Loss of PRC2 subunits primes lineage choice during exit of pluripotency.

Polycomb Repressive Complex 2 (PRC2) is crucial for the coordinated expression of genes during early embryonic development, catalyzing histone H3 lysine 27 trimethylation. Two distinct PRC2 complexes, PRC2.1 and PRC2.2, contain respectively MTF2 and JARID2 in embryonic stem cells (ESCs). In this study, we explored their roles in lineage specification and commitment, using single-cell transcriptomics and mouse embryoid bodies derived from Mtf2 and Jarid2 null ESCs. We observe that the loss of Mtf2 results in enhanced and faster differentiation towards cell fates from all germ layers, while the Jarid2 null cells are predominantly directed towards early differentiating precursors, with reduced efficiency towards mesendodermal lineages. These effects are caused by derepression of developmental regulators that are poised for activation in pluripotent cells and gain H3K4me3 at their promoters in the absence of PRC2 repression. Upon lineage commitment, the differentiation trajectories are relatively similar to those of wild-type cells. Together, our results uncover a major role for MTF2-containing PRC2.1 in balancing poised lineage-specific gene activation, whereas the contribution of JARID2-containing PRC2 is more selective in nature compared to MTF2. These data explain how PRC2 imposes thresholds for lineage choice during the exit of pluripotency.

Combinatorial transcription factor activities on open chromatin induce embryonic heterogeneity in vertebrates.

During vertebrate gastrulation, mesoderm is induced in pluripotent cells, concomitant with dorsal-ventral patterning and establishing of the dorsal axis. We applied single-cell chromatin accessibility and transcriptome analyses to explore the emergence of cellular heterogeneity during gastrulation in Xenopus tropicalis. Transcriptionally inactive lineage-restricted genes exhibit relatively open chromatin in animal caps, whereas chromatin accessibility in dorsal marginal zone cells more closely reflects transcriptional activity. We characterized single-cell trajectories and identified head and trunk organizer cell clusters in early gastrulae. By integrating chromatin accessibility and transcriptome data, we inferred the activity of transcription factors in single-cell clusters and tested the activity of organizer-expressed transcription factors in animal caps, alone or in combination. The expression profile induced by a combination of Foxb1 and Eomes most closely resembles that observed in the head organizer. Genes induced by Eomes, Otx2, or the Irx3-Otx2 combination are enriched for maternally regulated H3K4me3 modifications, whereas Lhx8-induced genes are marked more frequently by zygotically controlled H3K4me3. Taken together, our results show that transcription factors cooperate in a combinatorial fashion in generally open chromatin to orchestrate zygotic gene expression.

MTF2 recruits Polycomb Repressive Complex 2 by helical-shape-selective DNA binding.

ABSTACT: Polycomb-mediated repression of gene expression is essential for development, with a pivotal role played by trimethylation of histone H3 lysine 27 (H3K27me3), which is deposited by Polycomb Repressive Complex 2 (PRC2). The mechanism by which PRC2 is recruited to target genes has remained largely elusive, particularly in vertebrates. Here we demonstrate that MTF2, one of the three vertebrate homologs of Drosophila melanogaster Polycomblike, is a DNA-binding, methylation-sensitive PRC2 recruiter in mouse embryonic stem cells. MTF2 directly binds to DNA and is essential for recruitment of PRC2 both in vitro and in vivo. Genome-wide recruitment of the PRC2 catalytic subunit EZH2 is abrogated in Mtf2 knockout cells, resulting in greatly reduced H3K27me3 deposition. MTF2 selectively binds regions with a high density of unmethylated CpGs in a context of reduced helix twist, which distinguishes target from non-target CpG islands. These results demonstrate instructive recruitment of PRC2 to genomic targets by MTF2.

Activation of the antioxidant response in methionine deprived human cells results in an HSF1-independent increase in HSPA1A mRNA levels.

In cells starved for leucine, lysine or glutamine heat shock factor 1 (HSF1) is inactivated and the level of the transcripts of the HSF1 target genes HSPA1A (Hsp70) and DNAJB1 (Hsp40) drops. We show here that in HEK293 cells deprived of methionine HSF1 was similarly inactivated but that the level of HSPA1A and DNAJB1 mRNA increased. This increase was also seen in cells expressing a dominant negative HSF1 mutant (HSF379 or HSF1-K80Q), confirming that the increase is HSF1 independent. The antioxidant N-acetylcysteine completely inhibited the increase in HSPA1A and DNAJB1 mRNA levels upon methionine starvation, indicating that this increase is a response to oxidative stress resulting from a lack of methionine. Cells starved for methionine contained higher levels of c-Fos and FosB mRNA, but knockdown of these transcription factors had no effect on the HSPA1A or DNAJB1 mRNA level. Knockdown of NRF2 mRNA resulted in the inhibition of the increase in the HSPA1A mRNA, but not the DNAJB1 mRNA, level in methionine starved cells. We conclude that methionine deprivation results in both the amino acid deprivation response and an antioxidant response mediated at least in part by NRF2. This antioxidant response includes an HSF1 independent increase in the levels of HSPA1A and DNAJB1 mRNA.

A delayed antioxidant response in heat-stressed cells expressing a non-DNA binding HSF1 mutant.

To assess the consequences of inactivation of heat shock factor 1 (HSF1) during aging, we analyzed the effect of HSF1 K80Q, a mutant unable to bind DNA, and of dnHSF1, a mutant lacking the activation domain, on the transcriptome of cells 6 and 24 h after heat shock. The primary response to heat shock (6 h recovery), of which 30 % was HSF1-dependent, had decayed 24 h after heat shock in control cells but was extended in HSF1 K80Q and dnHSF1 cells. Comparison with literature data showed that even the HSF1 dependent primary stress response is largely cell specific. HSF1 K80Q, but not HSF1 siRNA-treated, cells showed a delayed stress response: an increase in transcript levels of HSF1 target genes 24 h after heat stress. Knockdown of NRF2, but not of ATF4, c-Fos or FosB, inhibited this delayed stress response. EEF1D_L siRNA inhibited both the delayed and the extended primary stress responses, but had off target effects. In control cells an antioxidant response (ARE binding, HMOX1 mRNA levels) was detected 6 h after heat shock; in HSF1 K80Q cells this response was delayed to 24 h and the ARE complex had a different mobility. Inactivation of HSF1 thus affects the timing and nature of the antioxidant response and NRF2 can activate at least some HSF1 target genes in the absence of HSF1 activity.

Urinary plasmin inhibits TRPV5 in nephrotic-range proteinuria.

Urinary proteins that leak through the abnormal glomerulus in nephrotic syndrome may affect tubular transport by interacting with membrane transporters on the luminal side of tubular epithelial cells. Patients with nephrotic syndrome can develop nephrocalcinosis, which animal models suggest may develop from impaired transcellular Ca(2+) reabsorption via TRPV5 in the distal convoluted tubule (DCT). In nephrotic-range proteinuria, filtered plasminogen reaches the luminal side of DCT, where it is cleaved into active plasmin by urokinase. In this study, we found that plasmin purified from the urine of patients with nephrotic-range proteinuria inhibits Ca(2+) uptake in TRPV5-expressing human embryonic kidney 293 cells through the activation of protease-activated receptor-1 (PAR-1). Preincubation with a plasmin inhibitor, a PAR-1 antagonist, or a protein kinase C (PKC) inhibitor abolished the effect of plasmin on TRPV5. In addition, ablation of the PKC phosphorylation site S144 rendered TRPV5 resistant to the action of plasmin. Patch-clamp experiments showed that a decreased TRPV5 pore size and a reduced open probability accompany the plasmin-mediated reduction in Ca(2+) uptake. Furthermore, high-resolution nuclear magnetic resonance spectroscopy demonstrated specific interactions between calmodulin and residues 133-154 of the N-terminus of TRPV5 for both wild-type and phosphorylated (S144pS) peptides. In summary, PAR-1 activation by plasmin induces PKC-mediated phosphorylation of TRPV5, thereby altering calmodulin-TRPV5 binding, resulting in decreased channel activity. These results indicate that urinary plasmin could contribute to the downstream effects of proteinuria on the tubulointerstitium by negatively modulating TRPV5.

Heat shock factor 1 is inactivated by amino acid deprivation.

Mammalian cells respond to a lack of amino acids by activating a transcriptional program with the transcription factor ATF4 as one of the main actors. When cells are faced with cytoplasmic proteotoxic stress, a quite different transcriptional response is mounted, the heat shock response, which is mediated by HSF1. Here, we show that amino acid deprivation results in the inactivation of HSF1. In amino acid deprived cells, active HSF1 loses its DNA binding activity as demonstrated by EMSA and ChIP. A sharp decrease in the transcript level of HSF1 target genes such as HSPA1A (Hsp70), DNAJB1 (Hsp40), and HSP90AA1 is also seen. HSPA1A mRNA, but not DNAJB1 mRNA, was also destabilized. In cells cultured with limiting leucine, HSF1 activity also declined. Lack of amino acids thus could lead to a lower chaperoning capacity and cellular frailty. We show that the nutrient sensing response unit of the ASNS gene contains an HSF1 binding site, but we could not detect binding of HSF1 to this site in vivo. Expression of either an HSF1 mutant lacking the activation domain (HSF379) or an HSF1 mutant unable to bind DNA (K80Q) had only a minor effect on the transcript levels of amino acid deprivation responsive genes.

Protein refolding in peroxisomes is dependent upon an HSF1-regulated function.

Post-heat shock refolding of luciferase requires chaperones. Expression of a dominant negative HSF1 mutant (dnHSF1), which among other effects depletes cells of HSF1-regulated chaperones, blocked post-heat shock refolding of luciferase targeted to the cytoplasm, nucleus, or peroxisomes, while refolding of endoplasmic reticulum (ER)-targeted luciferase was inhibited by about 50 %. Luciferase refolding in the cytoplasm could be partially restored by expression of HSPA1A and fully by both HSPA1A and DNAJB1. For full refolding of ER luciferase, HSPA1A expression sufficed. Neither nuclear nor peroxisomal refolding was rescued by HSPA1A. A stimulatory effect of DNAJB1 on post-heat shock peroxisomal luciferase refolding was seen in control cells, while refolding in the cytoplasm or nucleus in control cells was inhibited by DNAJB1 expression in the absence of added HSPA1A. HSPB1 also improved refolding of peroxisomal luciferase in control cells, but not in dnHSF1 expressing cells. HSP90, HSPA5, HSPA6, and phosphomevalonate kinase (of which the synthesis is also downregulated by dnHSF1) had no effect on peroxisomal refolding in either control or chaperone-depleted cells. The chaperone requirement for post-heat shock refolding of peroxisomal luciferase in control cells is thus unusual in that it can be augmented by DNAJB1 or HSPB1 but not by HSPA1A; in dnHSF1 expressing cells, expression of none of the (co)-chaperones tested was effective, and an as yet to be identified, HSF1-regulated function is required.

An atypical unfolded protein response in heat shocked cells.

BACKGROUND: The heat shock response (HSR) and the unfolded protein response (UPR) are both activated by proteotoxic stress, although in different compartments, and share cellular resources. How these resources are allocated when both responses are active is not known. Insight in possible crosstalk will help understanding the consequences of failure of these systems in (age-related) disease. RESULTS: In heat stressed HEK293 cells synthesis of the canonical UPR transcription factors XBP1s and ATF4 was detected as well as HSF1 independent activation of the promoters of the ER resident chaperones HSPA5 (BiP) and DNAJB9 (ERdj4). However, the heat stress activation of the DNAJB9 promoter, a XBP1s target, was not blocked in cells expressing a dominant negative IRE1α mutant, and thus did not require XBP1s. Furthermore, the DNA element required for heat stress activation of the DNAJB9 promoter is distinct from the ATF4 and ATF6 target elements; even though inhibition of eIF2α phosphorylation resulted in a decreased activation of the DNAJB9 promoter upon heat stress, suggesting a role for an eIF2α phosphorylation dependent product. CONCLUSIONS: The initial step in the UPR, synthesis of transcription factors, is activated by heat stress but the second step, transcriptional transactivation by these factors, is blocked and these pathways of the UPR are thus not productive. Expression of canonical ER chaperones is part of the response of heat stressed cells but another set of transcription factors has been recruited to regulate expression of these ER chaperones.

Explosive expansion of betagamma-crystallin genes in the ancestral vertebrate.

In jawed vertebrates, betagamma-crystallins are restricted to the eye lens and thus excellent markers of lens evolution. These betagamma-crystallins are four Greek key motifs/two domain proteins, whereas the urochordate betagamma-crystallin has a single domain. To trace the origin of the vertebrate betagamma-crystallin genes, we searched for homologues in the genomes of a jawless vertebrate (lamprey) and of a cephalochordate (lancelet). The lamprey genome contains orthologs of the gnathostome betaB1-, betaA2- and gammaN-crystallin genes and a single domain gammaN-crystallin-like gene. It contains at least two gamma-crystallin genes, but lacks the gnathostome gammaS-crystallin gene. The genome also encodes a non-lenticular protein containing betagamma-crystallin motifs, AIM1, also found in gnathostomes but not detectable in the uro- or cephalochordate genome. The four cephalochordate betagamma-crystallin genes found encode two-domain proteins. Unlike the vertebrate betagamma-crystallins but like the urochordate betagamma-crystallin, three of the predicted proteins contain calcium-binding sites. In the cephalochordate betagamma-crystallin genes, the introns are located within motif-encoding region, while in the urochordate and in the vertebrate betagamma-crystallin genes the introns are between motif- and/or domain encoding regions. Coincident with the evolution of the vertebrate lens an ancestral urochordate type betagamma-crystallin gene rapidly expanded and diverged in the ancestral vertebrate before the cyclostomes/gnathostomes split. The beta- and gammaN-crystallin genes were maintained in subsequent evolution, and, given the selection pressure imposed by accurate vision, must be essential for lens function. The gamma-crystallin genes show lineage specific expansion and contraction, presumably in adaptation to the demands on vision resulting from (changes in) lifestyle.

Co-chaperones are limiting in a depleted chaperone network.

To probe the limiting nodes in the chaperoning network which maintains cellular proteostasis, we expressed a dominant negative mutant of heat shock factor 1 (dnHSF1), the regulator of the cytoplasmic proteotoxic stress response. Microarray analysis of non-stressed dnHSF1 cells showed a two- or more fold decrease in the transcript level of 10 genes, amongst which are the (co-)chaperone genes HSP90AA1, HSPA6, DNAJB1 and HSPB1. Glucocorticoid signaling, which requires the Hsp70 and the Hsp90 folding machines, was severely impaired by dnHSF1, but fully rescued by expression of DNAJA1 or DNAJB1, and partially by ST13. Expression of DNAJB6, DNAJB8, HSPA1A, HSPB1, HSPB8, or STIP1 had no effect while HSP90AA1 even inhibited. PTGES3 (p23) inhibited only in control cells. Our results suggest that the DNAJ co-chaperones in particular become limiting in a depleted chaperoning network. Our results also suggest a difference between the transcriptomes of cells lacking HSF1 and cells expressing dnHSF1.

Manipulating heat shock factor-1 in Xenopus tadpoles: neuronal tissues are refractory to exogenous expression.

BACKGROUND: The aging related decline of heat shock factor-1 (HSF1) signaling may be causally related to protein aggregation diseases. To model such disease, we tried to cripple HSF1 signaling in the Xenopus tadpole. RESULTS: Over-expression of heat shock factor binding protein-1 did not inhibit the heat shock response in Xenopus. RNAi against HSF1 mRNA inhibited the heat shock response by 70% in Xenopus A6 cells, but failed in transgenic tadpoles. Expression of XHSF380, a dominant-negative HSF1 mutant, was embryonic lethal, which could be circumvented by delaying expression via a tetracycline inducible promoter. HSF1 signaling is thus essential for embryonic Xenopus development. Surprisingly, transgenic expression of the XHSF380 or of full length HSF1, whether driven by a ubiquitous or a neural specific promoter, was not detectable in the larval brain. CONCLUSIONS: Our finding that the majority of neurons, which have little endogenous HSF1, refused to accept transgene-driven expression of HSF1 or its mutant suggests that HSF1 levels are strictly controlled in neuronal tissue.

mRNA made during heat shock enters the first round of translation.

To determine whether mRNA synthesized during a heat shock is translated at least once in spite of the strong inhibition of translation by heat shock, we used nonsense-mediated decay (NMD) as an assay since NMD requires a round of translation. As NMD substrate we used the human psigammaE-crystallin gene, which contains a premature termination codon, and as control, its close relative, the human gammaD-crystallin gene, both placed under control of the Hsp70 promoter. We show that no spliced psigammaE-crystallin mRNA can be detected in heat shocked cells, suggesting that NMD resumes as soon as splicing is restored. We further show that newly synthesized mRNAs co-sediment with the 40S ribosomal subunits, indicating that the transcripts are recruited to the translation machinery but are stalled at the translation initiation stage. Using fluorescence loss in photobleaching (FLIP) we show that cytoplasmic EGFP-CBP20 is immobile in heat shocked cells. CBP20 is part of the cap binding complex which is thought to direct the first round of translation. Together our data suggest that all mRNAs made during heat shock enter the pioneer round of translation.

In vivo heteromer formation. Expression of soluble betaA4-crystallin requires coexpression of a heteromeric partner.

The beta-crystallins are a family of long-lived, abundant structural proteins that are coexpressed in the vertebrate lens. As beta-crystallins form heteromers, a process that involves transient exposure of hydrophobic interfaces, we have examined whether in vivobeta-crystallin assembly is enhanced by protein chaperones, either small heat shock proteins, Hsp27 or alphaB-crystallin, or Hsp70. We show here that betaA4-crystallin is abundantly expressed in HeLa cells, but rapidly degraded, irrespective of the presence of Hsp27, alphaB-crystallin or Hsp70. Degradation is even enhanced by Hsp70. Coexpression of betaA4-crystallin with betaB2-crystallin yielded abundant soluble betaA4-betaB2-crystallin heteromers; betaB1-crystallin was much less effective in solubilizing betaA4-crystallin. As betaB2-crystallin competed for betaA4-crystallin with Hsp70 and the proteasomal degradation pathway, betaB2-crystallin probably captures an unstable betaA4-crystallin intermediate. We suggest that the proper folding of betaA4-crystallin is not mediated by general chaperones but requires a heteromeric partner, which then also acts as a dedicated chaperone towards betaA4-crystallin.

Sequence and functional conservation of the intergenic region between the head-to-head genes encoding the small heat shock proteins alphaB-crystallin and HspB2 in the mammalian lineage.

An unexpected feature of the large mammalian genome is the frequent occurrence of closely linked head-to-head gene pairs. Close apposition of such gene pairs has been suggested to be due to sharing of regulatory elements. We show here that the head-to-head gene pair encoding two small heat shock proteins, alphaB-crystallin and HspB2, is closely linked in all major mammalian clades, suggesting that this close linkage is of selective advantage. Yet alphaB-crystallin is abundantly expressed in lens and muscle and in response to a heat shock, while HspB2 is abundant only in muscle and not upregulated by a heat shock. The intergenic distance between the genes for these two proteins in mammals ranges from 645 bp (platypus) to 1069 bp (opossum), with an average of about 900 bp; in chicken the distance was the same as in duck (1.6 kb). Phylogenetic footprinting and sequence alignment identified a number of conserved sequence elements close to the HspB2 promoter and two farther upstream. All known regulatory elements of the mouse alphaB-crystallin promoter are conserved, except in platypus and birds. The lens-specific region 1 (LSR1) and the heat shock elements (HSEs) lack in birds; in platypus the LSR1 is reduced to a Pax-6 site, while the Pax-6 site in LSR2 and a HSE are absent. Most likely the primordial mammalian alphaB-crystallin promoter had two LSRs and two HSEs. In transfection experiments the platypus alphaB-crystallin promoter retained heat shock responsiveness and lens expression. It also directed lens expression in Xenopus laevis transgenes, as did the HspB2 promoter of rat or blind mole rat. Deletion of the middle of the intergenic region including the upstream enhancer affected the activity of both the rat alphaB-crystallin and the HspB2 promoters, suggesting sharing of the enhancer region by the two promoters.

Influence of hormones and growth factors on lens protein composition: the effect of dexamethasone and PDGF-AA.

PURPOSE: To investigate the effect of hormones and ocular growth factors on the expression of alpha-, beta-, and gamma-crystallins in rat lens epithelial and fiber cells. METHODS: PDGF-AA, EGF, NGF, M-CSF, BMP-2, BMP-4, dexamethasone, and estrogen were tested for their ability to alter the spectrum of crystallins in explanted newborn rat lens epithelial cells or in vitro differentiating newborn rat lens fiber cells. The accumulation of alphaA-, aB-, betaA3/1-, betaB2-, and gamma-crystallin was measured by western blot and dot blot analysis. The morphology of the rat lens explants after culture was examined by hematoxylin-eosin staining, while crystallins were localized by immunofluorescence. RESULTS: Only dexamethasone and PDGF-AA showed an effect on relative crystallin levels. In the presence of dexamethasone the amount of alphaB-crystallin was increased in lens epithelial cells, but dexamethasone did not affect the crystallin spectrum in fiber cells. In rat lens epithelial explants cultured with PDGF-AA an increase in beta- and gamma-crystallin expression was seen. The spectrum of beta- and gamma-crystallins synthesized differed from that present in lens fiber cells. The cells expressing beta- and gamma-crystallin after culture with PDGF-AA were scattered in the epithelial cell layer and retained an epithelial morphology. PDGF-AA did not change the spectrum of crystallins synthesized in lens fiber cells but did enhance the rate of fiber cell differentiation, in agreement with results of others. CONCLUSIONS: Both dexamethasone and PDGF-AA influence crystallin gene expression in cultured rat lens epithelial cells. Dexamethasone enhances the expression of alphaB-crystallin while culturing in the presence of PDGF-AA caused an increase in beta- as well as gamma-crystallin synthesis. Since at least the gamma-crystallin genes are known to be silenced in epithelial cells by DNA methylation, PDGF-AA may be able to induce one of the steps towards fiber cell differentiation in some epithelial cells.

Translational thermotolerance provided by small heat shock proteins is limited to cap-dependent initiation and inhibited by 2-aminopurine.

Heat shock results in inhibition of general protein synthesis. In thermotolerant cells, protein synthesis is still rapidly inhibited by heat stress, but protein synthesis recovers faster than in naive heat-shocked cells, a phenomenon known as translational thermotolerance. Here we investigate the effect of overexpressing a single heat shock protein on cap-dependent and cap-independent initiation of translation during recovery from a heat shock. When overexpressing alphaB-crystallin or Hsp27, cap-dependent initiation of translation was protected but no effect was seen on cap-independent initiation of translation. When Hsp70 was overexpressed however, both cap-dependent and -independent translation were protected. This finding indicates a difference in the mechanism of protection mediated by small or large heat shock proteins. Phosphorylation of alphaB-crystallin and Hsp27 is known to significantly decrease their chaperone activity; therefore, we tested phosphorylation mutants of these proteins in this system. AlphaB-crystallin needs to be in its non-phosphorylated state to give protection, whereas phosphorylated Hsp27 is more potent in protection than the unphosphorylatable form. This indicates that chaperone activity is not a prerequisite for protection of translation by small heat shock proteins after heat shock. Furthermore, we show that in the presence of 2-aminopurine, an inhibitor of kinases, among which is double-stranded RNA-activated kinase, the protective effect of overexpressing alphaB-crystallin is abolished. The synthesis of the endogenous Hsps induced by the heat shock to test for thermotolerance is also blocked by 2-aminopurine. Most likely the protective effect of alphaB-crystallin requires synthesis of the endogenous heat shock proteins. Translational thermotolerance would then be a co-operative effect of different heat shock proteins.

c-Maf, the gammaD-crystallin Maf-responsive element and growth factor regulation.

The transcription factor c-Maf has been suggested to regulate the activity of gamma-crystallin promoters in lens fibre cells. We here show that the transactivation potential of c-Maf and MafB for the rat gammaD-crystallin Maf-responsive element (gammaD MARE) is dependent upon the cellular context and, using chimeric and single domain mutants, that c-Maf is most likely to be the cognate factor for the gammaD MARE in the lens. Transactivation of the gammaD MARE by c-Maf in lens cells was not enhanced by c-Fos or c-Jun and was not blocked by dominant negative c-Fos or c-Jun constructs. c-Maf can activate the gammaD MARE as a homodimer since activation of the gammaD-crystallin promoter in P19 embryonic carcinoma cells required only c-Maf, but none of a number of c-Fos and c-Jun family members tested. Transactivation by c-Maf was inhibited by activation of protein kinase A (PKA) (by signal transduction agonist forskolin) or of protein kinase C (PKC) (by signal transduction agonist tetradecanoyl phorbol acetate). Site-directed mutagenesis showed that this effect is not mediated by phosphorylation of the consensus PKA/PKC site in the extended DNA-binding domain, but likely involves activation of MAP kinase kinase, as inhibition by PD98059 increased transactivation by c-Maf.