Expression of the three human major histocompatibility complex class II isotypes exhibits a differential dependence on the transcription factor RFXAP

Abstract

Major histocompatibility complex class II (MHCII) molecules play a pivotal role in the immune system because they direct the development and activation of CD4(+) T cells. There are three human MHCII isotypes, HLA-DR, HLA-DQ, and HLA-DP. Key transcription factors controlling MHCII genes have been identified by virtue of the fact that they are mutated in a hereditary immunodeficiency resulting from a lack of MHCII expression. RFXAP-one of the factors affected in this disease-is a subunit of RFX, a DNA-binding complex that recognizes the X box present in all MHCII promoters. To facilitate identification of conserved regions in RFXAP, we isolated the mouse gene. We then delimited conserved domains required to restore endogenous MHCII expression in cell lines lacking a functional RFXAP gene. Surprisingly, we found that 80% of RFXAP is dispensable for the reactivation of DR expression. Only a short C-terminal segment of the protein is essential for this isotype. In contrast, optimal expression of DQ and DP requires a larger C-terminal segment. These results define an RFXAP domain with an MHCII isotype-specific function. Expression of the three MHCII isotypes exhibits a differential requirement for this domain. We show that this is due to a differential dependence on this domain for promoter occupation and recruitment of the coactivator CIITA in vivo.

FIG. 1

Sequence homology between human and mouse RFXAP. (A) Alignment of the human (H) and mouse (M) RFXAP amino acid sequences. Stars indicate identical residues. Triangles indicate positions at which the protein is truncated in the ABI cell line and in the two alleles of the 6.1.6 cell line. Arrows indicate ends of the N-terminal (N1 to N9) and C-terminal (C1 to C4) deletions. Regions rich in acidic amino acids (DE), basic amino acids (RK), and glutamine (Q) are underlined. Amino acid coordinates are indicated at the right. A secondary-structure prediction is indicated on the bottom line: H represents α-helices, the line represents coils, and E represents β-strands. (B) Schematic representations of human and mouse RFXAPs are shown. The acidic-amino-acid (DE)-, basic-amino-acid (RK)-, and glutamine (Q)-rich regions are represented as filled boxes. The ends of N- and C-terminal deletions are positioned on the human protein. Potential translation initiation codons are indicated. Homology (percent identity) between mouse and human RFXAPs within different regions is indicated below. (C) Sequences of human and mouse RFXAPs in the vicinity of the translation initiation codon and the next conserved downstream ATG codon. The translation initiation codon of human RFXAP aligns with an ACG codon in mouse RFXAP. Stars indicate nucleotides providing a good Kozak context.

FIG. 2

Mouse RFXAP is functionally equivalent to the human protein. (A) Schematic maps of human RFXAP and the two mouse RFXAP constructs (M-ACG and M-ATG). The M-ACG construct starts at the nonclassical ACG initiation codon and contains a second ATG codon situated 81 nucleotides further downstream. The M-ATG construct starts at the second ATG codon. (B) In vitro translation products of M-ACG and M-ATG were analyzed by SDS-PAGE. NS corresponds to nonspecific bands labeled in the lysate. (C) Binding of the RFX complex in extracts derived from cells complemented with mouse RFXAP. Binding was analyzed by EMSA with a DRA X box probe. Whole-cell extracts from Raji cells, nontransfected 6.1.6 cells, and 6.1.6 cells transfected with M-ACG or M-ATG were analyzed. All of the lanes are from the same gel at the same length of exposure. (D) Mouse RFXAP restores MHCII expression in 6.1.6 cells. Cell surface expression of HLA-DR, HLA-DP, and HLA-DQ was analyzed by FACS in Raji cells (gray histograms), nontransfected 6.1.6 cells (open histograms), and 6.1.6 cells transfected with M-ACG or M-ATG (black histograms). Expression of MHCI molecules was included as a control.

FIG. 3

A short C-terminal region of RFXAP is sufficient for binding of the RFX complex and activation of HLA-DR expression in RFXAP-deficient 6.1.6 cells. (A) Cell surface HLA-DR expression was analyzed by FACS in nontransfected 6.1.6 cells (open histogram) and in 6.1.6 cells complemented with bicistronic lentiviral vectors encoding wild-type RFXAP (wt) or RFXAPs with the N4, N5, N6, N7, N8, and N9 deletions (black histograms). The various versions of RFXAP are indicated schematically and their sizes in amino acids are provided. (B) Binding of the RFX complex was analyzed by EMSA with nuclear extracts derived from nontransfected 6.1.6 cells, 6.1.6 cells complemented with wild-type RFXAP, and 6.1.6 cells complemented with the versions of RFXAP with the N7, N8, or N9 deletion. All of the lanes are from the same gel at the same length of exposure. (C) Binding of the RFX complex was analyzed by EMSA with nuclear extracts derived from nontransfected 6.1.6 cells (6.1.6), Mann B cells, and 6.1.6 cells complemented with wild-type RFXAP. All lanes come from the same gel at the same length of exposure. (D) The composition of RFX complexes detected in EMSA was analyzed by supershift experiments using antibodies directed against RFX5, RFXANK, and the HA tag carried by the transfected RFXAP subunit. All three antibodies supershift the RFX complexes formed in 6.1.6 cells complemented with the constructs with the N7, N8, and N9 deletions.

FIG. 4

The minimal essential C-terminal region of RFXAP is sufficient for activation of HLA-DR expression in RFXAP-deficient ABI fibroblasts. (A) Schematic representation of the truncated RFXAP proteins that could be made in the 6.1.6 cell line and in the ABI fibroblast cell line derived from the MHCII deficiency patient. wt, wild type. (B) FACS analysis of HLA-DR and mCD8 expression on ABI cells transduced with CIITA alone (top) and together with wild-type RFXAP (middle) or RFXAP with the N9 deletion (bottom). Similar numbers of HLA-DR-positive cells are observed when ABI cells are cotransduced with CIITA and wild-type RFXAP (9.5% of mCD8-positive cells) or with CIITA and RFXAP with the N9 deletion (7.3% of mCD8-positive cells). No HLA-DR-positive cells are observed when only CIITA is transduced.

FIG. 5

The putative NLS present in the basic region is not essential for the function of RFXAP. (A) Sequences of the putative NLS in wild-type RFXAP (wt), of the NLS of nucleoplasmin (NPL), and of the two mutants (M1 and M2) are shown. Bold residues have been shown to be essential for NLS function in nucleoplasmin. (B) FACS analysis of HLA-DR expression on nontransduced 6.1.6 cells (open histograms) and 6.1.6 cells transduced with the M1 and M2 mutants of HRFXAP (black histograms).

FIG. 6

The glutamine-rich region is essential for the function of RFXAP. (A) FACS analysis of HLA-DR expression on nontransfected 6.1.6 cells (open histograms) and 6.1.6 cells complemented with wild-type RFXAP (wt) or with RFXAPs with deletions C1 to C4 (black histograms). Wild-type RFXAP and RFXAPs with C-terminal deletions are represented schematically, and their sizes in amino acids are provided. (B) Binding of the RFX complex was analyzed by EMSA with nuclear extracts derived from TK6 B cells, nontransfected 6.1.6 cells, 6.1.6 cells complemented with wild-type RFXAP, and 6.1.6 cells complemented with the version of RFXAP with the C1 or C2 deletion. All of the lanes are from the same gel at the same length of exposure.

FIG. 7

Identification of a segment within RFXAP that has an MHCII isotype-specific function. (A) FACS analysis of the expression of HLA-DR, HLA-DP, HLA-DQ, and MHCI at the surfaces of nontransduced 6.1.6 cells (open histograms) and 6.1.6 cells transduced with wild-type RFXAP (wt) or RFXAPs with the deletions N4 to N9 (black histograms). The transduced RFXAP proteins are represented at the left. (B) RNase protection analysis of the expression of the HLA-DRA, HLA-DRB, HLA-DQA, HLA-DQB, and HLA-A mRNAs in nontransduced 6.1.6 cells and in 6.1.6 cells transduced with wild-type RFXAP or versions of RFXAP with the N4, N7, and N9 deletions. GAPDH mRNA was included as an internal standard. (C) Expression levels measured by RNase protection experiments were quantified by phosphorimager analysis. All values were normalized with respect to GAPDH mRNA expression. For each mRNA, the level attained by complementation with wild-type RFXAP was defined as 100%. (D) Schematic representation of the regions within RFXAP that are required for expression of the three MHCII isotypes. A short 43-amino-acid C-terminal region is sufficient for expression of HLA-DR. In contrast, a 130-amino-acid C-terminal region is required for expression of HLA-DP and HLA-DQ. These two regions define an 87-amino-acid domain having HLA-DQ- and HLA-DP-specific functions.

FIG. 8

Isotype-specific role of RFXAP in inducing promoter occupation by RFX and CIITA in vivo. (A and B) The in vivo occupation of the DRA, DQB, and DPB promoters by RFX and CIITA was studied by chromatin immunoprecipitation experiments with Raji cells, 6.1.6 cells, and 6.1.6 cells complemented with the N7 construct. The CD20 promoter was chosen as a negative control. Results were visualized by gel electrophoresis of the PCR products (A) or quantified by real-time PCR (B). Results are given in percentages relative to values obtained for RFX and CIITA in Raji cells. PI, control immunoprecipitation performed with preimmune serum. (C) Dissociation rates were determined by EMSA for the higher-order protein-DNA complexes formed by simultaneous binding of RFX, NF-Y, and X2BP to the DRA promoter. The complexes were assembled using extracts from Raji cells containing wild-type RFX or 6.1.6 cells complemented with the N8 construct. The gels were quantified by phosphorimager analysis (graphs). The percentages of protein-DNA complexes remaining are plotted as a function of time. The curves represent the means of results from four independent experiments. Representative gels are shown as an inset.