, notably, we iden notably,we but, cells T ). Some of those genes, such such genes, of those Some ). γ δ β T cells may share aspects aspects share may cells T + 3 + was not unexpected, as as unexpected, not was T cells to model antigens 3 T cells with transgenic transgenic with cells T . . Many of the members γ + δ effector T effector cell popu T cells. T + T cell–activation cell–activation T + T cell differen

differentiated + memory memory + + ------

I, VII, VIII, IX and X but a repressor of cluster IV, which suggests its its IV, cluster of suggests repressor a but X and which IX VIII, VII, I, clusters of activator an as T-bet identified we example, For clusters. multiple repress and/or activate to used potentially regulators many T cell activation into clusters of common kinetic expression, we found clusters I, VI, VIII, IX and X. By parsing genes expressed during CD8 which all may inhibit later E-protein activity, as predicted activators of Conversely, populations. memory Id2, we ZEB1 and identified ZEB2, IV and VIII, which include genes expressed in naive and late effector- as known T cells helper STAT3-dependant of CD4 differentiation exposure, particularly at the peak of infection, and is important for the ROR as notable, is which CD8 of regulator potential additional an as X) and cells effector dif ferentiated terminally with associated closely mostly clusters the of one is of Prdm1 products the identified we work, published with consistent example, For activation. cell T in involved networks regulatory and formation. and memory differentiation in roles important larly likely other receptor-ligand pairs in identified our analyses have simi T-bet factor transcription the inducing by phenotype ­effector r R The authors declare no competing interests.financial and A.W.G. studies, designed analyzed data and wrote the manuscript. analyzed data and contributed to writing the manuscript; E.Y. sorted cell subsets; manuscript; D.A.B., V.M. and M.L.D. anddesigned did early infection experiments, J.K. sorted cell subsets and analyzed data; A.D. analyzed data and edited the J.A.B. did experiments, studies, designed analyzed data and wrote the manuscript; (A.W.G.) and the Cancer Research Institute (A.W.G. and V.M.). and Infectious to Diseases) the ImmGen Consortium), the Pew Scholars program P30 CA016087 for cell sorting; and R24 AI072073 (National Institute of Allergy AI067545 to A.W.G.; T32 AI060536 to J.A.B.; PN2 EY016586 to D.A.B. and M.L.D.; ImmGen Project. Supported by the US National Institutes of Health (AI072117 and We and thank ExpressioneBioscience, Affymetrix Analysis for support of the Note: Supplementary information is available in the codes. Accession the in v available are references associated any and Methods M studies. future for platform a provides ated with different stages of CD8 the identification and establishment of canonical gene clusters associ CD8 of context the in investigated be to genes and pathways new identifying vation, have study a our of view established comprehensive CD8 transcriptional of results The response. immune the in points ous that control of the differentiation and effector T memory cells at vari and will facilitate the elucidation of complex networks transcriptional promise holds signatures transcriptional of regulators potential ing gene-expression signatures. Thus, we believe this strategy for identify late-memory and naive repress to also but programs differentiation not may only to and serve promote effector-memory activity effector C AUTH A e e eprints and permissions information is available online at at online available is information permissions and eprints ck O p et r Our analyses allowed a broad, unbiased look at gene-expression gene-expression at look unbiased broad, a allowed analyses Our r MPETI s i n i no o t h s O n / , , i ods n wledgme R R C

o Stat4 d f e N

Tcfe2a x t ON aDV G G FI . h h e t and and m 3 TRIBUTI

5 p N A l . Notably, we identified the E proteins E2A ( E2A E proteins the . Notably, we identified + . a NCE ONLINE PUBLIC ONLINE NCE A immunity. The data set of the ImmGen Project and and Project ImmGen the of set data immunity. The ) and E2-2 ( E2-2 and ) p GEO: microarray data, data, microarray GEO: N n e Tbx21 r CIAL CIAL I t . s ONS 3 as potential regulators of cluster IV, which cluster of regulators potential as N 4 . Our data identified ROR identified data Our . α TERE expression is upregulated after antigen antigen after upregulated is expression Tcf4 S + T T cell activation and differentiation ) as potential repressors of clusters clusters of repressors potential as ) S A TION + G o cells of cells the T

n S l i nature immunology nature E n e 1

v 5 e 9 r h 0 s t i 7 t o p n . α : +

/ o / (clusters I, VI VI I, (clusters effector cells, cells, effector w f

H t w h + 17 subset 17 of subset w e T cell acti

. p n Tcf3 a a p t u e 1 o r r 9 . e n ; also ; also ; it is is it ; . c Id2 l i o n m + e ------, /

© 2013 Nature America, Inc. All rights reserved. 18. 17. Boston, Massachusetts, Boston, USA. Massachusetts, USA. 15 Women’s Hospital, Boston, USA. Massachusetts, Stanford University, Stanford, California, USA. School, Worcester, USA. Massachusetts, USA. Massachusetts, Department, Brown University, Providence, Rhode Island, USA. 4 Diane Mathis Tracy Heng Claudia Jakubzick C Charlie Kim Patrick Brennan Jamie Knell Kutlu Elpek Brian Paul Monach ImmGen Project Consortium: 16. 15. 14. 13. 12. 11. 10. 9. 8. 7. 6. 5. 4. 3. 2. 1. nature immunology nature Department Department of Medicine, Boston University, Boston, USA. Massachusetts,

Icahn Icahn Medical Institute, Mount Sinai Hospital, New York, New York, USA.

Gautier, E.L. G.J. Windt, der van rk, K. Araki, E.L. Pearce, Dietz, Whitaker-Menezes, S.B., of role Lessin, The S.R. & D. C. Gründemann, increased edge: Cutting R. Ahmed, & J.D. Altman, A.J., Zajac, J.M., Grayson, H. Beuneu, TranscriptionH.D. ELF4 Lacorazza, factor & Yamada,M. T.,Mamonkin, C.S., Park, C. Haluszczak, cell transgenic receptor cell T Initial J.T. Harty, & J.S. V.P., Haring, Badovinac, and inflammation of roles Differing E.G. Pamer, & K.M. Kerksiek, D.H., Busch, CD8 Memory R. Ahmed, & S.M. Kaech, T.C. Wirth, S. Sarkar, profiling functional and Molecular R. Ahmed, & E. Hemby,Kersh, S.M., S., Kaech, W.N. Haining, gene of networks Project: Genome Immunological M.W.The Painter, T.S.& Heng, 1118–1128 (2012). 1118–1128 tissue macrophages. mouse diversity identityunderlie and of the CD8 of (2009). 108–112 metabolism. (1996). 312–318 andE-cadherin in epidermotropism in cutaneous T-cell lymphoma. G1. receptor lectin-like cell killer murine (2000). 3950–3954 CD8 memory antigen-specific in Bcl-2 of expression nodes. lymph in KLF2. and KLF4 CD8 of homing and proliferation the controls Med. Exp. J. expansion. homeostatic of markers bearing cells phenotype memory includes mice Immunity precursor frequency dictates critical aspects of the CD8 (2000). antigen in T cell proliferation and memory generation. (2001). encounter triggers a developmental program in naive cells. Immunity CD8 memory of signature core a preserves but diversification fates. memory distinct with differentiation. cell T CD8 memory of 181 signature of CD8 memory differentiation that is shared by T and B cells. cells. immune in expression 17 , 1859–1868 (2008). 1859–1868 , Joslin Joslin Diabetes Center, Boston, USA. Massachusetts, S + Garrison T cell memory development. memory cell T

t al. et 26 33 t al. et et al. et t al. et et al.

Nature , 827–841 (2007). 827–841 , (2010). 128–140 , 18 206 1 t al. et 10 t al. et TR euae mmr C8 -el differentiation. T-cell CD8 memory regulates mTOR et al. et , , Ananda W Goldrath ucinl n gnmc rfln o efco C8 cellsubsets T CD8 effector of profiling genomic and Functional , Michio Painter 4 et al. et Immunity Visualizing the functional diversification of CD8 of diversification functional the Visualizing Gene-expression profiles and transcriptional regulatory pathways that 18 , , Angelique Bellemare-Pelletier eeiie nie siuain nue sews transcriptome stepwise induces stimulation antigen Repetitive , 435–448 (2009). 435–448 , Nat. Immunol. Nat. , , 14 8

S & Christophe Benoist nacn C8 -el eoy y ouaig at acid fatty modulating by memory T-cell CD8 Enhancing Immune Immune Diseases Institute, Children’s Hospital, Boston, USA. Massachusetts, 460 et al. et Identification of an evolutionarily conserved transcriptional conserved an evolutionarily of Identification The antigen-specific CD8 antigen-specific The , Lewis L , Lanier Lewis usan usan A 13 Cutting edge: identification of E-cadherin as a ligand for the for ligand a as E-cadherin of identification edge: Cutting 8 15 , Derrick , J Derrick Rossi , , Michael Brenner , 103–107 (2009). 103–107 , Mitochondrial respiratory capacity is a critical regulator critical a is capacity respiratory Mitochondrial

, Gwendalyn J Randolph 33 J. Exp. Med. Exp. J. aDV Nat. Immunol. Nat. , 412–423 (2010). 412–423 ,

10 S A hinton NCE ONLINE PUBLIC ONLINE NCE , 618–626 (2009). 618–626 , Immunity Cell 10 18

Dana-Farber Dana-Farber Cancer Institute and Harvard Medical School, Boston, USA. Massachusetts, 205 J. Immunol. J.

+ , Jeffrey Ericson 111

cl dfeetain iiil antigen initial differentiation: cell T 9 + 5 , 1091–1094 (2008). 1091–1094 , 12 , 625–640 (2008). 625–640 , 14 T cells via the Kruppel-like factors Kruppel-like the via cells T , , Richard R Hardy + , 837–851 (2002). 837–851 ,

Broad Broad Institute, Cambridge, USA. Massachusetts, T cell repertoire in unimmunized in repertoire cell T 14 36 8 , , Jennifer Miller 1 Department Department of Microbiology & Immunology, University of California San Francisco, San Francisco, California, USA. , , Kavitha , , Vladimir Jojic , 68–78 (2012). 68–78 , 13 J. Immunol. + +

cells. T 18 176 T cell response to infection. , , Taras Kreslavsky Nat. Immunol. , 1311–1315 (2006). 1311–1315 , + α A T cell differentiation. cell T 18 E TION β J.Cutan. Pathol. Division Division of Immunology, Department of Microbiology & Immunobiology, Harvard Medical School, 7 integrin (CD103)7 15 a. Immunol.Nat. . Immunol. J.

+ 164 Department Department of Biomedical Engineering, Howard Hughes Medical Institute, Boston University, Boston, N T cell responses cell T 10 , 18 16 arayan

, 4063–4070 , Deepali Malhotra , Deepali , Nature

, Francis Kim 2 J. Immunol. S , 415–422 cott Davis 15 11 5 5

, , Radu Jianu 16 , , Brian Brown Fox Fox Chase Cancer Center, USA. Pennsylvania, Philadelphia, 164 460 , , Daphne Koller

9 Department Department of Pathology & Immunology, Washington University, St. Louis, Missouri, 13 23 , , Katelyn , , , ,

13 , , 21. 20. 19. 31. 30. 29. 28. 27. 26. 25. 24. 23. 22. 35. 34. 33. 32. N 18

ac, S.M. Kaech, R.L. Rutishauser, N.S. Joshi, Agarwal, P.Agarwal, A.L. Marzo, J.J. Obar, CD8 naive Endogenous L. Lefrancois, & K.M. Khanna, Obar,J.J., Harty,J.T.V.P. Badovinac, & CD8 reshaping and Shaping D’Cruz, L.M., Lind, K.C., Wu, B.B., Fujimoto, J.K. & Goldrath, A.W. Loss of E protein M.A. Cannarile, C.Y.Yang, Huster, K.M. mediates Interleukin-7 L. Kieper,Lefrancois, W.C.,K.S., & Schluns, S.C. Jameson, Sundrud, M.S. & Rao, A. Regulation of T helper 17 differentiation by orphan nuclear Kallies, A. Distinct regulation of effector and memory T-cell differentiation. of Modulation J. Zerrahn, & S.H. Kaufmann, U., F.,Klemm, P.,Kaiser, Geserick, D. Parolaro, P.,& P.,Sacerdote, A.E. Massi, Panerai, fetr D T el ta gv rs t ln-ie mmr cells. memory long-lived to rise give that cells T CD8 effector cell T memory central of acquisition properties. the represses and differentiation terminal (2007). 281–295 CD8 commitment. CD8 memory infection. to (2008). responses 859–869 memory and primary regulates frequency Immunol. CD8 the during formation memory-precursor delays HEB and E2A factors transcription Immunol. Nat. CD8 memory distinct of USA Sci. Acad. Natl. CD8 distinct of generation CD40L-dependent early cells T CD8 memory (2000). 426–432 and naive of homeostasis the (2003). 1191–1198 IFN in programming for CD8 T cell effector function and memory. receptors: it’s not just ROR just not it’s receptors: Biol. Cell motif. receptors. (2004). helicase-like RNA CB2 an gene (slfn) harbouring Schlafen and the family of members novel CB1 by activation and development cell both T of involvement rat: Neuroimmunol. the J. in the decreases 940 macrophages CP55, cannabinoid synthetic the with (2009). 1695–1704 atalie atalie A Bezman 17 , Ayla Ergun , Tata + + S T cellT fates via the graded expression ofT-bet transcription factor. T-cell immune response. T-cellimmune 6 ylvia , , David Koller 10 15

86 et al. et 8 t al. et 13 , , , , Miriam Merad Immunity , 107–119 (2008). 107–119 , et al. et t al. et N 11 t al. et , 325–332 (2008). 325–332 , + S et al. 9 Division Division of Rheumatology, Immunology and Allergy, Brigham and 9 t al. et T cell differentiation. cell T Nat. Immunol. Nat. Department Department of Pathology, University of Medical Massachusetts

hannon Turley ageswara Rao , , Joonsoo Kang , , Tal 7 Pathogen-induced inflammatory environment controls effector and effector controls environment inflammatory Pathogen-induced et al. et h tasrpinl euaos d ad d cnrl h formation the control Id3 and Id2 regulators transcriptional The , 1317–1325 (2006). 1317–1325 , Gene regulation and chromatin remodeling by IL-12 and type I type and IL-12 by remodeling chromatin and regulation Gene Inflammation directs memory precursor and short-lived effector andshort-lived precursor memory directs Inflammation

Selective expression of IL-7 receptor on memory T cells identifies nta T el rqec dcae mmr CD8 memory dictates frequency cell T Initial t al. et 109 18 eetv epeso o te nelui 7 eetr identifies receptor 7 interleukin the of expression Selective

Transcriptional regulator Id2 mediates CD8 Transcriptionalmediates Id2 regulator

31 , Michael Mingueneau 101 S , 155–163 (2000). 155–163 , + rncitoa rpesr lm- pooe CD8 promotes Blimp-1 repressor Transcriptional hay , 296–308 (2009). 296–308 , T cell subsets. cell T 14 , 5610–5615 (2004). 5610–5615 , γ t anymore. t

, , Joseph C 6 6 12 , , Jim Collins , 793–799 (2005). 793–799 , Eur. J. Immunol. J. Eur. , , Aviv Regev 17 15 J. Immunol. J. 10 , Amy Wagers 9 , , Emmanuel L Gautier Immunity , , Anne Fletcher , , J Adam Best Nat. Immunol. Nat. S 11 un

Computer Computer Science Department,

8

7 n. Immunol. Int. 187 28 , 2031–2041 (2012). 2031–2041 , 6 14 + , , Roi Gazit Computer Computer Science 12 memory T cell subsets. cell T memory , 5–7 (2008). 5–7 , In vivo In , , 4967–4978 (2011). 4967–4978 , , , n vivo in 18

N + 12 e c r u o s e r T-cell memory. , adia adia Cohen 17

, 1221–1229 (2011). 1221–1229 , and n vitro in 1 , . ,

a. Immunol. Nat. 10 + a. Immunol. Nat.

+ J. Immunol. in vitro in 16 T cell immunity.cell T T cell precursor cell T + , cl lineage cell T 8 1535–1548 , Immunity

irto of migration , Immunity

treatment 15 Immunol. Nat. Rev.Nat. + cell T , 16

13 Proc. 183

28 27

,

, 1 4  , , , , ,

© 2013 Nature America, Inc. All rights reserved. (A7R34) (A7R34) and anti-KLRG1 (2F1). Antigen-specific CD8 anti- CD45.2 (104), anti-CD62L (MEL-14), anti-CD122 (TM-b1), anti-CD127 CD8; anti-CD27 anti-CD44 53-6.7), (LG-7F9), (IM7), (A20-1.7), anti-CD45.1 (anti- CD8 to antibody eBioscience): from (all antibodies following the used CD8 org/Protoc ( website on Project protocol the ImmGen to the sorting cytometry. and flow sorting Cell transfer. 2 d after mice from purified were and mice C57BL/6J into recipients. To obtain naive OT-I cells, 5 × 10 activation, 1 × 10 OVA or 1 × 10 response, C57BL/6J mice were infected with 5 × 10 days 5, 6 and 8 of For infection. evaluation of the endogenous polyclonal T cell to VSV-OVAfaster respond on to Lm-OVA,were collected than cells effector Lm-OVA or 5 × 10 after d 1 Then, mice. T recipient C57BL/6J into transferred were cells OT-I infection. and transfer Cell Diego. San California, of University the of Guidelines Use and Care Animal accordance Institutional in the with conditions pathogen–free specific in housed and bred were CD45.1 at weeks. 6 beginning use experimental before d 7–10 for conditions pathogen–free Mice. MET ONLINE nature immunology nature

cell transfer, mice were infected with either 5 either × with 10 transfer, were cell mice infected + Male C57BL/6J mice (Jackson Laboratories) were housed in specific specific in housed were Laboratories) (Jackson mice C57BL/6J Male T cells from single-cell splenocyte suspensions to assess phenotype phenotype assess to suspensions splenocyte single-cell from cells T ols/ImmGen Cell prep Cell ols/ImmGen 5 plaque-forming units plaque-forming VSV. To obtain cells 12, 24 and 48 h after 6 H + CD45.2 3 OT-I mice deficient in recombination-activating gene 1 gene recombination-activating in deficient mice OT-I ODS plaque-forming units of VSV-OVA. As OVA-specific T + OT-I cells were transferred into CD45.1 For days 5–100 after infection, 5 × 10 × 5 infection, after 5–100 For days Cells were purified and analyzed according according and analyzed were purified Cells and sorting SOP.pd sorting and 6 CD45.1 3 colony-forming units Lm- 3 + colony-forming units of units colony-forming OT-I cells were injected + f T cells were identified ). Flow cytometry of of cytometry Flow ). http://w ww.immgen. + 3 C57BL/6J CD45.1

cells +

eainhp ih rmr fr hueepn’ ee. eut wr normal were Results genes. ‘housekeeping’ linear for a primers in with amplify that relationship primers with Stratagene) Green; (SYBR detection product nonspecific with PCR quantitative by abun assessed The was mRNA of dance (Invitrogen). kit III SuperScript with generated was cDNA and with TRIzol extracted reagent (Invitrogen) and treated with DNAse (Ambion) PCR. Quantitative <0.5). variation, of (coefficient were by expressed at one least they and cell type with low in populations variability if only comparisons in included were probes and confidence), 95% (at of threshold 120 expression was to taken positive indicate post-normalization Project ImmGen general The quality. replicate good with Project ImmGen of the criteria quality-control MoGene 1.0 ST passed All data arrays. analyzed Affymetrix on generated were profiles gene-expression pipelines, ity-control qual and profiling ImmGen the With analysis. microarray for used were R GenePattern suite of genomic analysis software and the statistical environment was amplified and hybridized to the Mouse Affymetrix Gene 1.0 ST Array. The RNA reagent. TRIzol in Lm-OVAVSV-OVA or with prepared was infection analysis. microarray for CD8 preparation Sample with (TreeStar). analyzed software FlowJo were data and Biosciences) (BD FACSAria a or on FACSCalibur collected were Samples 780. Fluor Alexa or allophycocyanin erythrin, phyco isothiocyanate, fluorescein to conjugated were Antibodies Coulter). H-2K of tetramer a with ized to the expression of transcripts encoding GAPDH (glyceraldehyde (glyceraldehyde GAPDH encoding transcripts dehydrogenase). phosphate of expression the to ized + CD45.1 + cells (pooled from three mice) at various time points during during points time various at mice) three from (pooled cells Donor cells were sorted as described above. RNA was was RNA above. described as sorted were cells Donor b and OVA peptide (sequence, SIINFEKL; Beckman Beckman SIINFEKL; (sequence, OVA and peptide RNA obtained from from obtained RNA doi:10.1038/ni.2536 - - - -