LETTERS

The AP-1 transcription factor Batf controls TH17differentiationBarbara U. Schraml1*, Kai Hildner1,2*, Wataru Ise1,2, Wan-Ling Lee1, Whitney A.-E. Smith1, Ben Solomon1,Gurmukh Sahota3, Julia Sim4, Ryuta Mukasa5, Saso Cemerski1, Robin D. Hatton5, Gary D. Stormo3,Casey T. Weaver5, John H. Russell4, Theresa L. Murphy1 & Kenneth M. Murphy1,2

Activator protein 1 (AP-1, also known as JUN) transcription fac-tors are dimers of JUN, FOS, MAF and activating transcriptionfactor (ATF) family proteins characterized by basic region andleucine zipper domains1. Many AP-1 proteins contain definedtranscriptional activation domains, but BATF and the closelyrelated BATF3 (refs 2, 3) contain only a basic region and leucinezipper, and are considered to be inhibitors of AP-1 activity3–8. Herewe show that Batf is required for the differentiation of IL17-pro-ducing T helper (TH17) cells9. TH17 cells comprise a CD41 T-cellsubset that coordinates inflammatory responses in host defencebut is pathogenic in autoimmunity10–13. Batf2/2 mice have normalTH1 and TH2 differentiation, but show a defect in TH17 differenti-ation, and are resistant to experimental autoimmune encephalo-myelitis. Batf2/2 T cells fail to induce known factors required forTH17 differentiation, such as RORct11 (encoded by Rorc) and thecytokine IL21 (refs 14–17). Neither the addition of IL21 nor theoverexpression of RORct fully restores IL17 production in Batf2/2

T cells. The Il17 promoter is BATF-responsive, and after TH17differentiation, BATF binds conserved intergenic elements inthe Il17a–Il17f locus and to the Il17, Il21 and Il22 (ref. 18) promo-ters. These results demonstrate that the AP-1 protein BATF has acritical role in TH17 differentiation.

In a gene expression survey (Supplementary Fig. 1a), we identifiedthe basic leucine zipper transcription factor ATF-like7 Batf to behighly expressed in TH1, TH2 and TH17 cells compared to naive Tand B cells. BATF and BATF3 (refs 2, 3) form heterodimers withJUN6,7 and are considered to be repressors of AP-1 activity3,5,6,8,19.To assess the role of BATF in T cell differentiation20, we generatedBatf2/2 mice (Supplementary Fig. 2a, b). Batf2/2 mice lacked detect-able BATF protein, were fertile and appeared healthy. BATF proteinwas low in naive T cells, increased in TH2 cells, induced by activation(Supplementary Fig. 2), present in the nucleus and cytoplasm, butshowed increased nuclear translocation after activation (Fig. 1a andSupplementary Fig. 1b, c). Batf2/2 mice had normal thymus, spleenand lymph node development, and normal CD41 and CD81 T celldevelopment (Supplementary Figs 3, 4a, b). Although Batf-transgenicmice had altered natural killer T cell development21, Batf2/2 mice hadnormal development of natural killer T cells (Supplementary Fig. 4c),B cells (Supplementary Fig. 4d, e), and conventional and plasma-cytoid dendritic cells (Supplementary Fig. 5a, b).

Batf2/2 T cells showed normal TH1 and TH2 differentiation(Supplementary Fig. 6a). Under TH17 conditions, Batf2/2 T cells,but not Batf1/2 T cells, showed a marked reduction in IL17 produc-tion, but had normal levels of IL2, IFN-c and IL10 (Fig. 1b, c).

Batf2/2 DO11.101 T cells showed loss of IL17 even after severalpassages under TH17 conditions (Supplementary Fig. 6b). Batf2/2

CD81 T cells also failed to produce IL17 (Supplementary Fig. 6c). Wegenerated transgenic mice expressing Flag-tagged Batf under thecontrol of the Cd2 promoter22. Batf-transgenic DO11.101 CD41 Tcells and CD81 T cells had increased IL17 production under TH17conditions compared to controls (Supplementary Fig. 6d, e). Laminapropria CD41 T cells, which constitutively express IL17 in wild-typemice11, failed to produce IL17 in Batf2/2 mice (SupplementaryFig. 6f).

TH17 cells are the major pathogenic population in experimentalautoimmune encephalomyelitis10 (EAE), although factors other thanIL17A and IL17F can contribute to the disease23. Batf1/1 mice immu-nized with myelin oligodendrocyte glycoprotein peptide 35–55(MOG(35–55); Fig. 2) developed EAE, but Batf2/2 mice were com-pletely resistant (Fig. 2a). At peak disease, central nervous system(CNS)-infiltrating and splenic CD41 T cells from Batf1/1 mice pro-duced abundant IL17 and IFN-c, whereas T cells from Batf2/2 micedid not produce IL17 (Fig. 2b and Supplementary Fig. 7a). BecauseIL6-deficient mice are resistant to EAE owing to a compensatoryincrease in FOXP31 T regulatory (Treg) cells14, we analysed splenicBatf1/1 and Batf2/2 CD41 T cells for FOXP3 expression before andafter MOG(35–55) immunization (Supplementary Fig. 7b, c).Batf2/2 mice had lower basal numbers of splenic FOXP31 T cellscompared to Batf1/1 mice, but showed no change in FOXP31

expression after MOG(35–55) immunization (Supplementary Fig.7b, c), suggesting that their resistance to EAE is not due to an increasein Treg cells. To determine whether the resistance to EAE in Batf2/2

mice resulted from a defect within T cells or other immune cells, weinjected naive Batf1/1 CD41 T cells or PBS control buffer into micebefore MOG(35–55) immunization (Fig. 2c). Batf2/2 mice receivingPBS remained resistant to EAE, but Batf2/2 mice receiving naiveBatf1/1 CD41 T cells developed severe EAE (Fig. 2c and Supplemen-tary Table 1) with CNS-infiltrating IL17-producing CD41 T cells(Supplementary Fig. 7d). Thus, Batf2/2 mice have a T-cell-intrinsicdefect preventing EAE.

Batf could control TH17 development by regulating IL6 or TGF-bsignalling. IL6 receptor expression and IL6-induced STAT3 phos-phorylation were normal in Batf2/2 T cells (Supplementary Fig. 8a,b). TGF-b induced normal levels of FOXP3 in Batf2/2 CD41 T cells(Supplementary Fig. 8d). Whereas Batf2/2 T cells failed to fullydownregulate FOXP3 in response to IL6 (ref. 12), neutralization ofIL2 abrogated increased FOXP3 in Batf2/2 T cells, without restoringIL17 production (Supplementary Fig. 8d, e). Thus, Batf2/2 T cells

*These authors contributed equally to this work.

1Department of Pathology and Immunology, 2Howard Hughes Medical Institute, 3Department of Genetics, 4Department of Molecular Biology and Pharmacology, WashingtonUniversity School of Medicine, 660 South Euclid Avenue, Saint Louis, Missouri 63110, USA. 5Department of Pathology, University of Alabama at Birmingham, University Station,Birmingham, Alabama 35294, USA.

Vol 460 | 16 July 2009 | doi:10.1038/nature08114

405 Macmillan Publishers Limited. All rights reserved©2009

exhibit normal TGF-b signalling and proximal IL6 signalling, indi-cating Batf may regulate downstream target genes.

IL21, an early target of IL6 signalling in CD41 T cells17, is requiredfor TH17 development14–16. IL21 was reduced in Batf2/2 CD41 T cellsactivated under TH17 conditions (Fig. 3a). The addition of IL21failed to rescue TH17 development in Batf2/2 T cells (Fig. 3b), butIL21-induced STAT3 phosphorylation was intact (Supplementary

Fig. 8c), suggesting that Batf regulates other factors besides IL21during TH17 differentiation.

We performed DNA microarrays and quantitative PCR withreverse transcription (qRT–PCR) of Batf1/1 and Batf2/2 T cellsactivated with combinations of IL6 and/or TGF-b (Fig. 3c, d andSupplementary Fig. 9). This analysis identified several genes knownto regulate TH17 development as Batf-dependent (Fig. 3c, d,Supplementary Fig. 9c and Supplementary Table 2), includingRorc17, Rora24, the aryl hydrocarbon receptor (Ahr)25,26, Il22 (ref.18) and Il17. However, Irf4 (ref. 13) and suppressor of cytokinesignalling (Socs1–7) gene expression were unchanged in Batf2/2 Tcells (Supplementary Fig. 9b and Supplementary Table 4). Earlyinduction of RORct was normal in Batf2/2 T cells, but was notmaintained 62 h after stimulation (Supplementary Fig. 11a). Batfseemed to be necessary for the expression of a subset of IL6-inducedgenes, but was not required for the expression of TGF-b-inducedgenes (Fig. 3c, Supplementary Fig. 9a and Supplementary Tables 2

IFN

-γ

IL17

Batf –/–Batf +/–Batf +/+

8 0.1

91 0.8

0

103

104

105 3 0.3

79 17

3 0.3

84 13

100

101

102

103

1 0.1

88 11

100

101

102

103

1 0.1

98 0.5

52 9

37 2

60 0.5

39 0.3IL2

2 0

98 0.2

2 0.1

91 7

IL10

IL17

IFN

-γ

TH17

c

DIC

BATFDAPI

CD4BATF

CD4DAPI

TG PMA + iono WT PMA + ionoTG unstima

b

100 101 102 103 100 101 102 103 102100 103 104 105

104 10510300 0104 105103 104 105

Batf +/+

Batf –/–

103

Figure 1 | Loss of IL17 production in Batf2/2 T cells. a, DO11.101 CD41 Tcells from CD2-N-Flag-Batf transgenic mice or littermate controls werecultured with ovalbumin (OVA) and antigen-presenting cells (APCs) underTH2 conditions for 7 days, and stained with antibodies to CD4 and Flag.DAPI, 4,6-diamidino-2-phenylindole; DIC, differential interferencecontrast; iono, ionomycin; TG, transgenic, unstim, unstimulated; WT, wildtype. b, Batf1/1 and Batf2/2 CD41 CD62L1 CD252 T cells cultured underTH17 conditions were restimulated with PMA and ionomycin on days 7 (leftpanel) or 3 (middle and right panels), and stained for IL17, IFN-c, IL2 andIL10. c, IL17 and IFN-c expression in DO11.101 CD41 T cells from Batf1/1,Batf 1/2 and Batf2/2 mice activated with OVA and APCs under TH17conditions. Data are representative of at least two independent experiments.

a

b

c

53 3 45 12 84 0.5IFN

-γ

IL17

Batf +/+

(score 5)Batf +/+

(score 3)Batf –/–

(score 0)

Batf +/+ CD4+ Batf +/+

Batf +/+CD4+ Batf –/–

PBS Batf +/+

PBS Batf –/–

Time after immunization (days)

Clin

ical

sco

re

10 15 20 25 30 35 40

Time after immunization (days)

Clin

ical

sco

re

10 15 20 25 30

0

1

2

3

4 Batf +/+

Batf –/–

39 5 26 17 14 1

100 101 102 103 100 101 102 103 100 101 102 103100

101

102

103

0

1

2

3

4

Figure 2 | Batf2/2 mice are resistant to EAE. a, Batf1/1 (n 5 12) andBatf2/2 (n 5 13) mice were immunized with MOG(33–35) peptide. Meanclinical EAE scores 6 s.e.m are shown, and are representative of twoindependent experiments. b, Thirteen days after EAE induction, CNS-infiltrating lymphocytes were stimulated with PMA and ionomycin, gatedon CD41 cells, and stained for intracellular IL17 and IFN-c (clinical scoresare in parentheses, data are representative of 2–3 mice per group). c, Batf1/1

and Batf2/2 mice were injected with control PBS buffer (n 5 5) or with1 3 107 Batf1/1 CD41 T cells (n 5 6) 4 days before EAE induction. Meanclinical scores are shown.

LETTERS NATURE | Vol 460 | 16 July 2009

406 Macmillan Publishers Limited. All rights reserved©2009

and 3). However, Batf did not globally affect IL6-induced responses,because IL6-induced liver acute phase responses appeared normal inBatf2/2 mice (Supplementary Fig. 10).

Because RORct acts directly on the Il17 promoter27,28, weaddressed whether RORct could rescue TH17 development inBatf2/2 T cells. In Batf1/1 T cells, retroviral RORct expressioninduced 38% IL17 production, compared to only 1.6% IL17 produc-tion induced by control retrovirus (Fig. 3e and SupplementaryFig. 11c)11,13. However, in Batf2/2 T cells, retroviral RORct expres-sion induced only 5.7% IL17 production (Fig. 3e and SupplementaryFig. 11c). Even under TH17-inducing conditions, retroviral RORctexpression did not fully restore IL17 production in Batf2/2 T cells

(Supplementary Fig. 11b, c). Retroviral expression of both BATF andRORct in Batf2/2 T cells induced 26% IL17 production, compared toonly 5% with RORct alone and 14% with BATF alone(Supplementary Fig. 11d), suggesting that there is potential synergybetween RORct and BATF, and a possible direct action of BATF intranscription of Il17 and other TH17-specific genes.

We used a reverse-strand retroviral reporter29 to examine Il17promoter activity in primary Batf1/1 and Batf2/2 T cells (Fig. 4a).Three days after activation, Batf2/2 CD41 T cells showed consid-erably less reporter activity than Batf1/1 T cells, suggesting that theproximal Il17 promoter is Batf-responsive (Fig. 4a). Using chromatinimmunoprecipitation (ChIP) analysis of several conserved regionswithin the Il17a–Il17f locus (Supplementary Fig. 12a), we found thatBATF specifically bound to the 19.6 kilobase (kb) and 128 kb inter-genic regions within 24 h after activation (Fig. 4b and SupplementaryFig. 12b, c). By day 5 after stimulation, BATF bound specifically toseveral intergenic regions and to the proximal Il17a and Il17f pro-moters (Fig. 4b and Supplementary Fig. 12b, c), with distal elementsshowing more rapid and stronger binding than proximal elements.

T H17

TGF-β +

anti-

IL6

IL6 +

anti-

TGF-β

Anti-I

L6 +

anti-

TGF-βT H

17

TGF-β +

anti-

IL6

IL6 +

anti-

TGF-β

Anti-I

L6 +

anti-

TGF-β

Anti-IL2

0.2

80.1

94 0.3

4 0.1

88 8

10510410310200 105

105

104

104

103

0 105104103 0 105104103

103

102

1051041031020 1051041031020 1051041031020

102

0

105

104

103

102

0

105

104

103

102

0

105

104

103

0

105

104

103

102

0

6 0.1

94 0.2

TGF-β + IL6TGF-β + IL6

+ IL21

IFN

-γ

IL17

a

Batf +/+

Batf –/–

0

25

50

75

100

Il21

rela

tive

expr

essi

on

b

d

cRorcRoraAhrIl17Il22

Expression3 –3

0.2 0.9

22 77

3 21

29 470.6 0.3 0.3 0.6 0.4 4

1 0.1

99 0

GFP

IL17

Uninfected Empty-GFP-RV RORγt-GFP-RV

Ant

i-IL

4/an

ti-IL

12/a

nti-

IFN

-γ

Batf +/+

Batf –/–

e

01,0002,0003,0004,0005,000

<1,000IL21

(pg

ml–1

)

Batf +

/+

Batf –/–

Batf +

/+

Batf –/–

IL22

IL17

0.8 0.1

99 0.5

1.5 4

64 31

Batf +/+ Batf –/–

Batf +/+ Batf –/–

99 0 26 73 32 64

Figure 3 | BATF controls several TH17-associated genes. a, IL21 expressionin Batf1/1 or Batf2/2 T cells cultured under TH17 conditions determined byqRT–PCR and ELISA. The mean and s.d. are shown, from 3 mice. b, IL17 andIFN-c expression of CD41 CD62L1 CD252 T cells cultured as in a in thepresence or absence of IL21. c, Microarray analysis of anti-CD3/CD28-activated T cells at 72 h, presented as heat maps of genes fivefold-induced inBatf1/1 T cells under TH17 conditions. d, IL17 and IL22 expression inBatf1/1 or Batf2/2 CD41 T cells activated under TH17 conditions for 3 days.e, Anti-CD3/CD28-activated Batf1/1 or Batf2/2 CD41 T cells were leftuninfected or infected with retrovirus (RV) RORct-GFP-RV or with thecontrol empty-GFP-RV, and stained for IL17. GFP, green fluorescentprotein.

c

e

WT

BATFAP-1

KO IL17RORE

1

Ig

2 543

Anti-Flag

6 7 8 11109 12

WT

TG WT

TG WT

TG

AP-1 probe

WT

KO

Anti-BATF

WT

KO WT

KO

PMA + iono

BATF

Anti-Flag

WT

TG WT

TGIL17RORE

WT

KO

IL17 probe (–155 to –187)

1 2 543 6 7 8 11109

1 2 543 6 7 8 109

12 1413

Anti-BATF

WT

KO

d

BATF/JUN

WT

KO Anti-B

ATF

Anti-F

OS

Anti-J

UN

Anti-c

-JUN

Anti-J

UNB

Anti-J

UND

Anti-A

TF1

Anti-A

TF3

AP-1 probe

FOS/JUN

–5 kb

–243

to

–176

+9.6

kb

+23 kb

+28 kb

36 kb

–408

to

–340 –7

kb

Il17fIl17a

Batf +/+ unstimBatf +/+ PMA + ionoBatf –/– unstimBatf –/– PMA + iono

Rel

ativ

e bi

ndin

g

Il17a promoter

Empty-GFP

GFP

Per

cent

age

of m

ax

0

20

40

60

80

100hCD4+

0.00

0.05

0.10

0.15

a bBatf +/+

103100 102102

Batf –/–

Figure 4 | BATF directly regulates IL17 expression. a, Batf1/1 and Batf2/2

CD41 T cells cultured under TH17 conditions were infected with hCD4-pA-GFP-RV-IL17p reporter virus. GFP expression after PMA and ionomycinrestimulation is shown. b, Batf1/1 and Batf2/2 CD41 T cells cultured underTH17 conditions for 5 days were subjected to ChIP analysis of the indicatedregions using an anti-BATF antibody (mean and s.d.). c–e, EMSA supershiftanalysis of TH17 whole cell extracts using a consensus AP-1 (c, e) or the IL17(2155 to 2187) (d) probe. Batf1/1 (wild type, WT), Batf2/2 (knockout,KO), CD2-N-Flag-Batf transgenic (TG), IL17 (2155 to 2187) and retinoic-acid-receptor-related orphan receptor (ROR) response element (RORE)probes were used as competitors. Anti-JUN denotes antibodies withpan-JUN specificity.

NATURE | Vol 460 | 16 July 2009 LETTERS

407 Macmillan Publishers Limited. All rights reserved©2009

We next examined BATF binding to a consensus AP-1 probe6 byelectrophoretic mobility shift assays (EMSAs). This probe formedtwo complexes in Batf1/1 TH17 cell extracts (Fig. 4c) that weredependent on stimulation (Supplementary Fig. 13a). Only the uppercomplex formed in Batf2/2 TH17 cells (Fig. 4c). An anti-BATF anti-body inhibited the lower complex. In CD2-N-Flag-Batf-transgenicTH17 cell extracts, the lower complex was specifically supershifted byan anti-Flag antibody (Fig. 4c). Thus, only the lower complex bindingthe consensus AP-1 probe in TH17 cells contains BATF.

Several potential BATF-binding sites were identified by EMSAs inthe Il17, Il21 and Il22 proximal promoters, including the Il17 pro-moter region (2188 to 2210) that bound BATF in ChIP (Fig. 4b andSupplementary Fig. 13b–d). Another BATF-binding Il17 promoterregion (2155 to 2187) overlapped with a reported RORct-bindingelement27. As an EMSA probe, this region forms two complexes inTH17 cells (Fig. 4d), with the lower complex being selectively inhib-ited by an anti-BATF antibody, absent in Batf2/2 TH17 cells, andsupershifted by an anti-Flag antibody in Batf-transgenic TH17extracts (Fig. 4d). We confirmed BATF binding to the Il21 and Il22promoters by ChIP analysis (Supplementary Fig. 13e). The programCONSENSUS30 determined that the BATF-binding element in theIl17, Il21 and Il22 promoters resembles canonical AP-1 elements atpositions 1 to 3, with variation at the remaining nucleotides(Supplementary Fig. 13f). CONSENSUS did not identify other tran-scription-factor-binding sites enriched near BATF-binding elements.We determined the composition of the BATF-containing complexusing supershift analysis (Fig. 4e). The upper complex supershiftedwith a pan-anti-FOS antibody, whereas the lower complex super-shifted with a pan-anti-JUN and anti-BATF antibodies. Anti-JUNBsupershifted most of the lower complex, but antibodies against JUN(also known as c-JUN), JUND, ATF1 or ATF3 did not. Thus, BATFpreferentially forms heterodimers with JUNB during TH17 differ-entiation.

Although Batf and Batf3 were considered to be AP-1 inhibitors3–8,we have shown that they are required for the development of specificimmune lineages2. Batf is selectively required for TH17 development,but unlike Irf4 (ref. 13), it is not required for TH2 development.Because Batf is also expressed in TH1 and TH2 cells, it probablycooperates with other TH17-specific factors to regulate target genes.Future work will determine whether the actions of BATF involvedistinct DNA binding specificity or unique protein–protein interac-tions with TH17 specific factors.

METHODS SUMMARYMice. Batf2/2 mice were generated by homologous recombination, deleting

exons 1 and 2 of the Batf gene on the pure 129SvEv genetic background. The

neomycin-resistance cassette was removed from the targeted Batf allele in

embryonic stem cells before the generation of mice.

T cell differentiation assays. Naive CD41 CD62L1 CD252 T cells (also known

as CD41 SELL1 IL2RA2) were isolated by cell sorting and activated with plate-

bound anti-CD3 (also known as CD3E) and soluble anti-CD28 antibodies.

Cultures were supplemented with anti-IL4 (11B11; hybridoma supernatant),

IFN-c (Peprotech; 0.1 ng ml21) and IL12 (Genetics Institute; 10 U ml21) for

TH1; anti-IFN-c (H22; gift from R.D. Schreiber; 10 mg ml21), anti-IL12 (Tosh;

BioXcell; 10 mg ml21) and IL4 (Peprotech; 10 ng ml21) for TH2; anti-IL4, anti-

IL12, anti-IFN-c, IL6 (Peprotech; 20 ng ml21) and TGF-b (Peprotech;

0.5 ng ml21) for TH17 differentiation. Unless otherwise indicated, 3 days after

activation cells were restimulated with phorbol myristate acetate (PMA) and

ionomycin for 4 h for intracellular cytokine analysis by flow cytometry.

Intracellular staining. For intracellular cytokine staining, cells were stained for

surface markers followed by fixation with 2% formaldehyde for 15 min at room

temperature. Cells were then washed once in 0.05% saponin and stained with

anti-cytokine antibodies in 0.5% saponin. Anti-phospho-STAT3 antibody (BD

Pharmingen) was used according to the manufacturer’s recommendations. In

brief, cells were stained for surface markers followed by fixation with 90% meth-

anol at 220 uC overnight. Cells were then washed and stained for phosphory-

lated-STAT3 in PBS containing 3% FCS. FOXP3 staining was performed

according to the manufacturer’s recommendations using FOXP3 staining buf-

fers (eBioscience).

Induction of EAE. Mice (7–10 weeks old) were immunized subcutaneously with

100mg MOG(35–55) peptide (Sigma) emulsified in complete Freund’s adjuvant

(incomplete Freund’s adjuvant supplemented with 500mg Mycobacterium

tuberculosis). One and three days later, mice were given 300 ng Pertussis Toxin

(List Biological Laboratories) intraperitoneally. Clinical scores were assessed

as described in Methods. For T-cell transfer experiments, mice were injectedwith either PBS or 107 Batf1/1 CD41 T cells 4 days before MOG(35–55)

immunization13.

Full Methods and any associated references are available in the online version ofthe paper at www.nature.com/nature.

Received 9 April; accepted 5 May 2009.Published online 5 July 2009.

1. Wagner, E. F. & Eferl, R. Fos/AP-1 proteins in bone and the immune system.Immunol. Rev. 208, 126–140 (2005).

2. Hildner, K. et al. Batf3 deficiency reveals a critical role for CD8a1 dendritic cells incytotoxic T cell immunity. Science 322, 1097–1100 (2008).

3. Iacobelli, M., Wachsman, W. & McGuire, K. L. Repression of IL-2 promoteractivity by the novel basic leucine zipper p21SNFT protein. J. Immunol. 165,860–868 (2000).

4. Blank, V. Small Maf proteins in mammalian gene control: mere dimerizationpartners or dynamic transcriptional regulators? J. Mol. Biol. 376, 913–925 (2008).

5. Williams, K. L. et al. Characterization of murine BATF: a negative regulator ofactivator protein-1 activity in the thymus. Eur. J. Immunol. 31, 1620–1627 (2001).

6. Echlin, D. R., Tae, H. J., Mitin, N. & Taparowsky, E. J. B-ATF functions as a negativeregulator of AP-1 mediated transcription and blocks cellular transformation byRas and Fos. Oncogene 19, 1752–1763 (2000).

7. Dorsey, M. J. et al. B-ATF: a novel human bZIP protein that associates withmembers of the AP-1 transcription factor family. Oncogene 11, 2255–2265 (1995).

8. Thornton, T. M., Zullo, A. J., Williams, K. L. & Taparowsky, E. J. Directmanipulation of activator protein-1 controls thymocyte proliferation in vitro. Eur. J.Immunol. 36, 160–169 (2006).

9. Harrington, L. E. et al. Interleukin 17-producing CD41 effector T cells develop via alineage distinct from the T helper type 1 and 2 lineages. Nature Immunol. 6,1123–1132 (2005).

10. Langrish, C. L. et al. IL-23 drives a pathogenic T cell population that inducesautoimmune inflammation. J. Exp. Med. 201, 233–240 (2005).

11. Ivanov, I. I. et al. The orphan nuclear receptor RORct directs the differentiationprogram of proinflammatory IL-171 T helper cells. Cell 126, 1121–1133 (2006).

12. Bettelli, E. et al. Reciprocal developmental pathways for the generation ofpathogenic effector TH17 and regulatory T cells. Nature 441, 235–238 (2006).

13. Brustle, A. et al. The development of inflammatory TH17 cells requires interferon-regulatory factor 4. Nature Immunol. 8, 958–966 (2007).

14. Korn, T. et al. IL-21 initiates an alternative pathway to induce proinflammatoryTH17 cells. Nature 448, 484–487 (2007).

15. Nurieva, R. et al. Essential autocrine regulation by IL-21 in the generation ofinflammatory T cells. Nature 448, 480–483 (2007).

16. Wei, L., Laurence, A., Elias, K. M. & O’Shea, J. J. IL-21 is produced by Th17 cells anddrives IL-17 production in a STAT3-dependent manner. J. Biol. Chem. 282,34605–34610 (2007).

17. Zhou, L. et al. IL-6 programs TH17 cell differentiation by promoting sequentialengagement of the IL-21 and IL-23 pathways. Nature Immunol. 8, 967–974 (2007).

18. Liang, S. C. et al. Interleukin (IL)-22 and IL-17 are coexpressed by Th17 cells andcooperatively enhance expression of antimicrobial peptides. J. Exp. Med. 203,2271–2279 (2006).

19. Bower, K. E., Fritz, J. M. & McGuire, K. L. Transcriptional repression of MMP-1 byp21SNFT and reduced in vitro invasiveness of hepatocarcinoma cells. Oncogene 23,8805–8814 (2004).

20. Hess, J., Angel, P. & Schorpp-Kistner, M. AP-1 subunits: quarrel and harmonyamong siblings. J. Cell Sci. 117, 5965–5973 (2004).

21. Williams, K. L. et al. BATF transgenic mice reveal a role for activator protein-1 inNKT cell development. J. Immunol. 170, 2417–2426 (2003).

22. Zhumabekov, T., Corbella, P., Tolaini, M. & Kioussis, D. Improved version of ahuman CD2 minigene based vector for T cell-specific expression in transgenicmice. J. Immunol. Methods 185, 133–140 (1995).

23. Haak, S. et al. IL-17A and IL-17F do not contribute vitally to autoimmune neuro-inflammation in mice. J. Clin. Invest. 119, 61–69 (2009).

24. Yang, X. O. et al. T helper 17 lineage differentiation is programmed by orphannuclear receptors RORa and RORc. Immunity 28, 29–39 (2008).

25. Veldhoen, M. et al. The aryl hydrocarbon receptor links TH17-cell-mediatedautoimmunity to environmental toxins. Nature 453, 106–109 (2008).

26. Quintana, F. J. et al. Control of Treg and TH17 cell differentiation by the arylhydrocarbon receptor. Nature 453, 65–71 (2008).

27. Ichiyama, K. et al. Foxp3 inhibits RORct-mediated IL-17A mRNA transcriptionthrough direct interaction with RORct. J. Biol. Chem. 283, 17003–17008 (2008).

28. Zhang, F., Meng, G. & Strober, W. Interactions among the transcription factorsRunx1, RORct and Foxp3 regulate the differentiation of interleukin 17-producing Tcells. Nature Immunol. 9, 1297–1306 (2008).

29. Zhu, H. et al. Unexpected characteristics of the IFN-c reporters in nontransformedT cells. J. Immunol. 167, 855–865 (2001).

LETTERS NATURE | Vol 460 | 16 July 2009

408 Macmillan Publishers Limited. All rights reserved©2009

30. Hertz, G. Z. & Stormo, G. D. Identifying DNA and protein patterns with statisticallysignificant alignments of multiple sequences. Bioinformatics 15, 563–577 (1999).

Supplementary Information is linked to the online version of the paper atwww.nature.com/nature.

Acknowledgements We thank R. Lallone for anti-BATF antibody preparation, andB. Sleckman for Cre-expressing adenovirus. This work was supported by theHoward Hughes Medical Institute (K.M.M.), and grants from the NationalInstitutes of Health HG00249 and training grant GM07200 (G.D.S.), AI035783(C.T.W.), AR049293 (R.D.H.), and from Daiichi-Sankyo Co. Ltd (C.T.W.).

Author Contributions B.U.S. generated Batf2/2 mice, designed and analysed theexperiments, interpreted results and wrote the manuscript. K.H. constructed the

targeting vector and probes, transgenic vector, and recombinant BATF. W.I. helpedwith retroviral expression experiments. W.-L.L. helped with reverse-strandreporter analysis. W.A.-E.S. helped with mouse generation. B.S. helped with EMSAanalysis. G.S. and G.D.S. performed bioinformatics analysis for the BATF bindingelements. J.S. and J.H.R. helped with EAE experiments. R.M., R.D.H. and C.T.W.performed ChIP experiments. T.L.M. and S.C. performed confocal microscopy forBATF. K.M.M. directed the study and wrote the manuscript.

Author Information Microarray data are available at Array Express (http://www.ebi.ac.uk/array express/) under the accession numbers E-MEXP-1518,E-MEXP-2152 and E-MEXP-2153. Reprints and permissions information is availableat www.nature.com/reprints. Correspondence and requests for materials shouldbe addressed to K.M.M (kmurphy@wustl.edu).

NATURE | Vol 460 | 16 July 2009 LETTERS

409 Macmillan Publishers Limited. All rights reserved©2009

METHODSGeneration of Batf2/2 mice. Murine Batf exons 1–2 were deleted by homolog-

ous recombination via a targeting vector constructed in pLNTK31 using a 1 kb

genomic fragment (left arm) upstream of the Batf exon 1, and a 3.6 kb genomic

fragment (right arm) downstream of exon 2. The left arm was generated by PCR

from genomic DNA with the use of the oligonucleotides: left arm forward,

59-ATTACTCGAGTGAAACAAACAGGCAGTCGCAGTG-39; left arm reverse,

59-ATTACTCGAGCCTACTACCTTTCAGGGCTACTGC-39 (bold nucleotides

indicate Xho1 restriction-enzyme sites). The right arm was generated by PCR

with the use of the oligonucleotides: right arm forward, 59-ATTAGTCGAC-GCATTCTTCATGGTCCTTAGCCTTGG-39; right arm reverse, 59-ATTA-

GTCGACCAGAGAATGAGAAATGTTGGAGG-39 (bold nucleotides indicate

Sal1 restriction-enzyme sites). EDJ22 embryonic stem cells were transfected with

linearized targeting vector and targeted clones were identified by Southern blot

analysis using probes A and B located 59 to the left arm and 39 to the right arm,

respectively. Probe A was generated using the oligonucleotides 59-

CAACTGGGTCTGAGTCAAGAGGT-39 and 59-CGTAGCCGCTGATTGTTT-

TAGAAC-39 to generate a 531-bp product. Probe B was generated using

the oligonucleotides 59-ACAGCTTGAACTTCAGAGCCCTCC-39 and 59-

CACATTTAAGTCACAATAACACTGC-39 to generate a 772-bp product. The

neomycin-resistance cassette was deleted from successfully targeted clones by in

vitro treatment with Adeno-Cre virus (gift from B. Sleckman, Washington

University), and targeted clones with successful neo deletion were identified

by Southern blot using probes A and B (Supplementary Fig. 2a, b). Blastocyst

injections were performed with two distinct recombinant clones, each of which

generated germline transmission of the targeted Batf allele. Male chimaeras were

crossed with 129SvEv females to establish Batf mutants on the pure 129SvEv

genetic background. All experiments were performed with mice containing theneo-deleted mutant allele. Homozygous mice were obtained by intercrossing

heterozygous siblings, and littermates were used as controls in most experiments.

For some experiments 129SvEv wild-type mice purchased from Taconic served

as controls. For experiments with DO11.10 transgenic Batf 2/2 mice, mice were

crossed to BALB/c mice for at least five generations, and littermates were used as

control.

For the generation of transgenic mice, Batf complementary DNA was cloned

from CD41 T cell messenger RNA using primers 59-GGAAGATTAGAACCAT-

GCCTC-39 and 59-AGAAGGTCAGGGCTGGAAG-39, and subcloned into the

GFP-RV retrovirus32. An amino-terminal Flag tag was introduced by Quick

Change Mutagenesis kit (Stratagene) using the primers 59-GGACTACA-

AAGACGATGACGACAAGCCTCACAGCTCCGACAGCA-39 and 59-CTTG-

TCGTCATCGTCTTTGTAGTCCATGGTTCTAATCTTCCAGATC-39. The

underlined sequence indicates nucleotides used to introduce the Flag-tag. The

Flag-tagged Batf was cloned into the CD2 microinjection cassette33 via blunt-end

strategy into a Sma1-digested CD2 microinjection cassette. Transgene expres-

sion in CD41 T cells was tested by anti-Flag western blot. CD2-N-Flag-Batf

transgenic mice were crossed to C57BL/6 and BALB/c mice for at least fivegenerations. Transgene-negative littermates were used as control mice. Mice

were bred and maintained at the animal facilities at Washington University.

All animal experiments were approved by the Animal Studies Committee at

Washington University.

Visualization of lymph nodes. To visualize superficial inguinal lymph nodes,

mice were injected with 50 ml of 1% Evans Blue dye solution into each hind foot

pad. After 1.5 h mice were euthanized and lymph nodes were visualized using a

dissecting microscope34.

Western blot analysis. To test for residual BATF protein expression, total sple-

nocytes from Batf 1/1 and Batf 2/2 129SvEv mice were stimulated with anti-CD3

for 3 days under TH17 conditions. Cells were then lysed in RIPA buffer, electro-

phoresed on 15% polyacrylamide gels, transferred to nitrocellulose, and analysed

by western blot with rabbit anti-murine BATF polyclonal serum and HRP-con-

jugated anti-rabbit immunoglobulin antibody (Jackson ImmunoResearch).

Affinity purified rabbit anti-murine BATF polyclonal serum (Brookwood

Biomedical) was generated by immunization with full-length recombinant

BATF protein. Equal protein loading was assessed by subsequent immunoblot-

ting with antibody to b-actin (Santa Cruz Biotechnology) and HRP-conjugated

anti-mouse antibody (Jackson ImmunoResearch).

For analysis of BATF protein expression in naive CD41 T cells, CD41 T cells

from Batf 1/1 and Batf 2/2 129SvEv mice were magnetically purified. Equal cell

numbers were lysed in RIPA buffer and subjected to western blot analysis as

described above.

For analysis of BATF expression in TH2 cells, magnetically purified CD41 T

cells from Batf 1/1 and Batf 2/2 mice were activated with anti-CD3/CD28 in the

presence of IL4, anti-IL12 (Tosh) and anti-IFN-c (H22). On day 4, cells were left

unstimulated or stimulated with PMA and ionomycin for 4 h. Cells were col-

lected by centrifugation, washed with PBS, and resuspended (100 3 106

cells ml21) in Affymetrix Chip lysis buffer (10 mM Tris, pH 7.5, 10 mM NaCl,

3 mM MgCl2, 0.5% IGEPAL, with protease inhibitors (PMSF, aprotinin, leupep-

tin)). After 5 min at 4 uC, nuclei were collected by centrifugation (800g for 3 min,

4 uC) and lysed in RIPA (100 3 106 cell equivalents per ml) with protease inhi-

bitors. Nuclear lysates were centrifuged for 10 min at 4 uC 15,000g, diluted with

an equal volume of 23 SDS–PAGE sample buffer containing 2-mercaptoeth-

anol, and extracts from equal cell numbers were subjected to western blot ana-

lysis using rabbit anti-murine BATF polyclonal serum. Equal protein loading

was assessed by subsequent immunoblotting with antibody to Lamin B

(Santa Cruz Biotechnology) and HRP-conjugated anti-goat Ig (Jackson

ImmunoResearch).

Immunohistochemistry. CD41 T cells from CD2-N-Flag-Batf transgenic mice

were isolated by magnetic separation and either left untreated or stimulated with

PMA and ionomycin for 4 h. Cells were then allowed to settle on poly-L-lysine-

treated slides, fixed with 4% formaldehyde, permeabilized with 0.25%

Triton X-100, and stained with an anti-Flag antibody (M2, Sigma Aldrich)

according to the manufacturer’s recommendations. A goat anti-mouse AF-488

(Invitrogen) antibody was used to detect anti-Flag staining. For analysis of

cellular localization of BATF in TH2 cells, DO11.10 CD41 T cells from CD2-

N-Flag-Batf transgenic mice were isolated and differentiated with OVA and

APCs under TH2 conditions for 7 days. On day 7, cells were either left untreated

or stimulated with PMA and ionomycin for 4 h. Cells were stained with anti-Flag

antibody as described earlier. Cells were also stained with anti-CD4 APC anti-

body (BD Biosciences). Confocal images were obtained with the Olympus

FV1000 microscope and software using a 360 oil objective. The pinhole was

set to 110mm. The excitation/emission settings used for DAPI, Alexa 488 and

Alexa 633 were 405/461 nm, 488/520 nm and 635/668 nm, respectively.

Further methods can be found in the Supplementary Information.

31. Gorman, J. R. et al. The Igk enhancer influences the ratio of Igk versus Igl Blymphocytes. Immunity 5, 241–252 (1996).

32. Ranganath, S. et al. GATA-3-dependent enhancer activity in IL-4 gene regulation.J. Immunol. 161, 3822–3826 (1998).

33. Zhumabekov, T., Corbella, P., Tolaini, M. & Kioussis, D. Improved version of ahuman CD2 minigene based vector for T cell-specific expression in transgenicmice. J. Immunol. Methods 185, 133–140 (1995).

34. Sun, Z. et al. Requirement for RORc in thymocyte survival and lymphoid organdevelopment. Science 288, 2369–2373 (2000).

doi:10.1038/nature08114

Macmillan Publishers Limited. All rights reserved©2009