Sequential Modifications in Class II Transactivator Isoform 1 Induced by Lipopolysaccharide Stimulate Major Histocompatibility Complex Class II Transcription in Macrophages*

By presenting antigenic peptides on major histocompatibility complex class (MHC) II determinants to CD4+ T cells, macrophages help to direct the establishment of adaptive immunity. We found that in these cells, lipopolysaccharide stimulates the expression of MHC II genes via the activation of Erk1/2, which is mediated by Toll-like receptor 4. Erk1/2 then phosphorylates the serine at position 357, which is located in a degron of CIITA isoform 1 that leads to its monoubiquitylation. Thus modified, CIITA isoform 1 binds P-TEFb, which mediates the elongation of RNA polymerase II and co-transcriptional processing of nascent transcripts. This induction leads to the expression of MHC II genes. Subsequent polyubiquitylation results in the degradation of CIITA isoform 1. Thus, the signaling cascade from Toll-like receptor 4 to CIITA isoform 1 represents one connection between innate and adaptive immunity in macrophages.

The immune system is composed of innate immunity, which performs the function of immune surveillance, and adaptive immunity, which eliminates non-self-antigens and creates the immune memory. Constituents of the former are antigenpresenting cells and of the latter are B and T cells. Because the establishment of adaptive immunity is dependent on innate immunity, appropriate interactions between them are indispensable for the normal function of the immune system.
Macrophages are antigen-presenting cells that remove and digest invading pathogens as well as present antigenic peptides, thus directing the immune response against them. They sense the presence of these pathogens via their pathogen-associated molecular patterns, such as lipopolysaccharide (LPS), 2 flagellin, etc., which is mediated by Toll-like receptors (TLRs) (1). For example, the binding between LPS and TLR4 triggers a signaling cascade that results in the activation of p38 mitogen-activated protein kinase (p38 MAPK), extracellular regulated kinase 1/2 (Erk1/2), and Jun kinases (JNK), as well as of nuclear factor B (NF-B) (2). This signaling leads to the presentation of antigenic peptides in the groove of major histocompatibility complex class II (MHC II) molecules to CD4 ϩ T cells and subsequently to the establishment of adaptive immunity (1). Because MHC II determinants are involved directly in the establishment of adaptive immunity, it is not surprising that their expression is tightly regulated. One of the levels of this regulation is at transcription and is mediated by the class II transactivator (CIITA).
Transcription of CIITA can be initiated from three distinct promoters called PI, PIII, and PIV, which direct the synthesis of three isoforms (IF) of CIITA: IF1, IF3 and IF4 (3). Different isoforms are expressed following distinct stimuli in different cells. Whereas CIITA IF1 is expressed in macrophages and myeloid dendritic cells, CIITA IF3 is expressed in B cells and plasmacytoid dendritic cells (4). Only one study has been performed on CIITA IF1 (5). In contrast, CIITA IF3 has been studied in great detail (6 -12). CIITA isoforms do not bind to DNA. Rather they bind to the preformed enhanceosome on MHC II promoters, where they recruit general transcription factors (13)(14)(15)(16)(17) as well as coactivators (11, 18 -21) and function as transcriptional integrators.
Transcription of eukaryotic genes is a highly coordinated process and is often regulated by post-translational modifications of activators. The most studied are post-translational modifications of proteins containing class II B or acidic activation domains (AADs). The phosphorylation of proteolytic signaling elements, called degrons, which are located in or near these AADs, leads to their monoubiquitylation and higher transcriptional activity. Subsequent polyubiquitylation finally results in their degradation (22). Indeed, by binding the positive transcription elongation factor b (P-TEFb), which is composed of cyclin T1 (CycT1) and cyclin-dependent kinase 9 (Cdk9), the monoubiquitylated VP16 protein increases the rates of elongation rather than initiation of transcription (23). Of note, cyclindependent kinase 9 phosphorylates the C-terminal domain of RNA polymerase II and the negative transcription elongation factor, which contains 5,6-dichloro-1-␤-D-ribofuranosylbenzimidazole-sensitivity inducing factor and negative elongation factor (24). These changes lead additionally to co-transcriptional processing of nascent mRNA species.
LPS increases the levels of MHC II molecules on macrophages via an unknown mechanism (25,26). In this study, we found that TLR4, which binds LPS, directs several post-translational modifications of CIITA IF1. These modifications include its phosphorylation and monoubiquitylation. They lead to the binding between CIITA IF1 and P-TEFb, which increases the transcriptional activity of CIITA IF1 and thus the expression of MHC II genes. Subsequent polyubiquitylation results in a rapid degradation of CIITA IF1. Thus, these findings connect the innate and adaptive arms of the immune response.

EXPERIMENTAL PROCEDURES
Animals, Cells, and Cell Culture-C57BL/10ScN mice were purchased from the Jackson Laboratory (Bar Harbor, ME) and bred in our colony. HeLa, COS, and RAW 264.7 cells were maintained in Dulbecco's modified Eagle's medium, 5% fetal calf serum, and antibiotics. Bone marrow-derived macrophages were prepared and maintained as described previously (27). LPS Re595 was obtained from Sigma. MEK1/2 inhibitor UO 126 was obtained from Promega (Madison, WI), and proteasome inhibitor ALLN was purchased from Calbiochem (La Jolla, CA). -Phosphatase was obtained from Sigma.
Plasmid DNAs-Reporter plasmid pDRASCAT was described previously (12). Plasmid m:Ub was described previously (23). To construct a plasmid m:CycT1, the coding sequence of CycT1 was cloned into the expression vector pEF-Myc between EcoRI and XbaI restriction sites. Plasmids f:CIITA1 and h:CIITA1 were gift from Dr. Ting (University of North Carolina) and Dr. Chang (Indiana University School of Medicine). To construct a plasmid coding for f:CIITA1(S357A), the f:CIITA1 plasmid was subjected to sitedirected mutagenesis with the QuikChange II XL site-directed mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions. To construct plasmids coding for the h:Ub.CIITA1 and h:Ub(K48,63R).CIITA1, the Ub and Ub(K48,63R) plasmids were digested with EcoRI and cloned into h:CIITA1. Murine TLR4 was cloned into the pMX-pie bicistronic retroviral vector, as described previously (Onishi 1996). All of the plasmids were verified by DNA sequencing.
Transient Transfection and CAT Assay-The cells were seeded into 100-mm-diameter Petri dishes ϳ12 h prior to transfection. The cells were transfected with FuGene 6 reagent according to the manufacturer's instructions (Invitrogen). CAT enzymatic assays were performed as described (29). Fold transactivation represents the ratio between the CIITA-activated transcription and the activity of the reporter plasmid alone.
Immunoprecipitation Assay and Western Blot Analysis-The cells were transfected with 2 g of indicated plasmid vectors. About 18 h after transfection, the immunoprecipitations were performed as described previously (12). Precipitated proteins were resolved on SDS-PAGE and analyzed by immunoblotting with the appropriate antibody followed by horseradish peroxidase-conjugated secondary antibody. The blots were developed by chemiluminiscence assay from PerkinElmer Life Sciences.
Chromatin Immunoprecipitation (ChIP) Assays, RT-PCR, and Quantitative Real-time PCR-ChIP assays were performed as described previously (30). RNA was extracted from RAW 264.7 cells using the TRIzol protocol from Invitrogen. RT-PCR was described previously (31). cDNAs were then amplified by primers described previously (32). Quantitative PCR was performed by Stratagene Mx3005P quantitative real-time PCR system, according to the manufacturer's protocol.
In Vitro Transcription and Translation and in Vitro Kinase Assay-f:CIITA1 and f:CIITA1(S357A) were expressed in vitro by using coupled rabbit reticulocyte lysate transcription and translation system from Promega (Madison, WI). In vitro kinase assay with Erk1/2 was performed according to the manufacturer's instructions (Upstate Biotechnology Inc., Lake Placid, NY).
Pulse-Chase Analysis-18 h after transfection, COS cells were starved for 2 h in medium without cysteine and methionine. Radioactive labeling with [S 35 ]cysteine and [S 35 ]methionine was performed for 40 min. After the labeling, the cells were washed three times in phosphate-buffered saline and maintained in Dulbecco's modified Eagle's medium for indicated time periods. The cells were lysed for 45 min at 4°C. The cell lysates were then subjected to immunoprecipitation. Precipitated proteins were resolved on SDS-PAGE and analyzed by radiography.

LPS Increases the Expression of MHC II Determinants via TLR4 in Mouse
Macrophages-Upon stimulation with LPS, macrophages increase the surface levels of MHC II molecules (1). Via TLR4, LPS also activates Erk1/2, JNK, and p38 MAPKs (2). Thus, we asked two questions. First, does the expression of MHC II determinants depend on TLR4? Second, does this induction depend on any of these kinases?
To answer the first question, we infected primary bone marrow-derived macrophages from TLR4 Ϫ/Ϫ mice with TLR4 ( Fig. 1A, lanes 3 and 4) or an empty vector as the control (Fig.  1A, lanes 1 and 2). Two days later, we stimulated these macrophages with LPS for 2 h ( To answer the second question, we duplicated these experiments in RAW 264.7 cells, which are a mouse macrophage cell line. At identical time points, we determined levels of MHC II determinants (Fig. 1B, top panel, lanes 1 and 2). Indeed, LPS induced the expression of MHC II genes in RAW 264.7 cells equivalently to primary macrophages (Fig. 1B, top panel, compare lanes 1 and 2). In contrast, when these cells were treated with the MEK inhibitor UO 126 prior to the addition of LPS, the expression of MHC II determinants was abolished (Fig. 1B, top panel, compare lanes 2 and 3). These experiments were repeated with inhibitors of JNK and p38 MAPKs, but they had no effect on the induction of MHC II determinants by LPS (data not presented). UO 126 is a specific MEK1/2 inhibitor that prevents the phosphorylation and subsequent activation of Erk1/2. These findings were confirmed by Western blotting with antiphospho-Erk1/2 antibodies (20 min; Fig. 1B, bottom panel, compare lanes 2 and 3). Moreover, as presented in supplemental Fig. S1 (top panel, lanes [1][2][3][4][5][6], the activation of Erk1/2 was detectable by 10 min and disappeared by 2 h upon the addition of LPS in primary cells. We conclude that the activation of Erk1/2 precedes and is required for the induction of MHC II determinants by LPS. To determine effects of LPS on MHC II transcription, we stimulated RAW 264.7 cells with LPS and measured levels of I-A␣ transcripts starting at the zero time point. After 2 h, the cells were washed, and the medium was changed. We performed RT-PCR at the indicated time points (Fig. 1C, top panel). Initially, we did not detect MHC II transcripts in unstimulated cells (Fig. 1C, top panel, lane 1). They appeared 1 h and peaked 2 h after the addition of LPS (Fig. 1C, top panel, compare lanes 3 and 4). 4 h later, MHC II transcripts disappeared almost completely (Fig. 1C, top panel, lane 5). Importantly, the levels of actin mRNA, which we used as the internal control, were equivalent in all samples (Fig. 1C, bottom panel). Thus, LPS activates transiently Erk1/2, which is followed by the induction of MHC II transcription.
Next, we wanted to determine whether the expression of MHC II genes depended on the de novo synthesis of CIITA. Because different isoforms of CIITA had been described in RAW 264.7 cells (33, 34), we investigated first CIITA IF1, IF3, and IF4 transcripts by RT-PCR in the presence and absence of LPS in these cells. However, as presented in supplemental Fig.  S2, only CIITA IF1 mRNA encodes a functional CIITA protein in RAW 264.7 cells. Moreover, CIITA IF1 was the only isoform that we detected by immunoprecipitation and subsequent Western blotting in these cells (see Fig. 3D, bottom panel). Next, we performed quantitative real time RT-PCR to quantify mRNA levels of CIITA IF1 in stimulated and unstimulated cells (Fig. 1D). Time points 0, 2, and 4 h correspond to samples from Fig. 1C (lanes 1, 4, and 5). Surprisingly, the stimulation with LPS for 2 h decreased levels of CIITA IF1 mRNA up to 7-fold, compared with those in unstimulated cells. Thus, we conclude that LPS induces the transcription of MHC II genes without a concomitant increase in CIITA IF1 transcripts in macrophages. Moreover, this stimulation depends on TLR4 and the activation of Erk1/2 and leads to a transient induction of transcription of MHC II genes.
Erk1/2 Phosphorylates CIITA IF1 on the Serine at Position 357-Effects of LPS on levels of MHC class II determinants depend on the activation of Erk1/2. Moreover, CIITA IF1 directs the transcription of MHC II genes in macrophages. However, because the amounts of CIITA IF1 mRNA actually decreased (Fig. 1D), we hypothesized that Erk1/2 modifies the preformed CIITA IF1 for its increased transcriptional activity.
When analyzed by gel electrophoresis, CIITA IF3 migrates as a double band, which has been attributed to its phosphorylation (6,10,12). Moreover, treatment with -phosphatase and direct labeling revealed that the upper band represents the phosphorylated form of CIITA IF3 (12). Because it also appears as a double band, we wanted to determine whether CIITA IF1 is also modified by phosphorylation. Thus, we expressed the FLAG epitope-tagged CIITA IF1 protein (f:CIITA1) (Fig. 2A, lane 1) in COS cells. In these cells, Erk1/2 is active constitutively. Thus, they do not need the stimulation by LPS. An aliquot of cell lysates was incubated with -phosphatase, which removes phosphates from serines and threonines (Fig. 2A, lane  2). As expected, the upper f:CIITA1 band disappeared completely ( Fig. 2A, compare lanes 1 and 2). We conclude that CIITA IF1 is also phosphorylated in cells.
Because in Fig. 1 we showed that the increased expression of MHC II determinants after LPS stimulation depended on the activation of Erk1/2, we searched for putative Erk1/2 phosphorylation sites in CIITA IF1. Erk1/2 is a proline-directed kinase with the consensus phosphorylation sequence PX(S/T)P, where the serine or threonine is phosphorylated. Indeed, there is only one consensus Erk1/2 phosphorylation site in CIITA IF1, which contains a serine at position 357 (S357) (Fig. 2B).
To determine whether this site is phosphorylated, we created the mutant f:CIITA1 protein, where the serine at position 357 was changed to alanine (f:CIITA1(S357A)). Next, we expressed both the wild type f:CIITA1 and the mutant f:CIITA1(S357A) proteins in COS cells (Fig. 2C). We made two observations. First, expression levels of the mutant f:CIITA1(S357A) protein were 4-fold higher than those of the f:CIITA1 protein (Fig. 2C,  compare lanes 1 and 2). Second, the mutant f:CIITA1(S357A) protein migrated as a single lower band (Fig. 2C, compare lanes  1 and 2), which represents the nonphosphorylated form of CIITA IF1. Thus, we conclude that CIITA IF1 is phosphorylated on Ser 357 in cells.
Next, we wanted to determine whether Erk1/2 is involved in the phosphorylation of CIITA IF1 and whether the upper band of CIITA IF1 would diminish under the influence of UO 126.
Again we expressed f:CIITA1 (Fig. 2D, lanes 1 and 2) and treated COS cells with UO 126 (Fig. 2D, lane 2) or the solvent as the control (Fig. 2D, lane 1). As expected, in the presence of UO 126, the upper band of f:CIITA1 disappeared (Fig. 2D, top  panel, compare lanes 1 and 2), which correlated with the absence of phosphorylated Erk1/2 (Fig. 2D, bottom panel, compare lanes 1 and 2). This finding suggests that Erk1/2 is involved in the phosphorylation of CIITA IF1.
To extend our findings, we determined whether Erk1/2 could phosphorylate CIITA IF1 in vitro and whether this phosphorylation depended on Ser 357 . Thus, we performed in vitro kinase assays with Erk1/2. Wild type f:CIITA1 and mutant f:CIITA1(S357A) proteins were transcribed and translated using the rabbit reticulocyte lysate in vitro (Fig. 2E). Aliquots of the reaction were incubated with Erk1/2 (Fig. 2E, lanes 2 and 4) or with the reaction buffer as the control (Fig. 2E, lanes 1 and 3). Indeed, as indicated by the appearance of the upper band, Erk1/2 phosphorylated the wild type f:CIITA1 protein (Fig. 2E,  compare lanes 1 and 2). In contrast, Erk1/2 did not phosphorylate the mutant f:CIITA1(S357A) protein, which was confirmed by the absence of the upper band (Fig. 2E, compare lanes  3 and 4 with lane 2). Overall, we conclude that Erk1/2 phosphorylates Ser 357 in CIITA IF1 in vitro and in vivo.
CIITA IF1 Is Ubiquitylated in Cells-Thus far, we determined not only that Erk1/2 phosphorylates Ser 357 in f:CIITA1, but we also observed that this phosphorylation affects the stability of f:CIITA1. Namely, expression levels of the mutant f:CIITA1(S357A) protein were higher than those of f:CIITA1 (Fig. 2C, compare lanes 1 and 2). These findings led us to hypothesize that the phosphorylation and degradation of CIITA IF1 could be connected via ubiquitin.
To address this notion, we determined first whether different expression levels of f:CIITA1 and mutant f:CIITA1(S357A) proteins were due to differences in their rates of degradation. We expressed f:CIITA1 (Fig. 3A, top panel) and the mutant f:CIITA1(S357A) proteins (Fig. 3A, bottom panel) and performed pulse-chase analyses in COS cells. Indeed, the mutant f:CIITA1(S357A) protein had a half-life of ϳ2 h, which was four times longer than that of f:CIITA1 (Fig. 3A, compare top and bottom panels). We conclude that the phosphorylation of CIITA IF1 leads to its degradation.
Next, we wanted to determine whether CIITA IF1 was degraded via the proteasome. We expressed f:CIITA1 (Fig. 3B,  lanes 1 and 2) and the mutant f:CIITA1(S357A) (Fig. 3B, lanes 3  and 4) proteins in COS cells. Transfected cells were treated with the proteasomal inhibitor ALLN for 6 h (Fig. 3B, lanes 2 and 4) and the solvent as the control (Fig. 3B, lanes 1 and 3). ALLN  1 and 2) and the mutant f:CIITA1(S357A) proteins (lanes 3 and 4) were transcribed and translated using the rabbit reticulocyte lysate in vitro. Aliquots of these reactions were then incubated with Erk1/2 (lanes 2 and 4) or the reaction buffer as the control (lanes 1 and 3). The arrows on the right indicate the presence of phosphorylated and nonphosphorylated f:CIITA1 proteins. WB, Western blot.
increased levels of only the phosphorylated form of f:CIITA1 ϳ4-fold (Fig. 3B, lanes 1 and 2, compare top and bottom bands). It had almost no effect on levels of the mutant f:CIITA1(S357A) protein (Fig. 3B, compare lanes 3 and 4). Thus, we conclude that the phosphorylated CIITA IF1 is also degraded via the proteasome.
To determine whether CIITA IF1 is ubiquitylated, we performed ubiquitylation assays in vivo. In COS cells, we co-expressed f:CIITA1 with the Myc epitope-tagged ubiquitin (m:Ub) (Fig. 3C, lane 2) or an empty plasmid vector as the control (Fig. 3C, lane 1). Indeed, in the presence of f:CIITA1 and m:Ub, there was a strong ubiquitylation ladder (Fig. 3C, lane 2).
Because the mutant f:CIITA1(S357A) protein was more stable than f:CIITA1, we assumed that the ubiquitylation of f:CIITA1 was dependent on the phosphorylation of Ser 357 . Thus, we performed the ubiquitylation assay in vivo with the mutant f:CIITA1(S357A) protein (Fig. 3C, lanes 3 and 4). Indeed, by densitometry we determined that the mutant f:CIITA1(S357A) protein was ubiquitylated 4-fold less than f:CIITA1 (Fig. 3C,  compare lanes 2 and 4). We conclude that CIITA IF1 is ubiqui-tylated, which is promoted by the phosphorylation of Ser 357 in CIITA IF1.
Finally, we asked whether the endogenous CIITA IF1 is also ubiquitylated in mouse macrophages. To address this question, we repeated ubiquitylation assays in RAW 264.7 cells (Fig. 3D). Because of low levels of expression or accelerated degradation, we were not able to detect CIITA IF1 or its ubiquitylation in untreated cells (Fig. 3D, lane 1) or cells treated with LPS and UO 126, respectively (data not presented). However, we were able to observe the heavily ubiquitylated CIITA IF1 only after treatment with LPS and ALLN (Fig. 3D, lane 2). We treated cells with ALLN for 2 h prior to 40 min of treatment with LPS. Afterward, the cells were incubated with ALLN for an additional 2 h. Interestingly, CIITA IF1 was present in phosphorylated and unphosphorylated forms (Fig. 3D, bottom panel, lane 2), which can be explained by transient activation of Erk1/2 by LPS (Fig. 1D). Thus, we conclude that the endogenous CIITA IF1 is phosphorylated and ubiquitylated in RAW 264.7 cells.
In CIITA IF1, there is a predicted PEST sequence from positions 360 to 385 (Fig. 3E). The phosphorylation of these sequences decreases the stability of proteins (22,35). Because the phosphorylation of Ser 357 , which is flanked by a proline, a serine, and an acidic residue, led to the degradation of CIITA IF1, we surmise that the actual PEST sequence in CIITA IF1 is extended on its N terminus. Thus, Ser 357 becomes a part of the functional degron, which is located from positions 355 to 385 in CIITA IF1 (Fig. 3E). Overall, we conclude that the phosphorylation of Ser 357 in the degron precedes the ubiquitylation of CIITA IF1 that subsequently leads to its degradation.
Ubiquitylated CIITA IF1 Protein Is More Active Than Its Counterpart That Is Not Ubiquitylated-Thus far, we connected two post-translational modifications in CIITA IF1, namely phosphorylation and ubiquitylation, with the stability of CIITA IF1. It is known that ubiquitylation strongly influences the function of activators (23,(35)(36)(37)(38). Of note, monoubiquitylation increases the activity of transcription factors, whereas polyubiquitylation leads to their degradation. For example, by recruiting P-TEFb better, the monoubiquitylation of VP16 increases its effects on transcriptional elongation (23). Moreover, CIITA IF1 binds CycT1 (17). Thus, we hypothesized that the monoubiquitylated CIITA IF1 protein also binds P-TEFb better and is therefore more active than its nonubiquitylated counterpart.
It had been demonstrated previously that transcription factors fused to ubiquitin behave as endogenously ubiquitylated proteins (23,36). Therefore, we constructed the hemagglutinin epitope-tagged fusion protein between CIITA IF1 and ubiquitin (h:Ub.CIITA1). We expressed f:CI-ITA1 (Fig. 4A, lane 1) and the h:Ub.CIITA1 fusion proteins (Fig. 4A, lane 2) in COS cells. Of interest, the steady state levels of h:Ub.CIITA1 chimera were 10-fold lower than those of f:CIITA1 (Fig. 4A, compare lanes 1 and 2). This suggested that the h:Ub.CIITA1 fusion protein was extremely unstable. Indeed, pulse-chase analyses revealed that the half-life of the h:Ub.CIITA1 chimera was less than 10 min (Fig. 4B). As expected, the presence of ALLN for 6 h increased levels of the h:Ub.CIITA1 chimera substantially (Fig. 4C, compare  lanes 1 and 2). This finding is in agreement with data presented in Fig. 3D, where the endogenous CIITA IF1 protein was detectable only after the addition of ALLN in RAW 264.7 cells. Thus, we conclude that h:Ub.CIITA1 chimera is rapidly degraded via the proteasome.
To compare activities of the nonubiquitylated and ubiquitylated CIITA IF1 proteins, we performed transcriptional assays in cells. The mutant f:CIITA1(S357A) protein represented the nonubiquitylated CIITA IF1 protein, and the h:Ub.CIITA1 fusion protein represented the ubiquitylated CIITA IF1 protein. They were co-expressed with pDRASCAT reporter plasmid in COS cells. Prior to performing CAT assays, we equalized the levels of both proteins (Fig. 4D). Results revealed that the ubiquitylated CIITA IF1 protein stimulated transcription ϳ6-fold higher than its nonubiquitylated counterpart. We conclude that the ubiquitylation of CIITA IF1 increases its transcriptional activity.
To extend our findings, we performed ChIP assays in a mouse macrophage cell line with endogenous CIITA IF1 and CycT1 proteins (Fig. 5B). As presented in Fig. 1B, I-A␣ mRNA was detectable 1 h after the stimulation with LPS (Fig. 1B, top  panel, lane 3). To determine whether P-TEFb is on MHC II promoters before the appearance of MHC II mRNA species, we performed ChIP prior to this time point. We stimulated RAW 264.7 cells with LPS alone (Fig. 5B, lane 2) or with UO 126 (Fig.  5B, lane 3) for 40 min, or left them untreated (Fig. 5B, lane 1). After cross-linking and sonication, we performed immunoprecipitations with anti-CIITA and anti-CycT1 antibodies. Immunoprecipitated DNA was amplified by PCR with primers complementary to the I-A␣ promoter. When immunoprecipitated with anti-CycT1 antibodies, PCR products were present only in cells, which were stimulated with LPS (Fig. 5B, top panel, lane  2). In contrast, when immunoprecipitated with anti-CIITA antibodies, PCR products were detected only in cells treated with UO 126 (Fig. 5B, middle panel, lane 3). Thus, the treatment with UO 126 prevented the activation and binding of CIITA IF1 to CycT1 but led to its stabilization and subsequent accumulation on MHC II promoters (Fig. 5B, middle panel,  lane 3). As presented in supplemental Fig. S3, only monoubiquitylated CIITA IF1 bound CycT1. Overall, these results suggest that the phosphorylated, ubiquitylated CIITA IF1 recruits