Molecular Cell Article

Chromatin-Associated Protein C-q Regulates an Inducible Expression Program and MicroRNAs in Human T Lymphocytes

Elissa L. Sutcliffe,1 Karen L. Bunting,3,8 Yi Qing He,1,8 Jasmine Li,1 Chansavath Phetsouphanh,6 Nabila Seddiki,6 Anjum Zafar,1 Elizabeth J. Hindmarsh,1 Christopher R. Parish,1 Anthony D. Kelleher,6 Russell L. McInnes,4 Toshiki Taya,5 Peter J. Milburn,2 and Sudha Rao1,7,* 1Department of Immunology 2Biomolecular Resource Facility John Curtin School of Medical Research, Australian National University, Canberra, ACT 2601, Australia 3Division of Hematology and Medical Oncology, Department of Medicine, Weill Cornell Medical College, Cornell University, New York, NY 10065, USA 4Agilent Technologies, 347 Burwood Highway, Forest Hill VIC 3131, Australia 5Agilent Technologies Japan, Hachioji Site, 9-1 Takakura-Cho, Hachioji-Shi, Tokyo 192-8510, Japan 6National Centre in HIV Epidemiology and Clinical Research, St. Vincent’s Centre for Applied Medical Research, University of New South Wales, Sydney NSW 2010, Australia 7Discipline of Biomedical Sciences, Faculty of Applied Science, The University of Canberra, Canberra ACT 2601, Australia 8These authors contributed equally to this work *Correspondence: [email protected] DOI 10.1016/j.molcel.2011.02.030

SUMMARY provides the exquisite control necessary for the initiation of specific transcription programs. Traditionally, the action of Studies in yeast demonstrate that signaling signaling kinases was thought to occur predominantly in the have a surprisingly active role in the nucleus, where cytoplasm. Pioneering studies in yeast, however, have shown they tether to chromatin and modulate gene expres- that kinases translocate to the nucleus and sion programs. Despite these seminal studies, the stably associate with the promoter and transcribed regions of nuclear mechanism of how signaling kinases control to regulate expression (Pascual-Ahuir et al., 2006; Pokho- transcription of mammalian genes is in its infancy. lok et al., 2006). These chromatin-tethered kinases have been shown to have both a structural role as part of transcription Here, we provide evidence for a hitherto unknown q complexes, as well as an enzymatic role by phosphorylating function of C-theta (PKC- ), which their target proteins (de Nadal and Posas, 2010; Edmunds and physically associates with the regulatory regions of Mahadevan, 2006). inducible immune response genes in human T cells. There is growing evidence in higher eukaryotes that signal Chromatin-anchored PKC-q forms an active nuclear transduction kinases can exhibit a distinct function in both the complex by interacting with RNA polymerase II, the cytoplasm and the nucleus. For example, upstream kinases of histone kinase MSK-1, and the adaptor molecule the NF-kB pathway, IKK1 and NIK, shuttle between the cyto- 14-3-3z. ChIP-on-chip reveals that PKC-q binds to plasm and nucleus in resting cells to facilitate basal NF-kB tran- promoters and transcribed regions of genes, as scriptional activity (Birbach et al., 2002). IKK-a was shown to well as to microRNA promoters that are crucial for have an alternative role as a histone kinase that directly phos- cytokine regulation. Our results provide a molecular phorylates H3S10 at the promoters of NF-kB-responsive genes (Anest et al., 2003; Yamamoto et al., 2003). The p38 mitogen- explanation for the role of PKC-q not only in normal activated protein kinase (MAPK) is another signaling kinase T cell function, but also in circumstances of its that is recruited to the chromatin of muscle-specific loci where ectopic expression in cancer. it targets the SWI-SNF chromatin remodeling complex (Simone et al., 2004). It has also been demonstrated that inflammatory stimuli signal via p38 to chromatin (Saccani et al., 2002). INTRODUCTION Translocation of signaling kinases to the nucleus could provide an efficient mechanism whereby cells communicate Intracellular signal transduction often involves a complex extracellular signals generated at the plasma membrane to the cascade of phosphorylation events that enable cells to respond nucleus. Indeed, PKC isozymes have been shown to be capable appropriately to extracellular stimuli. Kinases play a crucial role of residency within the nucleus (Martelli et al., 1999), and it has in these signaling pathways as they transiently associate with been known for some time that these kinases can phosphorylate binding partners and transfer phosphate groups onto target histones in vitro (Inoue et al., 1977; Yu et al., 1998). Some of the substrates. The activation and concerted action of such proteins nuclear substrates identified for PKCs in vivo include DNA

704 Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

topoisomerase I, lamin B, myogenin, nucleolin, p53, TdT and the ciation between endogenous PKC-q and histones H3 and H4 in vitamin D3 receptor (Martelli et al., 1999). Importantly, phosphoi- T cells (Figure 1E). Additionally, PKC-q associates with two indi- nositide signaling, which is central to the PKC activation cators of active chromatin, Pol II and histone H3 acetylated at pathway, appears to be intact at both the plasma membrane lysine-9 (H3K9ac) (Figure 1F). The anti-phospho-Pol II antibody and in the nucleus (Visnjic and Banfic, 2007). Despite increasing recognizes this protein when it is phosphorylated at serine-5 (de- evidence that PKC signaling kinases localize to the nucleus, their noted anti-Pol IIp herein). PKC-q failed to coimmunoprecipite the nuclear-specific functions and substrate specificities have re- repressive marks H2A.Z and HP1-a (data not shown). Taken mained unclear. The Ca2+-independent PKC isoform, PKC-q, together, nuclear PKC-q exists in the proximity of nucleosomes has previously been shown to associate with centrosomes and within active chromatin. kinetochore structures of the mitotic spindle within the nucleus of murine erythroleukemia cells (Passalacqua et al., 1999). This PKC-q Is Preferentially Recruited to the Regulatory kinase is best known for its critical role in efficient T cell activation Regions of Activated Genes and Forms a Nuclear (Isakov and Altman, 2002) and for its importance in proliferation, Complex with Pol II and MSK-1 in Human T Cells differentiation, and survival of mature T cells (Hayashi and Alt- Since PKC-q colocalized in the vicinity of nucleosomes in both man, 2007). Dysregulation of the PKC family has also been impli- resting and activated T cells, we investigated whether PKC-q cated in pathologies such as cancer and tumor metastasis functions as a chromatin-tethered kinase, similar to those (Griner and Kazanietz, 2007; Martiny-Baron and Fabbro, 2007). recently reported (Pokholok et al., 2006; Proft et al., 2006). Here, we demonstrate a nuclear role for PKC-q as a compo- Sequential ChIP showed increased co-occupancy between nent of an activating transcriptional complex in human T cells. histone H3 not only with PKC-q and catalytically active PKC-q We have shown that PKC-q physically associates with the prox- (denoted as PKC-qp herein), but also with the histone kinase imal promoter and coding regions of immune response genes MSK-1 on the CD69 proximal promoter in activated T cells (ST) after T cell activation. Chromatin-tethered PKC-q is intimately (Figure 2A). PKC-q and MSK-1 were also found to coexist on associated with the presence of RNA polymerase II, the histone this stretch of DNA in activated T cells (Figure 2B). A stimulus- kinase MSK-1, and the adaptor protein 14-3-3z. This event dependent increase in PKC-q recruitment was observed across appears to be coupled with PKC-q catalytic activity. ChIP-on- the CD69 gene, with maximal occupancy detected at +0.5 Kb in chip showed that PKC-q also localizes to the regulatory regions the 50-transcribed region following 24 hr of P/I activation (Fig- of a distinct cluster of microRNAs and negatively regulates their ure 2C). Similarly, a transient increase in MSK-1 occupancy transcription. These findings exemplify a more general role for was detected across the CD69 gene, with levels peaking signaling kinases as regulators of gene transcription in mamma- at +0.5 Kb in the transcribed region following 4 hr of P/I (Fig- lian cells that act by two distinct mechanisms: (1) cytoplasmic ure 2D). PKC-q and MSK-1 occupancy were also detected in signaling to the nucleus and (2) direct association with chro- resting T cells in the 50-transcribed region (Figures 2C and 2D), matin-bound transcription complexes at activated target genes which may reflect basal transcription of this gene (Sutcliffe in the nucleus. et al., 2009). PKC-q catalytic activity is substantially influenced by phos- RESULTS phorylation of serine-695 in its hydrophobic motif (Liu et al., 2002). A transient increase was observed in active PKC-qp PKC-q Localizes in the Nucleus of T Cells and Coresides (phosphorylated serine-695) at the CD69 promoter after 4 hr with Active Chromatin P/I (Figure 2E). Recruitment of PKC-qp was closely coupled to To determine whether PKC-q resides in the nucleus of human the early stages of transcriptional activation at the CD69 T cells, we performed immunoblot of nuclear extracts from Ju- promoter, while being sustained in the 50-transcribed region (Fig- rkat T cells that were either untreated or stimulated with phorbol ure 2E). Recruitment of Pol II to gene promoters and coding 12-myristate 13-acetate/calcium ionomycin (P/I). PKC-q was regions is a well-characterized mark of active gene transcription. present in the nucleus of both nonstimulated (NS) and P/I-stim- Sequential ChIP revealed increased co-occupancy of both ulated (ST) T cells (Figure 1A). To confirm the purity of nuclear PKC-q and MSK-1 with Pol II on the CD69 gene promoter in acti- preparations, we carried out immunofluorescence staining for vated T cells (Figure 2F). These data suggest that PKC-q exists the nonnuclear antigen, Integrin b1, which was excluded from as part of an active transcriptional complex on chromatin with the isolated T cell nuclei (Figure 1B). Nuclear PKC-q was local- MSK-1 and Pol II on the CD69 gene. ized to the weakly DAPI-stained euchromatic compartment, in We next interrogated by ChIP whether PKC-q associated with both resting and activated T cells (data not shown). Consistent the promoters of CD69, heparanase, tumor necrosis factor-alpha with the detection of PKC-q in the euchromatin fraction, this (TNF-a), and interferon-gamma (IFN-g). These genes were kinase coresided with RNA polymerase II (denoted Pol II herein) chosen for their inducibility in T cells and for their critical immune in activated T cells (Figure 1C). In contrast, PKC-q did not coloc- functions, while the constitutively expressed GAPDH gene was alize with the repressive mark heterochromatin protein 1-a (HP1- used for comparison. Levels of PKC-q occupancy at the prox- a) or the heterochromatic histone modification, dimethylation of imal promoters of all genes increased in a stimulus-dependent lysine-9 of histone H3 (H3K9me2) (Figure 1C). PKC kinase manner, except for GAPDH (Table 1). Affymetrix profiling micro- assays confirmed catalytic activity of PKC in the nucleus and array data showed that Jurkat T cells closely mimic naive human in whole T cells prior to and after T cell activation (Figure 1D). CD4+ T cells, as they share at least 85% of their transcriptome PKC-q halfway ChIP on T cell nuclear extracts showed an asso- (data not shown). To further assess that PKC-q associates with

Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. 705 Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

DAPI Integrin 1 merged A Nuclear extracts B NS ST Whole cell

PKC- (82 kDa)

SP1 (95 kDa) nucleus

C DAPI PKC- Pol II Merged

D 1.2 DAPI PKC- HP1- Merged 0.9

0.6

0.3 DAPI PKC- H3K9me2 Merged Absorbance 450 nm at 0.0 ST ST NS NS WCL NE (-) Control (+) Control (+)

E -Ab IP PKC- IP F -Ab IP PKC- IP NS 4 NS 4 NS 4 NS 4

PKC- Pol II (82 kDa) (217 kDa)

Pol IIp H3 (17 kDa) (240 kDa)

H3K9ac H4 (17 kDa) (14 kDa)

Figure 1. PKC-q Associates with Active Chromatin in the Nucleus of Human T Cells (A) Immunoblot with anti-PKC-q (top) and anti-Sp1 (bottom) antibodies on nuclear extracts prepared from either nonstimulated (NS) or stimulated (ST with P/I for 4 hr) Jurkat T cells.

(B) Fluorescence microscopy on fixed whole cells and nuclei stained with DAPI and anti-integrin b1 antibody. (C) Fixed nuclei from Jurkat T cells ST stained with DAPI and anti-PKC-q antibody with either anti-Pol II, anti-HP1a or anti-H3K9me2 antibody. (D) PKC ELISA-based kinase assays of nuclear and whole-cell extracts of NS and ST-treated Jurkat T cells. Recombinant active PKC (postitive control) and secondary alone (negative control). Data represent the mean ± SE of two independent experiments performed in duplicate. (E and F) Halfway ChIP of either NS or 4 hr P/I-treated Jurkat T cells with the anti-PKC-q antibody. (–Ab) is in the absence of antibody. Immunoblot with anti- PKC-q, anti-histone H3 and anti-histone H4 antibodies (E) or anti-Pol II, anti-Pol IIp and anti-histone H3K9ac antibodies (F). chromatin, we performed ChIP on ex vivo-derived human ulations on the human IL-2 promoter (Figure 2G), consistent with primary CD4+ T cell subsets (Figure S1 available online), namely recent studies implicating this kinase in T cell memory (Marsland naive, resting, effector, and activated memory. PKC-q was en- et al., 2005). Furthermore, H3K9ac occupancy on the IL-2 riched predominantly in the effector and activated memory pop- promoter was similarly enriched in the memory T cell subsets

706 Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

over the naive subset (Figure 2H), consistent with PKC-q primarily To delineate the association of PKC-q with 14-3-3z or Pol II on associating with active chromatin and the recent observations in inducible genes, we performed ChIPs on T cells transiently trans- human CD8+ memory T cells (Araki et al., 2008). This chromatin fected with a PKC-q overexpression construct (WT), a PKC-q profile was mirrored at the transcript level, whereby IL-2 was ex- catalytically inactive mutant (KR), and a vector only (V) construct pressed at the highest levels in the effector and activated memory control. ChIP showed increased PKC-q recruitment to the CD69 populations (Figure 2I). To complement the PKC-q staining profile and heparanase promoters in WT cells compared to the vector in Jurkat T cells (Figure 1C), an accumulation of PKC-q was control (Figures 4E). Consistent with studies of the catalytically observed in the euchromatin fraction in ex vivo derived human inactive form of the yeast Hog1 kinase, which was unable to CD4+ T cells (Figure 2J). Less nuclear PKC-q was detected in engage with gene promoters (Alepuz et al., 2001), the PKC-q resting memory T cells compared to the other memory subsets KR mutant also failed to associate with inducible gene promoters (Figure 2J), which parallels the ChIP enrichments across these (Figures 4E). The inability of the PKC-q KR mutant to be recruited cells (Figure 2G). These data support the view that human onto inducible gene promoters cannot be attributed to lack of PKC-q selectively associates with the promoters of activated nuclear localization, since overexpression of both the His- immune response genes in human T cells. tagged PKC-q KR mutant and His-PKC-q displayed similar nuclear distribution in Jurkat T cells (data not shown). Binding PKC-q Recruitment Correlates with the Active profiles of 14-3-3z and Pol II closely mirrored those of PKC-q Transcription of Inducible Genes observed for WT and KR mutant samples (Figures 4E). These To assess the time-dependence of PKC-q recruitment after data demonstrate that binding of 14-3-3z and Pol II are depen- continuous T cell activation, an experiment was designed (sche- dent on catalytically active PKC-q. Thus, in addition to MSK-1 matically depicted in Figure 3A) to measure CD69, TNF-a, and and Pol II, PKC-q also appears to form a complex with 14-3-3z IFN-g gene transcription and the presence of PKC-q and at the promoters of activated genes in T cells. MSK-1 on these gene promoters before (NS), during (ST), and after (cells were washed extensively five times to remove Chromatin-Tethered PKC-q Forms an Active the stimulus; SW) T cell activation. Consistent with our previous Transcription Complex that Regulates Chromatin observations, transcripts were all induced after 4 hr of T cell acti- Accessibility, but Not H3S10p at the Promoters vation (Figure 3A). After stimulus removal, transcript levels of all of Inducible Genes of these genes diminished, albeit to different extents (Figure 3A). The nucleosomal kinase MSK-1 mediates serine-10 phosphory- Occupancy profiles of PKC-q and MSK-1 paralleled active tran- lation of histone H3 (H3S10p) on immediate-early response scription of these genes (Figures 3B and 3C, respectively), with genes, with this motif acting as a code for the recruitment of no MSK-1 detected on these promoters after stimulus removal 14-3-3z (Winter et al., 2008). Given that PKC-q and MSK-1 are (Figure 3C). Previous studies have shown that the histone variant tethered to the proximal regulatory elements of immune H2A.Z is preferentially enriched on transcriptionally inactive response genes, we investigated the ability of these kinases to promoters and depleted from highly transcribed genes. Accord- modulate H3S10p. Enforced expression of either PKC-q WT or ingly, the deposition of H2A.Z at these gene promoters was the catalytically inactive KR displayed minimal effects on inversely correlated with active gene transcription (Figure 3D). H3S10p levels compared to a vector only (V) control (Figure 5A). As expected, Pol IIp recruitment to these promoters increased However, when MSK-1 binding was monitored, only partial transiently after activation and diminished after stimulus with- impairment was detected in the KR mutant samples (Figure 5A). drawal (Figure 3E), consistent with the recruitment profiles of This suggests that the histone kinase MSK-1 may contribute to PKC-q and MSK-1. These data suggest that PKC-q and H3S10p on the CD69 promoter independently of PKC-q . Inhibi- MSK-1 require the continual presence of activating signals to tion of MSK-1 with H89 successfully blocked H3S10p at the maintain their association with chromatin. CD69 promoter (Figure 5B). This is consistent with previous studies of H89 for other early response genes (Thomson et al., PKC-q Colocalizes with the Adaptor Molecule 14-3-3z 1999) and can be attributed to the lack of MSK-1 recruitment on Transcriptionally Active Genes to this promoter in the presence of H89 (Figure 5C). In support Previous studies have shown that PKC-q and 14-3-3 adaptor of our earlier transfection studies, pretreatment with the PKC-q family members are interacting partners (Meller et al., 1998). ATP inhibitor Rottlerin (Springael et al., 2007) failed to block The 14-3-3z adaptor protein interacts with chromatin by binding H3S10p at the CD69 promoter (Figure 5D). In addition, Rottlerin to H3S10p, and this event is a prerequisite for inducible gene did not diminish PKC-q binding on this gene (Figure 5E). regulation (Winter et al., 2008). Halfway ChIP on nuclear extracts Together, these data suggest that the region encompassing showed that 14-3-3z associates with histone H3 (Figure 4A) and the KR mutation within the PKC-q catalytic core may be involved PKC-q (Figure 4B) in NS and in 4 hr-activated T cells. ChIP re- in tethering this kinase to chromatin. vealed that this adaptor protein is recruited in a stimulus-depen- Since active PKC-q appears to be essential for recruitment of dent manner to the CD69 and TNF-a gene promoters with transcription components, we assessed the ability of this kinase maximal occupancy at 8 hr after T cell activation (Figure 4C). to influence chromatin accessibility by CHART-PCR. Pretreat- Sequential ChIP with antibodies against PKC-qp and 14-3-3z ment with Rottlerin abrogated the chromatin accessibility across revealed the coexistence of these proteins on both the CD69 the CD69 and TNF-a proximal promoters as well as inhibiting and TNF-a gene promoters and this association increased in their messenger RNA (mRNA) production after T cell activation a stimulus-dependent manner (Figure 4D). (Figure 5F and Figure S2, respectively). In contrast, H89

Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. 707 Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

A B C 55 120 CD69 CD69 (-0.15 Kb) 9 (-0.15 Kb) CD69 0 hr PKC- 44 90 4 hr P/I PKC- p 6 24 hr P/I MSK-1 33 60

22 PKC- 3 30 ChIP (NS set to 1) set to ChIP (NS ChIP (NS set to 1) set to ChIP (NS

H3-enriched kinase H3-enriched 11 ChIP enrichment ratio enrichment ChIP PKC- MSK1 -enriched 0 0 0 NS ST NS ST -1 0 1 2 3 D E F Primer Position (Kb) 12 CD69 60 CD69 (-0.15 Kb) CD69 0 hr 7 4 hr P/I 6 9 24 hr P/I 5 40 6 4 MSK-1 3 20 3 2 PKC- (phospho) ChIP set (NS 1)to ChIP enrichment ratio enrichment ChIP ChIP enrichment ratio enrichment ChIP 1 II-enrichedPol kinase 0 0 0 -1 0 1 2 3 0 4 12 24 0 4 12 24 PKC- ChIP MSK-1 ChIP Primer Position (Kb) -0.15 Kb +0.5 Kb Stimulation time (hr) Primary human CD4+ T cell subsets G 250 IL-2 Promoter J DAPI PKC- Merged 200

150

100

50 Naïve

0 PKC- ChIPenrichment ratio

Naïve emory

Resting MemoryEffector M H Activated memory 8000 IL-2 promoter Resting 6000 Memory

4000

2000 T cell subsets: cell T

0 +

H3K9ac ChIP enrichment ratio enrichment ChIP H3K9ac Effector Naïve

CD4 Memory

Resting Memory Effector MemoryActivated memory I 15000 IL-2 cDNA

10000 Activated

5000 Memory

(Normalized to GAPDH) 0 IL-2 Arbitrary Copy Number

Naïve

Resting MemoryEffector Memory Activated Memory

Figure 2. PKC-q Forms an Active Transcription Complex with Pol II and MSK-1 across Proximal Regulatory Regions of Inducible Genes (A) Sequential ChIP of NS or ST Jurkat T cells with anti-histone H3 antibody, then anti-PKC-q, anti-PKC-qp, and anti-MSK-1 antibodies. Real-time PCR of CD69 (0.15 Kb). (B) Sequential ChIP as for (A) with anti-PKC-q followed by anti-MSK-1 antibody. (C and D) PKC-q (C) and MSK-1 (D) ChIPs with Jurkat T cells stimulated for the times indicated.

708 Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

q Table 1. PKC-q Is Recruited to the Proximal Promoters of Genome-wide ChIP-on-Chip Reveals that PKC- p Inducible Genes in a Stimulus-Dependent Manner Coresides with Pol II on the Promoter and Transcribed Gene Promotera Regions of Active Genes and MicroRNAs Having identified a nuclear role for PKC-q in the active tran- Stimulation CD69 Heparanase TNF-a IFN-g GAPDH scription of several key immune genes, we determined whether NS 1 1 1 1 1 PKC-q’s association with chromatin is a genome-wide 4 hr P/I 6.0 3.0 4.0 7.0 1.1 phenomenon in activated T cells. ChIP-on-chip (GEO acces- a Fold change in PKC occupancy with respect to nonstimulated Jurkat sion number GSE26035) identified only 141 genes bound by T cells. PKC-qp with a p value score of < 0.05 and an enrichment ratio > 5(Table S1). Constitutively expressed genes, including b-actin, treatment failed to inhibit the chromatin accessibility in the GAPDH, and cyclophilin A, were not enriched within the promoter regions after activation and had minimal effect on PKC-qp-bound subsets. The inducible immune response mRNA production (Figure 5F and Figure S2). Our results suggest genes, CD69, heparanase, TNF-a, and IFN-g, were all present that PKC-q activity, but not that of MSK-1, appears to facilitate in the PKC-qp-bound subset of genes, with enrichment ratios < chromatin accessibility and accompanying mRNA production 5 and p values < 0.05. Close examination of the classes of DNA of inducible immune response genes. This process seems to regions bound by PKC-qp showed that PKC-qp is enriched at be independent of serine-10 phosphorylation mediated by promoter, 50-transcribed, and downstream regions of a subpop-

MSK-1. A recent study found that PKC-b1 associates with chro- ulation of genes, with the majority of binding (61%) detected in matin and this dictates the action of the chromatin-modifying transcribed regions (Figure 6A). The top 50 ranked genes based LSD-1 in HeLa cells (Metzger et al., 2010). Thus, to on PKC-qp binding to either the 50-transcribed or promoter further support the interplay between PKC-q with chromatin regions are depicted in Figure 6B (p < 0.01). Functional annota- associated proteins (MSK-1, Pol II, 14-3-3z, and LSD-1) in tion of the PKC-qp target genes (Figure 6A) revealed that this T cells, we performed PKC-q-specific RNA interference (RNAi) kinase interacts with genes linked to specific cellular in resting and activated T cells in conjunction with ChIP. PKC-q processes, such as cell-cell communication, programmed was inhibited by RNAi treatment, as monitored by PKC-q Taq- cell-death, signal transduction, mesenchymal cell develop- Man real-time PCR (Figure S3). Inhibition of PKC-q with RNAi ment, and differentiation (Figure 6C). The top-ranking successfully diminished binding of this kinase to the IL-2 PKC-qp-bound gene in the promoter was the innate immune promoter, as well as inhibiting MSK-1, Pol II, 14-3-3z, and receptor TREML4, which has been shown to play a crucial LSD-1 occupancy on this target gene (Figure 5G). To gain a better role in fine-tuning inflammatory responses (Ford and McVicar, mechanistic understanding of the interplay of PKC-q with MSK- 2009). In addition, PARP-10, a novel interacting partner of 1, we performed ChIP for PKC-q, MSK-1, Pol II, 14-3-3z, and myc that is dysregulated in numerous human cancers (Yu LSD-1 with MSK-1 RNAi. Depletion of MSK-1 further supported et al., 2005) was identified as one of the top ranking genes in the interdependence of PKC-q and MSK-1, which also appeared the transcribed region. A subset of the top 50 ranked important for the inducible association of Pol II, 14-3-3z, and PKC-qp-bound genes encoded signaling proteins (e.g., LCK, LSD1 (Figure 5H). 14-3-3 RNAi only led to marginal diminution DGKD, CDC25A, CAMK2A, and SKIP), adhesion molecules, of the PKC-q chromatin associated complex at the ChIP level as well as the chromatin-associated enzyme RPS6KA1 and (Figure S4) and there was no effect on transcription of the induc- the epithelial to mesenchymal transition (EMT) inducer ible gene, IL-2, in these cells (Figure S4). This is not surprising DNMT3A (Doehn et al., 2009). PKC-qp also was found to asso- considering that this RNAi does not specifically target the 14- ciate with genes encoding proapoptotic molecules, such as 3-3z isoform, which we have shown thus far to form a complex YWHAZ, a 14-3-3 family member, and DDAH2. This kinase with PKC-q on a chromatin template. In contrast, depletion of bound to the regulatory regions of genes that express proteins either PKC-q or MSK-1 resulted in decreased transcription of critical for EMT progression and cancer development, including the IL-2 cytokine, consistent with the disassembly of an active the WNT family (WNT5A, WNT3, WNT2B, WNT7B, and complex in the absence of either of these two proteins (Figures WNT10A), cadherin members, and to oncogenic transcription 5I and 5J, respectively). These RNAi data complement our inhib- factors such as RUNX2 and KLF9. itor and PKC-q mutant data, which all demonstrate that PKC-q Since our initial studies suggested that the presence of association with chromatin is likely to be required for the stable PKC-qp and Pol II were intimately linked to the promoters and binding of MSK-1, Pol II, 14-3-3z, and LSD-1. transcribed regions of inducible genes in T cells, we performed

(E) ChIPs as above with anti-PKC-qp and primers for CD69 (0.15 and +0.5 Kb). (F) Sequential ChIP as for (A) with anti-Pol II then anti-PKC-q and anti-MSK-1. (G and H) ChIPs as above on ex vivo human CD4+ T cells with anti-PKC-q (G) or anti-H3K9ac antibody on IL-2 (H) (0.15 Kb). (I) Total RNA from CD4+ T cell subsets and TaqMan real-time PCR for IL-2. Data are expressed as arbitrary copy number normalized to GAPDH and representative of the mean ± standard error (SE) of two independent experiments. (J) Fixed nuclei from CD4+ T cell subsets stained as in Figure 1C. All ChIP data were calculated as n-fold enrichment ratio of immunoprecipitated DNA relative to no antibody control, normalized against total input DNA. NS ChIP enrichment was set to 1 for sequential ChIP. ChIP data with error bars are shown as the mean ± SE of two independent experiments, otherwise representative of at least two independent experiments performed in duplicate. See also Figure S1.

Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. 709 Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

Resting T cells Activated T cells Previously activated T cells A 4 hr P/I Stimuli withdrawal NS ST SW

40 1000 TNF- cDNA 3000 CD69 cDNA IFN- cDNA 30 750 2000 20 500 1000 10 250 FC (relative to NS) to (relative FC FC (relative to NS) to (relative FC NS) to (relative FC

0 0 0 NS ST SW NS ST SW NS ST SW

B CD69 (-0.15 Kb) TNF- (-0.15 Kb) 40 200 300 IFN- (-0.15 Kb)

30 150 200 20 100 PKC- PKC- PKC- 100 10 50 ChIP enrichment ratio enrichment ChIP ChIP enrichment ratio enrichment ChIP ChIP enrichment ratio enrichment ChIP 0 0 0 NS ST SW NS ST SW NS ST SW C 50 CD69 (-0.15 Kb) 40 7.5 TNF- (-0.15 Kb) IFN- (-0.15 Kb) 40 30 5.0 30 20 MSK-1 MSK-1 20 MSK-1 2.5 10 10 ChIP enrichment ratio enrichment ChIP ratio enrichment ChIP ChIP enrichment ratio enrichment ChIP 0 0 0.0 NS ST SW NS ST SW NS ST SW D CD69 (-0.15 Kb) 20 20 TNF- (-0.15 Kb) 15 IFN- (-0.15 Kb)

10 10 10 H2A.Z H2A.Z H2A.Z 5 ChIP enrichment ratio enrichment ChIP ratio enrichment ChIP ChIP enrichment ratio enrichment ChIP 0 0 0 NS ST SW NS ST SW NS ST SW

E CD69 (-0.15 Kb) TNF- (-0.15 Kb) 5000 3000 140 IFN- (-0.15 Kb) 120 4000 2000 100 3000 80 Pol IIPol P Pol IIPol P 2000 IIPol P 60 1000 40 1000 ChIP enrichment ratio enrichment ChIP ChIP enrichment ratio enrichment ChIP ChIP enrichment ratio enrichment ChIP 20 0 0 0 NS ST SW NS ST SW NS ST SW

Figure 3. PKC-q Recruitment Is Dependent on Active Transcription Jurkat T cells NS, ST, and SW. Depicted schematically (A, top). TaqMan real-time PCR for CD69, TNF-a, and IFN-g (A, bottom). Fold change (FC) relative to NS. Data show the mean ± SE of two independent experiments performed in duplicate. ChIPs on NS, ST and SW with anti-PKC-q (B), anti-MSK-1 (C), anti-histone H2A.Z (D), and anti-Pol IIp (E) antibodies across CD69, TNF-a, and IFN-g (0.15 Kb). ChIP data are shown as the mean ± SE of two independent experiments performed in duplicate.

710 Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

A -Ab IP H3 IP NS 4 NS 4

14-3-3ζ (30 kDa)

B -Ab IP PKC- IP NS 4 NS 4

14-3-3ζ (30 kDa)

C 1600 1200 CD69 (-0.15 Kb) TNF- (-0.15 Kb) 1200 900

800 600 14-3-3 14-3-3

400 300 ChIP enrichment ratio enrichment ChIP ChIP enrichment ratio enrichment ChIP 0 0 NS 4 8 NS 4 8

D 400 2000 CD69 (-0.15 Kb) TNF- (-0.15 Kb) 300 1500

200 1000 ChIP (NS set to 1) to set (NS ChIP 100 1) to set (NS ChIP 500 PKC- p-enriched 14-3-3 PKC- p-enriched 14-3-3 0 0 NS 4 8 NS 4 8

E CD69 (-0.15 Kb) Heparanase (-0.15 Kb)

oitar 1200 400 oitar 200 t tnemhcirnePIhC 50 ne 900

40 m hcirnePIhC 30 600 20 300 10 0 0 VWTKRVWTKR VWTKRVWTKRVWTKR

PKC- 14-3-3 PKC- 14-3-3 Pol II

Figure 4. Recruitment of PKC-q to Inducible Gene Promoters Is Required for 14-3-3z Occupancy (A and B) Halfway ChIP with anti-histone H3 (A) or anti-PKC-q (B) antibody with NS or 4 hr P/I-treated Jurkat T cells and immunoblot with anti-14-3-3z. (–Ab) is in absence of antibody. (C) ChIPs on Jurkat T cells NS or stimulated with P/I for 4 or 8 hr across CD69 and TNF-a (0.15 Kb). (D) Sequential ChIP as in Figure 2A, but with anti-PKC-qp and anti-14-3-3z antibody. (E) ChIPs on Jurkat T cells transfected with V, WT, or KR. Real-time PCR on DNA recovered with anti-PKC-q, anti-14-3-3z, or anti-Pol II across CD69 and heparanase (0.15 Kb). ChIP data are shown as the mean ± SE of two independent experiments performed in duplicate.

ChIP-on-chip of genome-wide Pol II binding in activated T cells. Of particular interest, PKC-q bound to the proximal regulatory Consistent with our initial findings, 62% of the 141 genes that regions of 60 microRNAs in activated T cells (listed in Table S2). bound PKC-qp(Table S1) were also occupied by Pol II, with PKC-q binding was most prevalent in the promoter, as opposed both molecules colocalizing to the same region of most genes to the coding region, of the microRNA genes and this was closely (Figure 6D). correlated with Pol II binding (Figure 6E and Table S2). The top

Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. 711 Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

A BC 400 CD69 (-0.15 Kb) CD69 (-0.15 Kb) 200 CD69 (-0.15 Kb) 18 4 150 15 3 100 12 9 2 MSK-1 -81 50 6 -86 -81 Histone H3S10p Histone -92 1 3 ChIP enrichment ratio enrichment ChIP ChIP enrichment ratio enrichment ChIP ChIP enrichment ratio enrichment ChIP 0 V WT KR V WT KR V WT KR 0 0 012012 012012 H3S10P PKC- MSK-1 H89 H89 DFE CD69 (-0.15 Kb) TNF- (-0.15 Kb) 30 8 60 50 CD69 CD69 (-0.15 Kb) 25 (-0.15 Kb) 40 6 20 40 - 30 15 4 PKC 20 10 20 (relative to NS) to (relative

Histone H3S10p Histone 2 5 10 ChIP enrichment ratio ChIP enrichment ratio enrichment ChIP % change in accessibility in % change 0 0 0 0 012012 012012 NS 4 NS 4 NS 4 NS 4 NS 4 NS 4 Rottlerin Rottlerin - +ROTT +H89 - +ROTT +H89

100 G IL-2 Promoter I 2000 80

60 1500

40 1000 20 500 ChIP enrichment ratio enrichment ChIP 0

(Normalized to GAPDH) to (Normalized 0 IL-2 Arbitrary Copy Number NS ST NS ST RNAi) RNAi) RNAi) RNAi) RNAi) RNAi) RNAi) RNAi) RNAi) RNAi) Mock PKC- RNAi ST (Mock) ST (Mock) ST (Mock) ST (Mock) ST (Mock) NS (Mock) NS (Mock) NS (Mock) NS (Mock) NS (Mock) ST (PKC- ST (PKC- ST (PKC- ST (PKC- ST (PKC- NS (PKC- NS (PKC- NS (PKC- NS (PKC- NS (PKC- PKC- MSK-1 Pol II 14-3-3 LSD1

25 H IL-2 Promoter J 20 600 15

10 400 5 ChIP enrichment ratio enrichment ChIP 0 200

(Normalized to GAPDH) 0 IL-2 Arbitrary Copy Number NS ST NS ST ST (Mock) ST (Mock) ST (Mock) ST (Mock) ST (Mock) NS (Mock) NS (Mock) NS (Mock) NS (Mock) NS (Mock) Mock MSK-1 RNAi ST (MSK-1 RNAi) ST (MSK-1 RNAi) ST (MSK-1 RNAi) ST (MSK-1 RNAi) ST (MSK-1 RNAi) NS (MSK-1 RNAi) NS (MSK-1 RNAi) NS (MSK-1 RNAi) NS (MSK-1 RNAi) NS (MSK-1 RNAi) PKC- MSK-1 Pol II 14-3-3 LSD1

Figure 5. PKC-q Is a Prerequisite for Chromatin Accessibility on Inducible Immune Response Genes in T Cells, but Is Not Required for H3S10p on the CD69 Gene (A) H3S10p, PKC-q, or MSK-1 ChIP on Jurkat T cells transfected with V, WT, or KR across CD69 (0.15 Kb). (B and C) H3S10p (B) and MSK-1 (C) ChIPs on Jurkat T cells untreated or pretreated with H89 across CD69 (0.15 Kb). Average percentage inhibition values shown above graphs. (D and E) H3S10p (D) and PKC-q (E) ChIPs repeated as above but with Rottlerin.

712 Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

five ranked microRNAs in each binding category are listed in Fig- we employed pharmacological blockade and overexpression ure 6E. Importantly, a number of these microRNAs have been strategies. Jurkat T cells treated with the PKC-q inhibitor Rot- closely linked with EMT and cancer, including miR-200b/c, tlerin increased expression of these microRNAs, whereas treat- miR-141, and miR-429 from the same family and miR-9-1, ment with the MSK-1 inhibitor H89 had no effect (Figure 7B). miR-9-3, and miR-205. Conversely, overexpression of PKC-q in cells transfected with Based on the specific functions of a cohort of PKC-q target the vector overexpressing PKC-q (WT) decreased levels of these genes identified in the ChIP-on-chip data, we surmised that microRNAs compared to the vector control (V) samples (Fig- PKC-q may function as a mediator in EMT transition. To test ure 7C). Together these findings support the notion that PKC-q this hypothesis, cell morphology was examined in control negatively regulates these microRNAs in T cells. (mock) and PKC-q wild-type (WT) transfected MCF-7 cells by Recent studies show that miR-200c and miR-9 repress the phase-contrast microscopy 48 hr posttransfection. As is typical transcription of two transcription factors, Zeb1/Zeb2 for miR- of this human epithelial breast cancer cell line, mock-transfected 200c and Onecut2 for miR-9 (Gregory et al., 2008; Plaisance cultures formed tightly packed clusters of cobblestone-shaped et al., 2006). Given that these transcription factors play a key cells (Figure 6F). In contrast, overexpression of PKC-q caused role in the regulation of inducible genes (Furuno et al., 2008), a clear morphological change, with transfected cells acquiring such as IL-2, we hypothesized that such microRNAs may partic- an irregular but elongated shape and loss of cell-cell contact ipate in the regulation of inducible immune response genes. (Figure 6F). There was visibly less cell clustering observed, Inducible profiles were measured in Jurkat with cell scattering suggesting increased motility. The effect of T cells after transfection of two of the most significant PKC-q- overexpression of PKC-q was comparable to that obtained by bound synthetic microRNA precursors (miR-200c and miR-9). treating cells with TNF-a (Figure 6F), which has been reported At 48 hr posttransfection, with either the synthetic Pre-miRs previously to induce motility and invasive behavior in MCF-7 or a scrambled Pre-miR (mock), total RNA was extracted for cells (Yin et al., 2009). Cell motility was further evaluated by an TaqMan real-time qPCR either before (NS) or after stimulation in vitro wound-healing assay. In PKC-q transfected cells, with P/I for 2 hr (ST). In parallel, miR-1 Pre-miR was used as substantial wound closure (80%, Figure 6G and Figure S5) a positive control to show that it successfully reduces PTK9 tran- was observed after 60 hr, compared with wound closure in script levels (Figure S6A). Enforced expression of Pre-miR-200c mock-transfected cultures (20%, Figure 6G and Figure S5). markedly increased IL-2 and IFN-g expression in stimulated Ju- PKC-q induced cell motility more effectively than TNF-a alone, rkat T cells (Figure 7D). The heparanase gene exhibited elevated which induced only 40% of wound closure in mock- expression in both resting and activated T cells as a conse- transfected cells (Figure S5); however, TNF-a did not further quence of Pre-miR-200c overexpression (Figure 7D). This is enhance the effect of PKC-q (Figure S5). Our genome-wide consistent with our previous report that heparanase has a high analysis, in conjunction with our functional studies, suggests level of basal transcription (Sutcliffe et al., 2009). In contrast, that PKC-q is coupled to processes that regulate EMT and miR-200c had little effect on TNF-a and CD69 mRNA production microRNA regulation. (Figure 7D). Previous studies have shown that miR-200c reduces transcript levels of the transcriptional repressors Zeb1/Zeb2, PKC-q Participates in a Signaling Complex consequently elevating E-cadherin mRNA production (Gregory that Negatively Regulates MicroRNAs et al., 2008). Since Zeb1 is also a critical repressor of IL-2 in MicroRNAs are small noncoding RNAs that generally act as T cells, we tested whether Zeb1 is also a target of miR-200c posttranscriptional repressors of gene expression either via in T cells. Consistent with previous findings, transfection of reduction of their target transcripts, prevention of the target Jurkat T cells with Pre-miR-200c had markedly repressed protein translation, or both. Since we have identified a distinct Zeb1 transcription, compared with mock samples (Figure 7E). cohort of microRNAs in our ChIP-on-chip that are direct gene Zeb2 was not found to be a target of miR-200c in Jurkat targets of PKC-q, we investigated the transcript expression T cells (data not shown). characteristics of the three microRNAs that exhibited the most Transfection of Pre-miR-9 significantly increased CD69 and significant PKC-q binding. When the transcript abundances of TNF-a expression in Jurkat T cells, whereas IL-2 and IFN-g these microRNAs (microRNA-9, -200c, and -183) were mea- were unaltered compared to the corresponding mock-trans- sured in resting and P/I-stimulated T cells, they decreased in fected control (Figure 7F). Heparanase also displayed increased a stimulus-dependent manner (Figure 7A). To confirm PKC-q expression in both resting and activated T cells when Pre-miR-9 mediated regulation of these microRNAs at the transcript level, was overexpressed (Figure 7F). Consistent with previous

(F) CHART on Jurkat nuclei from NS or 4 hr P/I left untreated or pretreated with Rottlerin (ROTT) or H89 across CD69 and TNF-a (0.15 Kb). Data are graphed as the percentage change in accessibility and plotted as the mean ± SE of two independent experiments performed in duplicate. (G) Jurkat T cells transfected with scrambled RNAi (mock) or PKC-q RNAi and subsequently left untreated (NS) or stimulated for 2 hr with P/I (ST). ChIP with PKC-q, MSK-1, Pol II, 14-3-3z, and LSD1 antibodies across IL-2 (0.15 Kb). (H) MSK-1 RNAi as in (G). ChIP data with error bars are shown as the mean ± SE of two independent experiments, otherwise representative of at least two individual experiments performed in duplicate. (I and J) IL-2 TaqMan real-time PCR for PKC-q RNAi-treated cells (I) and MSK-1 RNAi-treated cells (J) or mock samples. Data expressed as arbitrary copy number normalized to GAPDH and representative of the mean ± SE of two independent experiments performed in duplicate. See also Figure S2, S3, and S4.

Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. 713 Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

100 PKC bound genes A 1 4 D 80 PKC & Pol II bound genes 3% 1% 54 60 38% 40 Gene number Gene 82 20 58% 0

Promoter 5'-transcribed 3'-transcribed Unknown known Promoter Un transcribed 5'-transcribed 3'- B E

10.0 60

7.5 Binding: Transcribed regions 50

40

ratio 5.0 30 PKCTheta normalised 2.5 20 0.0 PKC & Pol II-bound Gene Number 10 MLN SKIP

TLR9 0 TMIE RNF25 SURF1 RING1 PTCH1 PRRX2 TCEA3 EPHA8 EPHA4 PKC-bound ZNF454 ZNF775 DDAH2 ZNF165 PARP10 LRRC42

SORCS2 Promoter MAPK15 DNMT3A DNMRT3 CAM2KB RPS6KA1

KLHDC8A hsa-mir-125a hsa-mir-9-1 Inside hsa-mir-141 hsa-mir-124a-1 hsa-mir-153-1 Downstream 7.5 hsa-mir-152 hsa-mir-150 hsa-mir-149 hsa-let-7b hsa-mir-129-1 Binding: Promoter regions hsa-mir-200b/c hsa-mir-93 hsa-mir-153-2 5.0 hsa-mir-182 hsa-mir-483 ratio PKCTheta normalised 2.5 (-TNF) (+TNF) F 0.0 LCK PXK KLF9 LCN2 IL17B CRAT WDR5 ZNF79 DGKD RGS12 PCBP4 FRAS1 RUNX2 GRRP1 ORC4L Mock POU4F2 ENTPD3 CDC25A EXOSC4 YWHAZ CNKSR1 WNT10A TREML4 CAMK2A CACNA1B C

Biological process and molecular function p-value Protein domain specific binding 4.11E-08 WT Anti-apoptosis 1.47E-07

Cell communication 7.05E-07 G Regulation of programmed cell-death 7.0E-05 Mock Cell-cell signalling 6.11E-05 0h 60h Overlay Cell-cell signalling during cell fate commitment 6.11E-05

Multicellular organismal movement 6.56E-05

Cell surface receptor linked signal transduction 8.52E-05

Neural crest cell migration 1.61E-04 Signal transduction 3.03E-04 WT Protein tyrosine kinases 3.29E-04 0h 60h Overlay Mesenchymal cell development 5.85E-04

Mesenchymal cell differentiation 5.85E-04

Establishment of protein localisation 8.76E-04

DNA replication 3.67E-04

Figure 6. Direct Targets of PKC-q Identified by Genome-wide ChIP-on-chip (A) Pie chart of PKC-qp bound genes (ratio > 5-fold, p < 0.05, n = 2). (B) The fifty most significant (p < 0.01, n = 2) bound PKC-q targeted genes within transcribed (top) and promoter (bottom) regions. (C) Table of most significant biological and molecular functions of PKC-q targeted genes from (A).

714 Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

findings, miR-9 appears to mediate its affect by inhibiting tran- PKC-q and MSK-1 appear to be highly dependent on one scription of the transcription factor onecut2 in T cells (Fig- another to remain tethered to chromatin in activated T cells. ure S6B). Transfection of the Pre-miR200c into Jurkat T cells Knockdown of either kinase resulted in the disassembly of the that were subsequently activated with P/I for 2 hr inhibited active transcription complex at the IL-2 promoter. MSK-1 has Zeb-1 occupancy on the IL-2 promoter and PKC-q concomi- previously been shown to modify chromatin architecture by tantly increased in occupancy (Figure 7G). This correlates with phosphorylating H3S10 on the nucleosomes of the immediate- the ability of PKC-q to positively regulate its target genes. Trans- early genes (Thomson et al., 1999). While the MSK-1 inhibitor, fection of the Pre-miR200c also diminished the repressive mark H89, reduced H3S10p across the CD69 promoter, levels were H3K9me2 across the IL-2 promoter, collectively correlating with unaffected by PKC-q activity blockade by Rottlerin or by the increased transcription of this gene (Figure 7D). kinase-dead PKC-q mutant. This is in contrast to another study that has shown that histone H3 phosphorylation of osmotic- DISCUSSION stress response genes is blocked by Rottlerin (Burkhart et al., 2007). It is possible that the role of PKC-q is gene and cell It is widely understood that protein kinases are cytoplasmic specific, as has been demonstrated for nuclear p38a (Simone intermediates of signal transduction cascades, relaying informa- et al., 2004). While PKC proteins have been shown to phosphor- tion from the cell surface to the nucleus for modulation of gene ylate histones in vitro, the exact nature of the histone posttrans- expression programs. We have now shown that the signal trans- lational modifications (PTMs) induced by PKC-q at activated duction kinase PKC-q, which is predominantly expressed in T cell gene targets remains to be identified. T cells and is best known for its critical role in CD4+ T cell The specific mechanisms that control the recruitment of sig- responses, belongs to an emerging class of signaling kinases naling kinases to chromatin remain elusive. Numerous models in mammalian cells that possess additional roles beyond modifi- have been proposed to explain the way in which signaling cation of DNA binding regulatory proteins. We provide evidence kinases are targeted to DNA (Pokholok et al., 2006). PKC-q of a nuclear role for PKC-q as an integral component of transcrip- and MSK-1 could be recruited to chromatin by PTMs or by tran- tion complexes assembled at the control regions of immune scription factors (TFs) in a T cell activation-dependent manner. genes and genes encoding microRNAs. Specifically, we have Detailed TFBS analyses of PKC-q gene targets identified by shown that recruitment of PKC-q to these inducible T cell genes ChIP-on-chip revealed a cluster of TFs that are overrepresented (1) is dependent on continuous T cell activation signals, (2) in this gene subset (data not shown). Thus, PKC-q and MSK-1 requires the catalytically active domain of PKC-q, (3) forms an may associate with gene promoters through direct interactions active transcription complex with MSK-1, Pol II, 14-3-3z, and with TFs, such as MAZ, Sp1, EKLF, and the NF-kB family LSD1, and (4) is required for inducible chromatin accessibility members (data not shown). In yeast, the Hog1 kinase binds to at proximal promoters. These results parallel the previously TFs bound to target gene promoters and can also stimulate unrecognized nuclear roles for the NF-kB signaling kinases recruitment of Pol II, associate with components of the elonga- IKK-a and NIK (Anest et al., 2003; Yamamoto et al., 2003) and tion complex and regulate transcription initiation (Alepuz et al., the protein kinases A-Cdk2 identified in promoter 2003; Proft et al., 2006). As such, kinases may enter the coding complexes on mammalian genes (Lawrence et al., 2008). regions of genes by ‘‘piggybacking’’ with Pol II (Edmunds and The enzymatically active form of PKC-q bound transiently to Mahadevan, 2006). It is possible that PKC-q may travel with the promoter region of CD69 in response to T cell activation, the Pol II elongation complex through the coding regions of but showed even greater binding to the transcribed region of genes upon T cell activation. In support, we have shown that the gene, which persisted even after 24 hr of T cell stimulation. recruitment of PKC-q and MSK-1 to the transcribed regions of In contrast, the PKC-q kinase inactive mutant was not recruited the heparanase and CD69 genes is dependent on Pol II elonga- to CD69, suggesting that the catalytic activity of PKC-q is tion, as treatment of cells with Actinomycin D prevented the required for binding to chromatin. This is also true of other yeast recruitment of these kinases beyond the transcriptional start protein kinases, whose stable association with chromatin in site of the CD69 gene (data not shown). Moreover, our ChIP- promoter complexes requires them to be in an active form (Law- on-chip data shows a close correlation between the occupancy rence et al., 2008). ChIP-on-chip further showed that PKC-qp profiles of both PKC-q and Pol II in their active states. We can localizes to the promoter and transcribed regions of active speculate that PKC-q may behave like a chromatin remodeller genes. Recent work has also shown that protein kinases, such or structural adaptor by traveling with Pol II through the coding as the stress-induced MAP kinase, Hog1, associate with the regions of genes upon transcriptional activation. Indeed, we transcribed regions of activated target genes (Alepuz et al., found that PKC-q was essential for inducing chromatin accessi- 2001; Proft et al., 2006). These data suggest that such features bility across the proximal promoters of several immune genes. of PKC-q may facilitate efficient transcriptional output by binding An alternate means of modulating gene expression is to gene control regions in the nucleus. post-transcriptionally through the action of microRNAs. Our

(D) Correlation between PKC-q with Pol II-bound genes from duplicate microarray experiments. (E) PKC-q targeted microRNAs (p < 0.05, n = 2) either PKC-q bound or PKC-q & Pol II bound. (F) Phase-contrast microscopy of MCF-7 cells transfected with mock or WT and left untreated or treated for 24 hr with TNF-a. (G) MCF-7 cells transfected with mock or WT and analyzed by the wound-healing assay. Images were taken at zero hr (red line) and after 60 hr (green line). See also Tables S1 and S2 and Figure S5.

Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. 715 Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

MicroRNA-9 cDNA MicroRNA-200c cDNA MicroRNA-183 cDNA A 1000 90 140

750 105 60 500 70 30 250 35 Arbitrary copyno. (normalised to RNU6B) 0 0 0 0 0.5 1.5 4 6 8 12 24 0 0.5 1.5 4 6 8 12 24 0 0.5 1.5 4 6 8 12 24 Stimulation time (hr) Stimulation time (hr) Stimulation time (hr)

MicroRNA-9 cDNA MicroRNA-200c cDNA MicroRNA-183 cDNA

B 1200 80 180

900 60 120 600 40 60 300 20 Arbitrary copyno. (normalised to RNU6B) 0 0 0 (-) ROTT H89 (-) ROTT H89 (-) ROTT H89 MicroRNA-9 cDNA MicroRNA-200c cDNA MicroRNA-183 cDNA

C 250 30 30

200 20 20 150

100 10 10

Arbitrary copyno. 50 (normalised to RNU6B) 0 0 0 VWT VWT VWT

D 200000 miR-200c E Mock (scrambled Pre-miR) 80000 150000 60000

100000 40000

50000 20000 Arbitary Copy Number (normalised to GAPDH) (Normalize to GAPDH) 0 0 Zeb1 Arbitary Copy Number Copy Arbitary Zeb1 ST ST NS ST NS ST NS ST NS ST NS ST NS NS mir-200c Mock CD69 IL-2 TNF IFN HPSE 800 F G 400 IL-2 Promoter 150000 200 160 miR-9 100000 Mock (scrambled Pre-miR) 120 80

50000 40 ChIP enrichment ratio enrichment ChIP 0 Arbitary Copy Number (normalised to GAPDH)

0 200c 200c 200c Mock Mock T Mock NS ST NS S NS ST NS ST NS ST CD69 IL-2 TNF IFN HPSE ZEB1 PKC- H3K9me2

Figure 7. Functional Analysis of a Subset of MicroRNAs Targeted by PKC-q in T Cells (A) TaqMan microRNA complementary DNA (miR-9, miR-200c, and miR-183) from resting and activated Jurkat T cells for time points indicated. (B) As in (A), but in the presence or absence of Rottlerin (ROTT) or H89.

716 Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

ChIP-on-chip data showed that PKC-q bound preferentially to mined whether nuclear roles in gene expression are intrinsic to all the control regions of not only protein-coding genes, but also PKC family members. the promoter regions of microRNAs. The mRNA expression of q three microRNA targets of PKC- (miR-200c, miR-9, and miR- EXPERIMENTAL PROCEDURES 183) were inversely correlated with both T cell activation and with overexpression of PKC-q, and were upregulated in Cell Culture response to Rottlerin (but not MSK-1) treatment of cells, sug- Human Jurkat T cells and mammary adenocarcinoma MCF-7 cells were gesting a role for PKC-q in the negative regulation of these cultured and stimulated as previously described (Sutcliffe et al., 2009). For RNAs. Furthermore, enforced expression of miR-200c and H89 (Sigma-Aldrich) and Rottlerin (Calbiochem) studies, cells were pretreated with 15 mM of each for 1 hr prior to stimulation. In the stimulation withdrawal miR-9 resulted in the induction of activated T cell genes, sug- experiment, cells were washed extensively 5 times in warm complete media gesting that these microRNAs are essential for inducible gene and recultured in medium without stimulus for 16 hr (SW). regulation. Previous studies have shown that miR-200c and miR-9 may operate via the regulation of T cell repressor proteins, Immunofluorescence Zeb1, Blimp1, and Onecut2. Indeed, enforced expression of Intracellular staining on 1 3 106 whole cells or isolated nuclei was performed miR-200c repressed Zeb1 expression in both resting and acti- as detailed previously (Rao et al., 2001). These were fixed in 2% paraformalde- vated T cells, while miR-9 successfully repressed Onecut2 and hyde (pH 7–8) and centrifuged onto 0.1% poly-L-lysine coated coverslips. Blimp1 (N.S. and A.D.K., unpublished data). Given that Zeb1 Whole cells or nuclei were permeabilized with 1% Triton X-100 and incubated and several of these PKC-q-bound microRNAs are central with blocking buffer for 45 min at room temperature. Primary and secondary players in tumor metastasis (Gregory et al., 2008), we used antibody details are provided in the Supplemental Experimental Procedures. wound-healing assays on human epithelial breast cancer cells to show that PKC-q may play a functionally relevant role in Total RNA Isolation and Quantitative Real-Time PCR EMT processes. Total RNA was extracted, and real-time PCR was performed as previously described (Sutcliffe et al., 2009). MicroRNA assays were performed with the Our findings lend support to a model (Figure S7) whereby TaqMan MicroRNA Reverse Transcription Kit (ABI 4366596). MicroRNA data chromatin anchored PKC-q is likely to regulate inducible T cell were normalized to RNU6B. gene transcription by two opposing mechanisms: (1) positive regulation of inducible immune response genes as an active Chromatin Accessibility by Real-Time PCR participant in chromatin-tethered transcription complexes in Accessibility of DNA to digestion with micrococcal nuclease (MNase) was addition to its signaling role and (2) negative regulation of genes analyzed with chromatin accessibility by real-time PCR (CHART-PCR) as encoding transcriptional repressor proteins through microRNA- described previously (Rao et al., 2001; Sutcliffe et al., 2009). mediated processes. Since the overall outcome is increased transcription of inducible genes, it is plausible that PKC-q func- ChIP Assays tions as a modulator to fine-tune inducible gene regulation, ChIPs were performed according to the protocol supplied by Upstate Biotech- both to initiate transcription, but also to control the response nology, as previously detailed (Sutcliffe et al., 2009). Antibody details are by increasing the level of cytokine repressor proteins by downre- provided in the Supplemental Experimental Procedures. gulating microRNA expression in a T cell activation-dependent manner. The timing of these competitive processes would ulti- ChIP-on-Chip mately prevent aberrant transcription ensuring the production Catalytically active PKC-q and Pol II ChIP DNA generated from P/I-stimulated of appropriate levels of cytokines required for T cell activation. Jurkat T cells were pooled from five independent ChIPs that were first individ- In summary, we have identified an alternate mechanism of ually validated by real-time PCR. Pooled DNA was subsequently amplified based on one round of the whole genome amplification method with the action of the signal transduction kinase, PKC-q, as a component WGA2 kit (Sigma-Aldrich). Duplicate samples were hybridized onto Agilent of the nuclear transcription machinery assembled at gene human promoter microarrays that span 17,000 well-characterized genes control regions in activated T cells. Future studies will aim to (5.5 to +2.5 Kb from TSS) from the , as described in the identify other components of the nuclear PKC-q transcription Mammalian ChIP-on-chip protocol (version 9.1; Agilent Technologies). complex, and to determine the way in which the nuclear and cytoplasmic actions of PKC-q are coordinated and integrated ACCESSION NUMBERS to achieve the appropriate physiological response, especially in human memory T cell subsets. Finally, it remains to be deter- Microarray data have been under GEO accession number GSE26035.

(C) As above from Jurkat T cells transfected with V or WT. Data are expressed as arbitrary copy number normalized to RNU6B and are representative of the mean ± SE of three independent experiments. (D) Synthetic miR-200c or scrambled Pre-miR (mock) transfected into Jurkat T cells and left untreated (NS) or stimulated for 2 hr (ST) with P/I. TaqMan real-time PCR for CD69, IL-2, TNF-a (TNF), IFN-g (IFN), and heparanase (HPSE). (E) Zeb1 real-time PCR on samples as in (D). (F) Repeated as for (D) with synthetic miR-9. Data are expressed as arbitrary copy number normalized to GAPDH and are representative of the mean ± SE of two independent experiments performed in duplicate. (G) Zeb1, PKC-q, and H3K9me2 ChIP on Jurkat T cells transfected with miR-200c or mock and stimulated 2 hr P/I. ChIP data are shown as the mean ± SE of two independent experiments performed in duplicate. See also Figure S6.

Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. 717 Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

SUPPLEMENTAL INFORMATION Inoue, M., Kishimoto, A., Takai, Y., and Nishizuka, Y. (1977). Studies on a cyclic nucleotide-independent protein kinase and its proenzyme in mammalian Supplemental Information includes Supplemental Experimental Procedures, tissues. II. Proenzyme and its activation by calcium-dependent protease seven figures, and two tables and can be found with this article online at from rat brain. J. Biol. Chem. 252, 7610–7616. doi:10.1016/j.molcel.2011.02.030. Isakov, N., and Altman, A. (2002). (theta) in T cell activation. Annu. Rev. Immunol. 20, 761–794. ACKNOWLEDGMENTS Lawrence, M.C., McGlynn, K., Shao, C., Duan, L., Naziruddin, B., Levy, M.F., and Cobb, M.H. (2008). Chromatin-bound mitogen-activated protein kinases We thank Andrew Pinson for extensive bioinformatic analysis and Ian Parish transmit dynamic signals in transcription complexes in beta-cells. Proc. Natl. for performing Genomatix analysis. We acknowledge Torsten Juelich for Acad. Sci. USA 105, 13315–13320. comments on this project. We are grateful to Vijay Randev for establishing Liu, Y., Graham, C., Li, A., Fisher, R.J., and Shaw, S. (2002). Phosphorylation of our collaboration with Agilent Technologies specialists and for continuous the protein kinase C-theta activation loop and hydrophobic motif regulates its support that has allowed smooth progression of the ChIP-on-chip experi- kinase activity, but only activation loop phosphorylation is critical to in vivo ments. S.R. is supported by an Australian National Health and Medical nuclear-factor-kappaB induction. Biochem. J. 361, 255–265. Research Council (NHMRC) New Investigator project grant. Marsland, B.J., Nembrini, C., Schmitz, N., Abel, B., Krautwald, S., Bachmann, Received: February 6, 2010 M.F., and Kopf, M. (2005). Innate signals compensate for the absence of Revised: July 20, 2010 PKC-theta during in vivo CD8(+) T cell effector and memory responses. Accepted: December 24, 2010 Proc. Natl. Acad. Sci. USA 102, 14374–14379. Published: March 17, 2011 Martelli, A.M., Sang, N., Borgatti, P., Capitani, S., and Neri, L.M. (1999). Multiple biological responses activated by nuclear protein kinase C. J. Cell. REFERENCES Biochem. 74, 499–521. Martiny-Baron, G., and Fabbro, D. (2007). Classical PKC isoforms in cancer. Alepuz, P.M., Jovanovic, A., Reiser, V., and Ammerer, G. (2001). Stress- Pharmacol. Res. 55, 477–486. induced map kinase Hog1 is part of transcription activation complexes. Mol. Meller, N., Altman, A., and Isakov, N. (1998). New perspectives on PKCtheta, Cell 7, 767–777. a member of the novel subfamily of protein kinase C. Stem Cells 16, 178–192. Alepuz, P.M., de Nadal, E., Zapater, M., Ammerer, G., and Posas, F. (2003). Metzger, E., Imhof, A., Patel, D., Kahl, P., Hoffmeyer, K., Friedrichs, N., Mu¨ ller, Osmostress-induced transcription by Hot1 depends on a Hog1-mediated J.M., Greschik, H., Kirfel, J., Ji, S., et al. (2010). Phosphorylation of histone recruitment of the RNA Pol II. EMBO J. 22, 2433–2442. H3T6 by PKCbeta(I) controls demethylation at histone H3K4. Nature 464, Anest, V., Hanson, J.L., Cogswell, P.C., Steinbrecher, K.A., Strahl, B.D., and 792–796. Baldwin, A.S. (2003). A nucleosomal function for IkappaB kinase-alpha in Pascual-Ahuir, A., Struhl, K., and Proft, M. (2006). Genome-wide location anal- NF-kappaB-dependent gene expression. Nature 423, 659–663. ysis of the stress-activated MAP kinase Hog1 in yeast. Methods 40, 272–278. Araki, Y., Fann, M., Wersto, R., and Weng, N.P. (2008). Histone acetylation Passalacqua, M., Patrone, M., Sparatore, B., Melloni, E., and Pontremoli, S. facilitates rapid and robust memory CD8 T cell response through differential (1999). Protein kinase C-theta is specifically localized on centrosomes and expression of effector molecules (eomesodermin and its targets: perforin kinetochores in mitotic cells. Biochem. J. 337, 113–118. and granzyme B). J. Immunol. 180, 8102–8108. Plaisance, V., Abderrahmani, A., Perret-Menoud, V., Jacquemin, P., Lemaigre, Birbach, A., Gold, P., Binder, B.R., Hofer, E., de Martin, R., and Schmid, J.A. F., and Regazzi, R. (2006). MicroRNA-9 controls the expression of Granuphilin/ (2002). Signaling molecules of the NF-kappa B pathway shuttle constitutively Slp4 and the secretory response of insulin-producing cells. J. Biol. Chem. 281, between cytoplasm and nucleus. J. Biol. Chem. 277, 10842–10851. 26932–26942. Burkhart, B.A., Kennett, S.B., and Archer, T.K. (2007). Osmotic stress-depen- Pokholok, D.K., Zeitlinger, J., Hannett, N.M., Reynolds, D.B., and Young, R.A. dent repression is mediated by histone H3 phosphorylation and chromatin (2006). Activated signal transduction kinases frequently occupy target genes. structure. J. Biol. Chem. 282, 4400–4407. Science 313, 533–536. de Nadal, E., and Posas, F. (2010). Multilayered control of gene expression by Proft, M., Mas, G., de Nadal, E., Vendrell, A., Noriega, N., Struhl, K., and Posas, stress-activated protein kinases. EMBO J. 29, 4–13. F. (2006). The stress-activated Hog1 kinase is a selective transcriptional elon- Doehn, U., Hauge, C., Frank, S.R., Jensen, C.J., Duda, K., Nielsen, J.V., gation factor for genes responding to osmotic stress. Mol. Cell 23, 241–250. Cohen, M.S., Johansen, J.V., Winther, B.R., Lund, L.R., et al. (2009). RSK Rao, S., Procko, E., and Shannon, M.F. (2001). Chromatin remodeling, is a principal effector of the RAS-ERK pathway for eliciting a coordinate measured by a novel real-time polymerase chain reaction assay, across the promotile/invasive gene program and phenotype in epithelial cells. Mol. Cell proximal promoter region of the IL-2 gene. J. Immunol. 167, 4494–4503. 35, 511–522. Saccani, S., Pantano, S., and Natoli, G. (2002). p38-Dependentmarking ofinflam- Edmunds, J.W., and Mahadevan, L.C. (2006). Cell signaling. Protein kinases matory genes for increased NF-kappa B recruitment. Nat. Immunol. 3,69–75. seek close encounters with active genes. Science 313, 449–451. Simone, C., Forcales, S.V., Hill, D.A., Imbalzano, A.N., Latella, L., and Puri, P.L. Ford, J.W., and McVicar, D.W. (2009). TREM and TREM-like receptors in (2004). p38 pathway targets SWI-SNF chromatin-remodeling complex to inflammation and disease. Curr. Opin. Immunol. 21, 38–46. muscle-specific loci. Nat. Genet. 36, 738–743. Furuno, K., Ikeda, K., Hamano, S., Fukuyama, K., Sonoda, M., Hara, T., Springael, C., Thomas, S., Rahmouni, S., Vandamme, A., Goldman, M., Sasazuki, T., and Yamamoto, K. (2008). Onecut transcription factor OC2 is Willems, F., and Vosters, O. (2007). Rottlerin inhibits human T cell responses. a direct target of T-bet in type-1 T-helper cells. Genes Immun. 9, 302–308. Biochem. Pharmacol. 73, 515–525. Gregory, P.A., Bert, A.G., Paterson, E.L., Barry, S.C., Tsykin, A., Farshid, G., Sutcliffe, E.L., Parish, I.A., He, Y.Q., Juelich, T., Tierney, M.L., Rangasamy, D., Vadas, M.A., Khew-Goodall, Y., and Goodall, G.J. (2008). The miR-200 family Milburn, P.J., Parish, C.R., Tremethick, D.J., and Rao, S. (2009). Dynamic and miR-205 regulate epithelial to mesenchymal transition by targeting ZEB1 histone variant exchange accompanies gene induction in T cells. Mol. Cell. and SIP1. Nat. Cell Biol. 10, 593–601. Biol. 29, 1972–1986. Griner, E.M., and Kazanietz, M.G. (2007). Protein kinase C and other diacylgly- Thomson, S., Clayton, A.L., Hazzalin, C.A., Rose, S., Barratt, M.J., and cerol effectors in cancer. Nat. Rev. Cancer 7, 281–294. Mahadevan, L.C. (1999). The nucleosomal response associated with imme- Hayashi, K., and Altman, A. (2007). Protein kinase C theta (PKCtheta): a key diate-early gene induction is mediated via alternative MAP kinase cascades: player in T cell life and death. Pharmacol. Res. 55, 537–544. MSK1 as a potential histone H3/HMG-14 kinase. EMBO J. 18, 4779–4793.

718 Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. Molecular Cell PKC-q Binds Target Genes and MicroRNAs in T Cells

Visnjic, D., and Banfic, H. (2007). Nuclear phospholipid signaling: phosphatidy- Yin, Y., Chen, X., and Shu, Y. (2009). Gene expression of the invasive pheno- linositol-specific and phosphoinositide 3-kinase. Pflugers type of TNF-alpha-treated MCF-7 cells. Biomed. Pharmacother. 63, 421–428. Arch. 455, 19–30. Yu, H., Fukami, K., Watanabe, Y., Ozaki, C., and Takenawa, T. (1998). Winter, S., Simboeck, E., Fischle, W., Zupkovitz, G., Dohnal, I., Mechtler, K., Phosphatidylinositol 4,5-bisphosphate reverses the inhibition of RNA tran- Ammerer, G., and Seiser, C. (2008). 14-3-3 proteins recognize a histone scription caused by histone H1. Eur. J. Biochem. 251, 281–287. code at histone H3 and are required for transcriptional activation. EMBO J. 27, 88–99. Yu, M., Schreek, S., Cerni, C., Schamberger, C., Lesniewicz, K., Poreba, E., Yamamoto, Y., Verma, U.N., Prajapati, S., Kwak, Y.T., and Gaynor, R.B. Vervoorts, J., Walsemann, G., Gro¨ tzinger, J., Kremmer, E., et al. (2005). (2003). Histone H3 phosphorylation by IKK-alpha is critical for cytokine- PARP-10, a novel Myc-interacting protein with poly(ADP-ribose) polymerase induced gene expression. Nature 423, 655–659. activity, inhibits transformation. Oncogene 24, 1982–1993.

Molecular Cell 41, 704–719, March 18, 2011 ª2011 Elsevier Inc. 719