Essential role of TAK1 in thymocytedevelopment and activationHong-Hsing Liu*, Min Xie†, Michael D. Schneider†, and Zhijian J. Chen*‡§

*Department of Molecular Biology and ‡Howard Hughes Medical Institute, University of Texas Southwestern Medical Center, Dallas, TX 75390-9148;and †Department of Medicine, Center for Cardiovascular Development, Baylor College of Medicine, Houston, TX 77030

Edited by Eric N. Olson, University of Texas Southwestern Medical Center, Dallas, TX, and approved June 9, 2006 (received for review April 15, 2006)

The protein kinase TAK1 mediates the activation of NF-�B inresponse to stimulation by proinflammatory cytokines and micro-bial pathogens in the innate immunity pathways. However, thephysiological function of TAK1 in the adaptive immunity pathwaysis unclear. By engineering mice lacking TAK1 in T cells, here, weshow that TAK1 is essential for thymocyte development andactivation in vivo. Deletion of TAK1 prevented the maturation ofsingle-positive thymocytes displaying CD4 or CD8, leading toreduction of T cells in the peripheral tissues. Thymocytes lackingTAK1 failed to activate NF-�B and JNK and were prone to apoptosisupon stimulation. Our results provide the genetic evidence thatTAK1 is required for the activation of NF-�B in thymocytes andsuggest that TAK1 plays a central role in both innate and adaptiveimmunity.

I�B kinase � JNK � NF-�B � T cell

The Rel�NF-�B family of transcription factors regulates theexpression of a plethora of genes involved in inflammation,

immunity, and apoptosis (1, 2). NF-�B is normally sequestered inthe cytoplasm of unstimulated cells through its association with theI�B family of inhibitory proteins. Stimulation of cells with a varietyof agents leads to the rapid phosphorylation and subsequentdegradation of I�B by the ubiquitin–proteasome pathway, thusallowing NF-�B to enter the nucleus to turn on various target genes.

Phosphorylation of I�B is catalyzed by a large kinase complexconsisting of I�B kinase (IKK)�, IKK�, and NEMO (also knownas IKK� or IKKAP). The IKK complex integrates signals fromdiverse pathways, including those emanating from the receptors forTNF� and IL-1�, Toll-like receptors (TLRs), and T cell receptors(TCRs) (3–6). Stimulation of IL-1R and some TLRs leads to therecruitment of several proteins, including the adaptor MyD88, thekinases IRAK4 and IRAK1, and the ubiquitin ligase TRAF6.TRAF6 functions in conjunction with the ubiquitin-conjugatingenzyme (E2) complex Ubc13–Uev1A to catalyze the synthesis ofLys-63-linked polyubiquitin chains on certain protein targets, in-cluding TRAF6 itself (7, 8). Polyubiquitinated TRAF6 activates aprotein kinase complex consisting of the TAK1 kinase and theadaptor proteins TAB1 and TAB2 (8, 9). The activation of TAK1by TRAF6 requires the binding between the K63 polyubiquitinchains and a conserved novel zinc finger (NZF) domain of TAB2or its homologue TAB3 (10). After TAK1 is activated, it phos-phorylates IKK� within the activation loop, resulting in the acti-vation of IKK. TAK1 also phosphorylates and activates MKK6 andMKK7, leading to the activation of p38 and JNK kinase pathways.

Recent studies have shown that TRAF-mediated polyubiquiti-nation and the TAK1 kinase complex also play an important rolein NF-�B activation in T cells (11). Stimulation of TCR by anantigenic peptide and its cognate MHC activates a tyrosine kinasephosphorylation cascade, which, in turn, leads to the activation ofprotein kinase (PK)C�. PKC� then triggers the recruitment of theCARD domain proteins CARMA1 and BCL10 and the para-caspase MALT1 to lipid rafts (12–14). MALT1 binds to TRAF6and promotes TRAF6 oligomerization, which activates its ubiquitinligase activity (11). TRAF6-mediated polyubiquitination then leadsto the activation of TAK1 and subsequent activation of IKK. This

T cell signaling pathway from BCL10 to IKK activation can bereconstituted in vitro by using purified recombinant proteins, in-cluding Ubc13–Uev1A (E2), TRAF6 (E3), and the TAK1 kinasecomplex (11). Furthermore, RNAi-mediated silencing of TAK1,TRAF2, and TRAF6 inhibits IKK activation and IL-2 productionin Jurkat T cells. However, it has been shown that MALT1 canfunction as a ubiquitin ligase that binds directly to Ubc13–Uev1Aand promotes the polyubiquitination of NEMO, thereby leading toIKK activation (15). According to this model, TRAF proteins andTAK1 are not required for IKK activation by TCR.

The role of TAK1 in NF-�B activation by receptors of the innateimmunity pathways, including TNFR, IL-1R, and TLR, has beenvalidated in vivo through the isolation of Drosophila TAK1 mutants(16) and the generation of TAK1-knockout mice (17, 18). However,conditional deletion of TAK1 in B cells by using Cd19-Cre did notabolish NF-�B activation by B cell receptors (BCRs) (17), whichalso signal through the CARMA1–BCL10–MALT1 complex (19).This result is discordant with another recent study that usedhomologous recombination in chicken DT40 cells to delete TAK1and showed that the complete absence of TAK1 abolished IKK andNF-�B activation by BCRs (20).

In this report, we investigated the role of TAK1 in T celldevelopment and activation by engineering a mouse model in whichTAK1 was conditionally deleted in T cells. We showed thatthymocytes lacking TAK1 failed to survive during the progressionfrom double-positive (DP) (CD4�CD8�) to single-positive (SP)(CD4� or CD8�) stages, resulting in significant reduction of naı̈veT cells in the peripheral tissues. The loss of TAK1 in the thymocytesprevented the activation of IKK, NF-�B, and JNK and sensitizedthe mutant cells to activation-induced apoptosis. Our results pro-vide the genetic evidence that TAK1 is essential for thymocytedevelopment and activation.

ResultsConditional Knockout of TAK1 in T Cells. To engineer conditionalalleles of Tak1 in mice, we constructed a targeting vector in whichexon 1 of Tak1 was flanked between a loxP site before thetranscriptional initiation site and another loxP site within intron 1(Fig. 1A). The FRT-neo-FRT selection cassette was inserted beforethe intronic loxP site. The 5� and 3� homologous regions were 2.5and 3.0 kb, respectively. ES cell targeting and the generation ofheterologous floxed Tak1 mice (Tak1�/flox) were carried out byusing standard protocols. The Tak1flox/flox mice were born and livednormally, and they expressed TAK1 protein as expected (data notshown). To delete the Tak1 allele specifically in T cells, we crossedTak1flox/flox mice with the Lck-Cre transgenic mice, which expressthe Cre recombinase under the control of the T cell-specific Lckpromoter (21). Southern blotting and PCR showed that the floxed

Conflict of interest statement: No conflicts declared.

This paper was submitted directly (Track II) to the PNAS office.

Abbreviations: DN, double-negative; DP, double-positive; IKK, I�B kinase; SP, single-posi-tive; TCR, T cell receptor.

§To whom correspondence should be addressed. E-mail: zhijian.chen@utsouthwestern.edu.

© 2006 by The National Academy of Sciences of the USA

www.pnas.org�cgi�doi�10.1073�pnas.0603089103 PNAS � August 1, 2006 � vol. 103 � no. 31 � 11677–11682

IMM

UN

OLO

GY

Dow

nloa

ded

by g

uest

on

Janu

ary

24, 2

021

Tak1 alleles were excised in thymocytes, but not in the tail (Fig. 1B and C). Western blotting confirmed that TAK1 was not detect-able in the thymocytes of Lck-Cre�Tak1flox/flox mice, but its expres-sion level in splenocytes was similar to that in control littermates(Fig. 1D). Surprisingly, the lymph node T cells from Lck-Cre�Tak1flox/flox mice had normal levels of TAK1 but lacked the expres-sion of Cre, whereas the cells from Lck-Cre�Tak1�/� mice still hadhigh levels of Cre expression (Fig. 1D). Genomic PCR confirmedthe presence of the floxed Tak1 allele in T cells derived from lymphnodes and blood (data not shown), indicating that these cells haveescaped from Cre-mediated recombination. Thus, the loss of Creexpression and the retention of the floxed Tak1 alleles in Lck-Cre�Tak1flox/flox mice likely resulted from counterselection during T celldevelopment in the thymus (see below and Discussion). For thecontrol groups, we observed no phenotypic difference amongTak1floxed/floxed, Lck-Cre�Tak1�/�, and Lck-Cre�Tak1floxed/�, indicat-ing that one copy of the Cre transgene did not have any confoundingeffect on the functional analyses of the mice. For simplicity, the

Lck-Cre�Tak1flox/flox mice with deletion of the Tak1 alleles areherein referred to as Tak1D or knockout, whereas the control micestill containing the floxed Tak1 allele are referred to as Tak1FL orcontrol.

Reduction of Peripheral T Cells in TAK1 Conditional Knockout Mice.We analyzed peripheral B (B220�) and T cells (CD3�) in Tak1D

and Tak1FL by FACS. Although the percentage of B cells wassimilar in both genotypes, the percentage of T cells in the peripherallymphoid organs, including lymph nodes, spleens, and blood, wassignificantly lower in Tak1D mice as compared with controls (Fig.2A). This decrease of T cell percentage was not due to an increaseof B cell number, because the number of splenocytes was similar inthe knockout and control mice (4.0 � 1.0 � 108 in Tak1D vs. 4.1 �0.8 � 108 in Tak1FL; n � 5). The decrease of T cell number was notobserved in various control animals, including floxed mice withoutthe Lck-Cre transgene and the Lck-Cre mice without the floxedTak1 allele. The percentage of T cells in the blood of Tak1D micewas about one-fourth that of the control mice (Fig. 2B; 5.9 � 1.1%in Tak1D vs. 26.4 � 1.6% in Tak1FL; n � 13). The relativeabundance of helper (CD4�) vs. cytotoxic (CD8�) T cells in theblood was similar for both Tak1D and Tak1FL mice (CD4��CD8�

ratio, 1.2 � 0.2 in Tak1D vs. 1.3 � 0.2 in Tak1FL; n � 6).

TAK1 Is Required for the Development and Maturation of Single-Positive Thymocytes. The reduction of T cells in the peripherallymphoid organs of Tak1D mice may be due to defective T celldevelopment in the thymus. Intrathymic T cell precursors developthrough several stages before entering the peripheral mature T cellpool (2, 22). The most immature cells transit from the double-negative (DN) stage CD4�CD8� into the DP stage (CD4�CD8�)after completion of �-selection. DP thymocytes go through furtherselections before committing to SP cells (CD4� or CD8�). To

Fig. 1. Conditional deletion of Tak1 in mouse thymocytes. (A) Strategies forthe generation and deletion of floxed Tak1 alleles. (B) Southern blotting ofgenomic DNA after digestion with NheI. (C) PCR of genomic DNA isolated fromthe tails or thymocytes of the mice, as indicated. (D) Western blotting ofwhole-cell lysates from splenocytes, thymocytes, and lymph node T cells.Lysates from 1.6 � 106 cells were loaded on each lane. In lane 6, lymph nodeT cells for Lck-Cre�Tak1flox/flox were pooled from three mice. The expression ofTAK1 and loss of expression of Cre in these cells likely resulted from theselective expansion of TAK1-expressing cells that escaped from Cre-mediatedexcision.

Fig. 2. Reduction of peripheral T cells in Tak1D mice. (A) Suspension cellsfrom lymph nodes, spleens, and blood were analyzed by FACS using antibod-ies against CD3 and B220, respectively. (B) The percentages of CD3� T cells inthe blood from the TAK1-knockout (Tak1D) mice and their control littermates(n � 13).

11678 � www.pnas.org�cgi�doi�10.1073�pnas.0603089103 Liu et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

24, 2

021

determine whether TAK1 is required for thymocyte development,we analyzed the expression of CD4 and CD8 by FACS. As shownin Fig. 3A and Table 1, both CD4� and CD8� SP thymocytes inTak1D mice were reduced by �50% as compared with the Tak1FL

mice. In contrast, there was no significant difference in the numberof CD4�CD8� DP thymocytes between Tak1D and Tak1FL mice.We also analyzed the CD24highCD4�CD8� thymocytes to examinethe transition of thymocytes from DN1 to DN4 (DN1,CD44�CD25�; DN2, CD44�CD25�; DN3, CD44�CD25�; andDN4, CD44�CD25�). No apparent defect was observed in any ofthese developmental stages in Tak1D mice (Fig. 3B), consistent withnormal TAK1 protein expression in DN thymocytes in whichLck-Cre was not turned on until the later stages of DN thymocytedevelopment (data not shown) (23).

To investigate the mechanism underlying the reduction of SPthymocytes in Tak1D mice, we analyzed the CD69 surface marker,which is expressed on positively selected cells (24). As shown in Fig.3C, although Tak1D mice contained fewer CD4� and CD8� SPthymocytes, the percentages of CD69� thymocytes were compa-rable to those in the wild-type mice, indicating that loss of TAK1did not compromise the positive selection of thymocytes. To

determine whether the maturation of SP thymocytes is affected bythe loss of TAK1, we used FACS to examine the expression ofCD24, a surface marker that is gradually down-regulated duringmaturation of SP thymocytes (25). As shown in Fig. 3D, the numberof CD24low and CD24intermediate CD4� or CD8� SP cells wassignificantly less in Tak1D than in Tak1FL mice, indicating that thematuration of SP cells was impaired in Tak1D thymocytes.

Loss of TAK1 Sensitizes SP Thymocytes to Apoptosis. The reductionin the number of SP thymocytes could be due to survival disad-vantages in Tak1D thymocytes. To investigate this possibility, wecarried out an in vitro survival assay for thymocytes at DP or SPstages. These cells were sorted by FACS and cultured in vitro withor without anti-CD3� stimulation. At indicated time points, non-surviving cells were stained by Annexin-V and analyzed by FACS(Fig. 4). After stimulation with anti-CD3� for 40 h, both CD4� andCD8� SP thymocytes from Tak1D mice had a significant increasein apoptosis as compared with thymocytes from the control litter-mates, as shown by enhanced Annexin-V staining (Table 2). In theabsence of stimulation, the SP thymocytes from Tak1D mice alsodisplayed increased Annexin-V staining compared with those fromthe control mice (Fig. 4). In contrast to SP thymocytes, the Tak1D

DP thymocytes were surviving as well as control DP thymocytes inthe absence of anti-CD3� stimulation. After stimulation, the num-ber of Annexin-V-positive DP thymocytes in Tak1D mice wasslightly less than that of control mice, suggesting that TAK1 mightfacilitate the apoptosis of DP thymocytes. Cell cycle analysis by7-amino-actinomycin D (7-AAD) staining (26) showed that Tak1D

SP thymocytes did not have proliferation defects (data not shown),indicating that the decrease in SP thymocytes was primarily due toenhanced apoptosis.

Fig. 3. Defective development of CD4� and CD8� SPthymocytes in Tak1D mice. (A) Thymocytes from Tak1D

(knockout) and wild-type (control) littermates wereanalyzed by FACS using antibodies against CD4 andCD8. (Insets) The percentages of thymocytes at differ-ent stages. The absolute number of each type of thy-mocytes is shown in Table 1. (B) CD24high CD4� CD8�

DN thymocytes were analyzed by FACS using antibod-ies against CD25 and CD44. Different developmentalstages of thymocytes (DN1–4) are indicated in key tothe right of B. (C) SP and DP thymocytes were analyzedfor CD69 expression. Percentages of CD69� popula-tions were 89% in Tak1D vs. 87% in Tak1FL for CD4� SP,60% in Tak1D vs. 64% in Tak1FL for CD8� SP, and 11%in Tak1D vs. 10% in Tak1FL for CD4�CD8� DP. (D) FACSanalysis of CD24 expression in CD4� or CD8� SP thy-mocytes. High expression of CD24 is inversely corre-lated with the maturation of SP thymocytes. The re-sults are normalized for the total numbers ofthymocytes.

Table 1. Comparison of thymocyte numbers in Tak1 (knockout)mice and control littermates

Thymusgenotype

Totalthymocytes

(�108)CD4�CD8�

DP (�108)CD4� SP*

(�107)CD8� SP*

(�107)

Control (n � 5) 2.81 � 0.27 2.40 � 0.26 1.98 � 0.21 0.71 � 0.07Knockout (n � 5) 2.04 � 0.47 1.82 � 0.45 0.82 � 0.22 0.32 � 0.03

*t test P � 0.005.

Liu et al. PNAS � August 1, 2006 � vol. 103 � no. 31 � 11679

IMM

UN

OLO

GY

Dow

nloa

ded

by g

uest

on

Janu

ary

24, 2

021

TAK1 Is Required for the Activation of NF-�B and JNK in Thymocytes.The defective SP thymocyte development observed in Tak1D miceis reminiscent of the phenotypes observed in mice lacking NEMOor expressing a dominant-negative mutant of IKK� in T cells (27).Because in vitro and ex vivo studies have suggested that TAK1 isrequired for NF-�B activation in T cells (11), we used EMSA todetermine whether NF-�B activation was impaired in Tak1D thy-mocytes (Fig. 5A). As reported (28), NF-�B is active in CD4� orCD8� SP thymocytes of wild-type mice. In contrast, the NF-�Bactivity was greatly diminished in the SP thymocytes of Tak1D mice.The DP thymocytes from wild-type mice also exhibited weakNF-�B activity; this activity was not detectable in the DP thymo-cytes of Tak1D mice. The loss of TAK1 did not affect the DNAbinding of the control transcription factor Oct-1. Thus, TAK1 isrequired for NF-�B activation during the normal development ofmouse thymocytes.

To determine whether TAK1 is required for the activation ofIKK and JNK, we stimulated thymocytes with phorbol ester(phorbol 12-myristate 13-acetate) and ionomycin, which mimicthe stimulation of TCR in T cells (Fig. 5 B and C), or with TNF�(Fig. 5D). In both cases, the degradation of I�B� and activationof IKK and JNK were severely impaired in thymocytes derivedfrom Tak1D mice, whereas the activation of ERK occurrednormally in these cells. We also examined NF-�B activation after

stimulation of thymocytes with an antibody against CD3�, whichcross-links TCR. As shown in Fig. 5E, NF-�B activation wasimpaired in thymocytes from Tak1D mice. Finally, we usedreal-time PCR to measure the expression of c-myc, an NF-�B-dependent gene required for the survival of thymocytes (29–31).When thymocytes were stimulated with anti-CD3�, c-myc wasinduced by �4-fold in the wild-type cells but not in Tak1D cells(Fig. 5F). Collectively, these results indicate that TAK1 isessential for the activation of NF-�B and JNK in thymocytes.

DiscussionIn this report, we showed that specific deletion of TAK1 in T cellsprevented the development of CD4� and CD8� SP thymocytes,resulting in significant reduction of T cells in the peripheral tissues,including lymph nodes, spleens, and blood. The defective develop-ment of SP thymocytes was due, at least in part, to the increasedapoptosis of these cells, especially under conditions of anti-CD3stimulation. We further showed that TAK1 was essential for theactivation of IKK, NF-�B, and JNK, demonstrating the role ofTAK1 in T cell development and activation. Thus, TAK1 is anessential IKK kinase in both innate and adaptive immunity.

The defective thymocyte development observed in the condi-tional Tak1D mice is similar to the phenotypes of mice lackingNEMO or expressing a kinase-dead mutant of IKK� in T cells (27),further supporting the role of TAK1 in IKK activation. However,knockouts of some components of the TCR signaling pathway, suchas CARMA1, BCL10, and MALT1, which affect IKK activation inmature T cells, do not severely affect T cell development in thethymus (32–37). Thus, TAK1 and IKK may be activated by aTCR-independent signaling pathway in thymocytes. Indeed, NF-�Bis constitutively active during intrathymic development at both DNand SP stages. The constitutive activation of NF-�B in DN thymo-cytes is thought to be mediated by pre-TCR, which is assembledafter the rearrangement of TCR �-chain during the transition fromDN3 to DN4 stages. Pre-TCR signaling is ligand-independent andmay be initiated by the autonomous oligomerization of pre-TCR�-chain (38). However, the mechanism underlying the constitutiveactivation of NF-�B at the SP stage is currently unknown. Becauseour studies of Tak1D mice have now shown that TAK1 is requiredfor IKK and NF-�B activation in SP thymocytes, further studiesshould be directed toward understanding how TAK1 is activated inthese cells.

Previous studies using transgenic mice expressing an I�B� su-perrepressor under the control of the Cd2 promoter have demon-strated that NF-�B is required for the positive selection of CD8�

thymocytes (39). Furthermore, it was found that the I�B� trans-genic mice exhibited a developmental block in the transition fromDN3 to DN4 thymocytes (28). However, we did not observe anyobvious developmental defect in the DN thymocytes of Tak1D mice(Fig. 3B). A possible explanation for these distinct phenotypes isthat, in the I�B� transgenic mice, I�B� can immediately inhibitNF-�B once it is synthesized, whereas, in the Tak1D mice, the TAK1protein remains in the DN thymocytes until the endogenous TAK1is degraded after the induction of Cre and the deletion of the floxedTak1 locus. Indeed, immunoblotting experiments showed thatTAK1 is present in the DN thymocytes of Tak1D mice (data notshown). Thus, the role of TAK1 in the early stages of thymocytedevelopment remains to be determined.

A recent study using Cd19-Cre to delete Tak1 in B cells showedthat TAK1 is required for JNK, but not NF-�B, activation inresponse to B cell receptor (BCR) stimulation (17). However,another recent study using chicken DT40 cells to completelyremove Tak1 demonstrated that TAK1 is required for both IKKand JNK activation after BCR stimulation (20). It is not clearwhether these different results reflect the difference of cells(chicken vs. mouse) or the knockout strategies used in the studies.It is possible that the Cd19-Cre-mediated deletion may not be veryefficient, resulting in a low level of TAK1 activity that is sufficient

Fig. 4. Survival disadvantage of CD4� and CD8� SP thymocytes in Tak1D mice.SP and DP thymocytes from Tak1D mice (knockout) or control littermates werepurified and sorted by FACS and then cultured in 96-well plates coated withan anti-CD3� antibody or PBS buffer (as a control). Cells were harvested atindicated time points, incubated with Annexin-V, and analyzed by FACS. Atleast 4,000 events were analyzed for each sample. Probabilities shown in thediagrams represent samples stimulated with anti-CD3� for 40 h. (Inset LowerRight) is a typical FACS diagram from CD4� thymocytes stimulated withanti-CD3� for 40 h.

Table 2. Percentage of Annexin-V-positive cells

anti-CD3� stimulation (10 �g�ml)

Genotype Thymocytes 0 h, % 16 h, % 40 h, %

Knockout CD4� SP 5.39 � 1.28 43.14 � 3.64 72.04 � 7.79CD8� SP 8.38 � 0.51 45.32 � 3.14 73.18 � 3.96

CD4�CD8� 6.65 � 0.65 34.96 � 2.78 70.17 � 1.88Control CD4� SP 5.35 � 0.18 19.74 � 0.42 48.81 � 3.05

CD8� SP 4.74 � 0.82 25.80 � 6.07 57.78 � 3.38CD4�CD8� 5.11 � 0.39 35.09 � 3.22 77.79 � 3.11

11680 � www.pnas.org�cgi�doi�10.1073�pnas.0603089103 Liu et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

24, 2

021

for IKK activation but insufficient for JNK activation. Our currentstudy shows clearly that TAK1 is required for IKK, NF-�B, and JNKactivation, at least in thymocytes.

The defective T cell development in Tak1D mice results in asignificant decrease of mature T cells in the peripheral tissues. Infact, when T cells isolated from the lymph nodes of Tak1D mice wereanalyzed, they were found to express TAK1 and lack the expressionof Cre (Fig. 1D), implying that only T cells that escape fromCre-mediated excision were able to emigrate from the thymus andpopulate the peripheral tissues. The requirement of TAK1 for thedevelopment of mature T cells precludes the analysis of the role ofTAK1 in the activation of these cells. Conditional deletion of TAK1specifically in mature T cells, such as the use of a tamoxifen-inducible Cre, will be required to examine the function of TAK1 inthese cells.

In sum, our results provide the genetic evidence that TAK1 isrequired for the activation of IKK, NF-�B, and JNK in mousethymocytes and that TAK1 plays an essential role in thymocytedevelopment and activation. These results extend the pivotal role ofTAK1 in the innate immune system to the adaptive immune system.

Materials and MethodsGene Targeting and Genotyping of Mice. AB2.2 mouse ES cells weretargeted by a construct containing one loxP site before the tran-scription initiation site of Tak1 and the other loxP site in intron 1.The targeting construct also contained a FRT-neo-FRT selectioncassette before the intronic loxP site. The 5� and 3� homologousregions spanned 2.5 and 3.0 kb, respectively. Targeted ES cells werescreened by Southern blotting with both 5� and 3� probes (Fig. 1)after digestion with EcoRV and NheI, respectively. Blastocystinjection was performed at Baylor College of Medicine. Lck-Cretransgenic mice were obtained from The Jackson Laboratory (21).Floxed Tak1 mice were crossed to Lck-Cre mice at University ofTexas Southwestern Medical Center. Mice were genotyped by PCR

using the following primer pairs: GCACAGAAAATGCACAGT-GCTC and GCTTGGGACAGGCTGGTAAAG (for the wild-type allele), GCACAGAAAATGCACAGTGCTC and CTTA-CAAGCCGAATTCCAGCA (for the f loxed allele), andGCACAGAAAATGCACAGTGCTC and CTCCTCCACTC-CGCCCCTAC (for the excised allele). The PCR conditions were94°C for 30 s, 58°C for 30 s, and 72°C for 1 min; 35 cycles. The miceused in this study were 5–10 weeks old. All mice were housed inconventional animal facilities at University of Texas SouthwesternMedical Center or Baylor College of Medicine.

FACS. Spleens, thymi, and lymph nodes were mechanically disruptedby a syringe pump and filtered through cell strainers (100 �m; BDBiosciences) to obtain suspension cells. Blood cells were isolated byfollowing the online protocol at The Jackson Laboratory (www.jax.org�imr�facs.html), except FACS buffer was replaced by 5%FBS in PBS. Cells were stained with a monoclonal antibody for 15min on ice and washed once before FACS analysis. In the eventwhen staining with a secondary antibody was required, cells werestained with the antibody for another 15 min on ice, followedby another wash step. Data were collected by FACSCalibur orFACScan (Becton Dickinson) flow cytometers and analyzed byusing CellQuest software. Primary antibodies against B220 (RA3-6B2), CD3 (17A2), CD24 (M1�69), CD4 (GK1.5), CD8a (53-6.7),and CD69 (H1.2F3) were from BD Biosciences; these antibodiesare conjugated with different markers, such as FITC, phycoerythrin(PE), allophycocyanin (APC), or biotin. Streptavidin coupled toperidinin chlorophyll protein (BD Biosciences) was used as asecondary antibody.

Isolation and Purification of Thymocytes and Lymph Node T Cells.CD4� SP and CD4�CD8� DP thymocytes were directly sorted byFACSVantage SE (with DIVA upgrade) after CD4 and CD8staining. CD8� SP thymocytes were purified by depleting CD4�

Fig. 5. TAK1 is required for the activation of NF-�B and JNK in thymocytes. (A) (Upper) EMSAs for NF-�B DNA-binding activity using whole-cell extracts fromSP or DP thymocytes isolated from Tak1D mice (lanes 1, 3, and 5) or control littermates (lanes 2, 4, and 6). (Lower) The same extracts were assayed for DNA bindingof the constitutive transcription factor Oct-1. (B) Thymocytes were stimulated with phorbol 12-myristate 13-acetate (100 ng�ml) and ionomycin (200 ng�ml) forthe indicated time periods, and cell lysates were harvested for analysis by immunoblotting using antibodies specific for I�B�, JNK, phosphorylated JNK or ERK,or tubulin (as a loading control). (C) Thymocytes were stimulated with PMA and ionomycin as described above, and the IKK complex was immunoprecipitatedby using a NEMO-specific antibody. IKK activity was measured by using GST-I�B�-NT (N terminus) and �-32P-ATP as the substrates. Aliquots of the immunopre-cipitated complexes were subject to immunoblotting using an antibody against IKK�. (D) Thymocytes were stimulated with TNF� for the indicated time periods,and cell lysates were harvested for analysis by immunoblotting using antibodies specific for TAK1, I�B�, JNK, phosphorylated JNK, or tubulin. (E) Thymocyteswere incubated with PBS or plate-bound anti-CD3� for 16 h, and whole-cell extracts were prepared for analyses of NF-� B or OCT-1 DNA binding by EMSA. Thesame extracts were also subjected to immunoblotting with an antibody against p65. (F) Thymocytes from Tak1D mice or control littermates were incubated withPBS or plate-bound anti-CD3� for 20 h before total RNA was extracted for real-time PCR analyses. Two mice were used in each group. The c-myc expression levelswere normalized to the levels of �-actin. The error bars indicate standard errors.

Liu et al. PNAS � August 1, 2006 � vol. 103 � no. 31 � 11681

IMM

UN

OLO

GY

Dow

nloa

ded

by g

uest

on

Janu

ary

24, 2

021

cells with a magnetic column, followed by FACS sorting for CD8�

cells. Briefly, thymocytes were incubated with anti-CD4-PE andanti-PE magnetic beads (Miltenyi Biotec) before applying to amagnetic column. The unbound materials were incubated withanti-CD8a-FITC and then sorted for the CD8� SP thymocytes byFACS. The purity of the sorted cells was at least 95%. Lymph nodeT cells were purified by using a Pan T Cell Isolation kit (MiltenyiBiotec) from a pool of popliteal, axillary, and mesentery lymphnodes. The purity of the sorted CD3� cells was at least 96%.

Annexin-V Cell Death Assay. Purified thymocytes at various stageswere pelleted and resuspended in complete media (RPMI medium1640, 10% FBS, penicillin�streptomycin, and 50 �M �-mercapto-ethanol) at a density of 5 � 105 per ml. Aliquots of the cells (5 �104 cells per well) were grown in 96-well plates precoated witheither PBS or 10 �g�ml anti-CD3� (145-2C11; BD Biosciences). Atindicated times, cells were incubated with Annexin-V-APC (BDBiosciences) in staining buffer (10 mM Hepes, pH 7.4, 140 mMNaCl, and 2.5 mM CaCl2) for 15 min at room temperature andanalyzed by FACS. At least 4,000 events were recorded for eachsample.

Biochemical Analyses. Immunoblotting was carried out by usingstandard procedures. In Fig. 1D, cells (1.6 � 106) were lysed directlyin SDS sample buffer supplemented with 25 units of Benzonase(Novagen), which digests genomic DNA to reduce viscosity. Afterincubation at 4°C for 30 min, the samples were boiled and subjectedto SDS�PAGE and Western transfer. In Fig. 5 B and C, cells werelysed in 200 �l of kinase assay buffer per 107 cells [20 mM Tris�HCl,pH 7.5, 100 mM NaCl, 25 mM �-glycerophosphate, 1 mM sodiumvanadate, 10% glycerol, 0.02% Nonidet P-40, and proteinase in-hibitor (Roche)]. The antibodies against phospho-ERK and JNKwere from Cell Signaling Technology, and the antibody againstphospho-JNK was from BioSource International, Camarillo, CA.Antibodies for TAK1 (M579), I� B� (C21), and NEMO (FL-419)were from Santa Cruz Biotechnology. Antibodies for tubulin andCre were from Sigma and Novagen, respectively.

For EMSA, whole-cell extracts [3–4 �g of protein in 20 mM Tris,pH 7.5, 10% glycerol, 0.4 M KCl, 1 mM DTT, 1 mM EDTA, 0.1%Nonidet P-40, and proteinase inhibitor (Roche)] were incubatedwith radiolabeled DNA probes containing the consensus NF-�B- or

Oct-1-binding sites (Promega). After incubation at room temper-ature for 15 min, the DNA–protein complexes were resolvedby electrophoresis on 5% polyacrylamide gel and analyzed byPhosphorImaging.

For NF-�B and JNK activity assays, thymocytes were preparedat a density of 2 � 107 per ml in complete media (RPMI medium1640, 10% FBS, penicillin�streptomycin, and 50 �M �-mercap-toethanol) and stimulated with phorbol 12-myristate 13-acetate(100 ng�ml) and ionomycin (200 ng�ml) or mouse TNF� (25ng�ml; Chemicon) for the indicated time periods. The IKKkinase assay was carried out as described (11). For plate-boundanti-CD3� stimulation, 106 per 100 �l thymocytes in completemedia were stimulated for 16 h at 37°C in 96-well plates that hadbeen coated by either PBS or 10 �g�ml anti-CD3�.

Real-Time PCR. Thymocytes were stimulated with anti-CD3� for 20 has described above, and total RNA was extracted by using theQiagen RNeasy Mini kit. First-strand cDNA was synthesized byusing SuperScript III SuperMix for quantitative RT-PCR (Invitro-gen). Real-time PCR was performed in duplicates in the iQ5multicolor detection system using SYBR green supermix (Bio-Rad). c-myc primers were GCCCAAATCCTGTACCTCGTC andTGCCTCTTCTCCACAGACACC. �-actin primers wereTGACGTTGACATCCGTAAAGACC and AAGGGTGTA-AAACGCAGCTCA. The PCR conditions were 94°C for 30 s, 58°Cfor 30 s, and 72°C for 30 s; 40 cycles. The expression of c-myc wasnormalized by using �-actin as an internal control.

Statistics and Graph Preparation. Student’s t tests were used forstatistical analysis. Data are presented as average � SE.Graphs, except FACS analyses, were prepared by using theprograms gnuplot (www.gnuplot.info), Adobe Photoshop, orMicrosoft Excel.

We thank Gabriel Pineda for help with the IKK kinase assay. This workwas supported by National Institutes of Health Grant R01-AI60919,American Cancer Society Grant RSG0219501TBE, and Welch Foun-dation Grant I-1389. Z.J.C. is an Investigator of the Howard HughesMedical Institute and a Burroughs Wellcome Fund Investigator inPathogenesis of Infectious Diseases.

1. Silverman, N. & Maniatis, T. (2001) Genes Dev. 15, 2321–2342.2. Siebenlist, U., Brown, K. & Claudio, E. (2005) Nat. Rev. Immunol. 5, 435–445.3. Chen, G. & Goeddel, D. V. (2002) Science 296, 1634–1635.4. Dunne, A. & O’Neill, L. A. (2003) Sci. STKE 2003, re3.5. Akira, S. & Takeda, K. (2004) Nat. Rev. Immunol. 4, 499–511.6. Weil, R. & Israel, A. (2004) Curr. Opin. Immunol. 16, 374–381.7. Deng, L., Wang, C., Spencer, E., Yang, L., Braun, A., You, J., Slaughter, C., Pickart,

C. & Chen, Z. J. (2000) Cell 103, 351–361.8. Wang, C., Deng, L., Hong, M., Akkaraju, G. R., Inoue, J. & Chen, Z. J. (2001) Nature

412, 346–351.9. Ninomiya-Tsuji, J., Kishimoto, K., Hiyama, A., Inoue, J., Cao, Z. & Matsumoto, K.

(1999) Nature 398, 252–256.10. Kanayama, A., Seth, R. B., Sun, L., Ea, C. K., Hong, M., Shaito, A., Chiu, Y. H., Deng,

L. & Chen, Z. J. (2004) Mol. Cell 15, 535–548.11. Sun, L., Deng, L., Ea, C. K., Xia, Z. P. & Chen, Z. J. (2004) Mol. Cell 14, 289–301.12. Lin, X. & Wang, D. (2004) Semin. Immunol. 16, 429–435.13. Lucas, P. C., McAllister-Lucas, L. M. & Nunez, G. (2004) J. Cell Sci. 117, 31–39.14. Thome, M. & Tschopp, J. (2003) Trends Immunol. 24, 419–424.15. Zhou, H., Wertz, I., O’Rourke, K., Ultsch, M., Seshagiri, S., Eby, M., Xiao, W. &

Dixit, V. M. (2004) Nature 427, 167–171.16. Vidal, S., Khush, R. S., Leulier, F., Tzou, P., Nakamura, M. & Lemaitre, B. (2001)

Genes Dev. 15, 1900–1912.17. Sato, S., Sanjo, H., Takeda, K., Ninomiya-Tsuji, J., Yamamoto, M., Kawai, T.,

Matsumoto, K., Takeuchi, O. & Akira, S. (2005) Nat. Immunol. 6, 1087–1095.18. Shim, J. H., Xiao, C., Paschal, A. E., Bailey, S. T., Rao, P., Hayden, M. S., Lee, K. Y.,

Bussey, C., Steckel, M., Tanaka, N., Yamada, G., Akira, S., Matsumoto, K. & Ghosh,S. (2005) Genes Dev. 19, 2668–2681.

19. Simeoni, L., Kliche, S., Lindquist, J. & Schraven, B. (2004) Curr. Opin. Immunol. 16,304–313.

20. Shinohara, H., Yasuda, T., Aiba, Y., Sanjo, H., Hamadate, M., Watarai, H., Sakurai,H. & Kurosaki, T. (2005) J. Exp. Med. 202, 1423–1431.

21. Hennet, T., Hagen, F. K., Tabak, L. A. & Marth, J. D. (1995) Proc. Natl. Acad. Sci.USA 92, 12070–12074.

22. Rothenberg, E. V. & Taghon, T. (2005) Annu. Rev. Immunol. 23, 601–649.23. Cantrell, D. A. (2002) Nat. Rev. Immunol. 2, 20–27.24. Ziegler, S. F., Ramsdell, F. & Alderson, M. R. (1994) Stem Cells 12, 456–465.25. Tian, T., Zhang, J., Gao, L., Qian, X. P. & Chen, W. F. (2001) Int. Immunol. 13,

313–320.26. Toba, K., Winton, E. F., Koike, T. & Shibata, A. (1995) J. Immunol. Methods 182,

193–207.27. Schmidt-Supprian, M., Courtois, G., Tian, J., Coyle, A. J., Israel, A., Rajewsky, K. &

Pasparakis, M. (2003) Immunity 19, 377–389.28. Voll, R. E., Jimi, E., Phillips, R. J., Barber, D. F., Rincon, M., Hayday, A. C., Flavell,

R. A. & Ghosh, S. (2000) Immunity 13, 677–689.29. Wang, W., Wykrzykowska, J., Johnson, T., Sen, R. & Sen, J. (1999) J. Immunol. 162,

314–322.30. Duyao, M. P., Buckler, A. J. & Sonenshein, G. E. (1990) Proc. Natl. Acad. Sci. USA

87, 4727–4731.31. Broussard-Diehl, C., Bauer, S. R. & Scheuermann, R. H. (1996) J. Immunol. 156,

3141–3150.32. Ruland, J., Duncan, G. S., Elia, A., del Barco Barrantes, I., Nguyen, L., Plyte, S.,

Millar, D. G., Bouchard, D., Wakeham, A., Ohashi, P. S. & Mak, T. W. (2001) Cell104, 33–42.

33. Egawa, T., Albrecht, B., Favier, B., Sunshine, M. J., Mirchandani, K., O’Brien, W.,Thome, M. & Littman, D. R. (2003) Curr. Biol. 13, 1252–1258.

34. Hara, H., Wada, T., Bakal, C., Kozieradzki, I., Suzuki, S., Suzuki, N., Nghiem, M.,Griffiths, E. K., Krawczyk, C., Bauer, B., et al. (2003) Immunity 18, 763–775.

35. Newton, K. & Dixit, V. M. (2003) Curr. Biol. 13, 1247–1251.36. Ruef li-Brasse, A. A., French, D. M. & Dixit, V. M. (2003) Science 302, 1581–

1584.37. Ruland, J., Duncan, G. S., Wakeham, A. & Mak, T. W. (2003) Immunity 19, 749–758.38. Yamasaki, S., Ishikawa, E., Sakuma, M., Ogata, K., Sakata-Sogawa, K., Hiroshima,

M., Wiest, D. L., Tokunaga, M. & Saito, T. (2006) Nat. Immunol. 7, 67–75.39. Hettmann, T. & Leiden, J. M. (2000) J. Immunol. 165, 5004–5010.

11682 � www.pnas.org�cgi�doi�10.1073�pnas.0603089103 Liu et al.

Dow

nloa

ded

by g

uest

on

Janu

ary

24, 2

021