This information is current as
of June 18, 2017.
A Hierarchy of Nuclear Localization Signals
Governs the Import of the Regulatory Factor
X Complex Subunits and MHC Class II
Expression
Uma M. Nagarajan, Alyssa B. Long, Michelle T. Harreman,
Anita H. Corbett and Jeremy M. Boss
J Immunol 2004; 173:410-419; ;
doi: 10.4049/jimmunol.173.1.410
http://www.jimmunol.org/content/173/1/410
Subscription
Permissions
Email Alerts
This article cites 39 articles, 20 of which you can access for free at:
http://www.jimmunol.org/content/173/1/410.full#ref-list-1
Information about subscribing to The Journal of Immunology is online at:
http://jimmunol.org/subscription
Submit copyright permission requests at:
http://www.aai.org/About/Publications/JI/copyright.html
Receive free email-alerts when new articles cite this article. Sign up at:
http://jimmunol.org/alerts
The Journal of Immunology is published twice each month by
The American Association of Immunologists, Inc.,
1451 Rockville Pike, Suite 650, Rockville, MD 20852
Copyright © 2004 by The American Association of
Immunologists All rights reserved.
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
References
The Journal of Immunology
A Hierarchy of Nuclear Localization Signals Governs the
Import of the Regulatory Factor X Complex Subunits and
MHC Class II Expression1
Uma M. Nagarajan,2,3* Alyssa B. Long,2* Michelle T. Harreman,† Anita H. Corbett,† and
Jeremy M. Boss4*
M
ajor histocompatibility complex class II proteins play
the important role of presenting peptide Ags to the
CD4 Th cells. Recognition of MHC/peptide complexes by CD4 T cells results in T cell proliferation and activation
of the regulatory function of the T cell. The importance of MHC
class II expression is exemplified in the bare lymphocyte syndrome
(BLS),5 a severe primary immunodeficiency in which patients fail
to mount CD4 T cell-mediated immune responses (reviewed in
Refs. 1 and 2). BLS patients cannot transcribe MHC class II genes
due to mutations in one of four transcription factors required for
MHC class II expression. Cell lines derived from BLS patients were
used to identify, clone, and characterize the four BLS-specific MHC
II transcription factors: CIITA (3), RFX5 (4), RFXAP (5), and RFXB/ANK (6, 7).
MHC class II genes are coordinately regulated by a set of conserved, upstream elements, termed the X1, X2, and Y boxes (reviewed in Refs. 8 and 9). The three BLS factors, RFX5, RFXAP,
and RFX-B, form the regulatory factor X (RFX) complex. RFX
Departments of *Microbiology and Immunology and †Biochemistry, Emory University School of Medicine, Atlanta, GA 30322
Received for publication September 29, 2003. Accepted for publication April
21, 2004.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1
This work was supported by National Institutes of Health Grants AI34000 (to
J.M.B.) and GM58728 (to A.H.C.).
2
U.M.N. and A.B.L. are co-first authors.
3
Current address: Department of Microbiology and Immunology, University of Arkansas for Medical Sciences, Little Rock, AR 72205
4
Address correspondence and reprint requests to Dr. Jeremy M. Boss, Department of
Microbiology and Immunology, Emory University School of Medicine, Atlanta, GA
30322. E-mail address: [email protected]
5
Abbreviations used in this paper: BLS, bare lymphocyte syndrome; 3AT, 3-amino1,2,4-triazole; CRM-1, chromosomal region maintenance-1; DAPI, 4⬘,6-diamido-2phenylindole hydrochloride; LMB, leptomycin B; NES, nuclear export sequence;
NLS, nuclear localization sequence; RFX, regulatory factor X; X-gal, 5-bromo-4chloro-3-indolyl-␤-D-galactoside.
Copyright © 2004 by The American Association of Immunologists, Inc.
binds to the X1 box (reviewed in Refs. 1 and 8) in a cooperative
manner with X2BP/CREB and NF-Y, which bind to the downstream X2 and Y boxes, respectively (10 –13). The BLS factor
CIITA does not bind DNA directly, but interacts with the factors
bound to the X-Y box region. CIITA functions to aid in the recruitment/stability of the basal transcription machinery through its
potent transcriptional activation domain (14 –16). Through the recruitment of histone acetyltransferases, CIITA association with
MHC class II promoters results in the opening of the local chromatin structure (17–19).
In recent years associations between the subunits of the RFX
complex and CIITA have been described (20 –22). RFX5 is the
only member of the RFX complex with a distinct DNA binding
domain, but it cannot function unless it is associated with RFX-B
and RFXAP. Interactions between the subunits of the RFX complex have been mapped to the N terminus of RFX-5, the ankyrin
repeats of RFX-B, and the C terminus of RFXAP. Although the
interaction domains have been mapped, the question of how this
transcription complex is targeted to its site of action within the
nucleus has not been addressed.
Although diffusion can play a role in the nuclear localization of
small molecules, the import or export of large molecules into and
out of the nucleus is a regulated event that requires specific signaling sequences within the protein cargos (23). The most wellcharacterized nuclear targeting signal is the classical nuclear localization sequence (NLS), which mediates nuclear protein import
(24). Classical NLS can be either monopartite, a single cluster of
basic amino acids, or bipartite, two clusters of basic amino acids
separated by a nonconserved linker sequence of 6 –12 aa (25, 26).
Classical NLSs are recognized in the cytoplasm by a heterodimeric
receptor composed of importin/karyopherin ␣ and importin/karyopherin ␤ (24). Importin ␣ binds to and recognizes the NLS, and
importin ␤ targets the complex to the nuclear pore. Once in the
nucleus, the small GTPase, RanGTP, binds to importin ␤ and triggers dissociation of the complex and release of the NLS cargo into
the nucleus (24). Export from the nucleus is also a signal-mediated
0022-1767/04/$02.00
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
Comprised of RFX5, RFXAP, and RFX-B/ANK, the regulatory factor X (RFX) complex is an obligate transcription factor
required for the expression of MHC class II genes. RFX functions by binding to the conserved X1 box sequence located upstream
of all MHC class II genes. Using a mutagenesis scheme and a yeast heterologous reporter system, the mechanism by which the RFX
complex is transported into the nucleus was examined. The results have identified specific nuclear localization signals (NLS) in
both RFX5 and RFXAP that direct the nuclear translocation and expression of MHC class II genes. Additionally, a nuclear export
signal was identified in the N terminus of RFXAP. RFX-B was poorly localized to the nucleus, and no specific NLS was identified.
Whereas RFX5 could import an RFXAP NLS mutant into the nucleus, it had no effect on the import of RFX-B. The results
suggest that although RFX5 and RFXAP could assemble before nuclear import, RFX-B association with the complex does
not take place until after the subunits enter the nucleus. The identification of nuclear import and export sites on RFX
molecules provides potential targets to modulate MHC class II expression. The Journal of Immunology, 2004, 173: 410 – 419.
The Journal of Immunology
Materials and Methods
Cell lines
The SJO and 6.1.6 B cell lines are defective for RFX5 and RFXAP, respectively (4, 5, 29). SJO cells were grown in Ham’s F-12-DMEM with
20% FBS, 2 mM glutamine, penicillin (5 U/ml), and streptomycin sulfate
(5 ␮g/ml). 6.1.6 and COS-7 (American Type Culture Collection, Manassas,
VA) cells were grown in IMDM and DMEM, respectively (Cellgro/Mediatech, Herndon, VA), with 10% FBS and the above supplements.
GFP-RFX constructs
GFP-tagged constructions were made using the vector pEGFP-C2 (Clontech Laboratories, Palo Alto, CA) with a GFP tag on the N terminus.
Wild-type RFX5 was cloned into the HindIII-SmaI site after PCR amplification of the coding region, creating G-RFX5. The 5⬘ and 3⬘ primers used
were CCCAAGCTTGGCAGAAGATGAGCCT and GATCAGATCT
TCATGGGGGTGTTGC, respectively. All point mutations were generated
by overlap PCR using the primers described below and were confirmed by
sequencing. G-RFX5⌬461– 476 was made by overlap PCR using the primers
TACAGCAAGTGATGCCTCAGGTGGAAGTGG and CCACTTCCAC
CTGAGGCATCACTTGCTGTA. RFX5 N-terminal deletions were made
using the following forward primers: G-RFX5410 – 616, GGCCCAAGCTT
GGGCACAGAGAACAGAGAGG; and G-RFX5453–616, CCCAAGCTTG
CAGGATATAGAGGAT. G-RFX51– 409 was subcloned from the plasmid
RFX5⌬6 (21). RFX5 C-terminal deletions were made using the following
3⬘ primers: G-RFX51– 427, CAATGCATTGGTTCTGCAGTCAGTCATG
TGGTCCTTGGTC; G-RFX51– 455, CAATGCATTGGTTCTGCAGTCAA
TCCTCTATATCCTGCTT; and G-RFX51– 487, CAATGCATTGGTTCT
GCAGTCATGACTTGAGAGGGGTAGA. Point mutations at RFX5
NLS were made using the following sets of overlap primers for
PCR: G-RFX5NLS1, GACCACATGACGCGGGTGTCGCGCGGACAGC
TGAA and TTCAGCTGTCCGCGCGACACCCGCGTCATGTGGTC; and
G-RFX5NLS2, GCTCCACCAGCTGCAGCAGCAGCGCAGGATATAG and
CTATATCCTGCGCTGCTGCTGCAGCTGGTGGAGC.
RFXAP and its deletion constructs (122–272, 179 –272, and 1–245)
were made by subcloning from the wild-type pEcoHis6 RFXAP vector
described previously (21) using EcoRI and PmeI and cloning into the
EcoRI and SmaI sites of the pEGFP-C2 vector. All point mutations were
made by overlap PCR using the primers described below and were cloned
into the EcoRI and SmaI sites. The overlap primers used for point mutations were: G-APNLS1, CAAACCGTGGATGTGCGCGGCACACCG
CAACAAGA and TCTTGTTGCGGTGTGCCGCGCACATCCACGGT
TTG; and G-APNLS2, CAAGGACAAGTATGCAGCGGCGAAGAGC
GACCAG and CTGGTCGCTCTTCGCCGCTGCATACTTGTCCTTG.
The 5⬘ and 3⬘ primers used for these PCR were described previously (21).
RFXAP N-terminal deletions for the nuclear export study were made by
PCR using the following forward primers: G-AP19 –272, GCGGAATTC
CCCCACCCCGCGGCCCTA; G-AP41–272, GCGGAATTCGCCTCT
CAATTCACCCTGC; G-AP 63–272 , GCGGAATTCGGCAGCGTT
GGGGCGGGCAAG; and G-AP85–272, GCGGAATTCGCTGGGGAGG
ACGAGGCGGA. G-RFX-B was constructed by subcloning RFX-B from
pEcoHis6 vector (21), using the EcoRI and PmeI sites and cloning into
EcoRI- and SmaI-digested pEGFP-C2. After transfection into COS-7 cells,
all GFP-tagged constructs were checked for expression by Western blot
using an anti-GFP Ab (Santa Cruz Biotechnology, Santa Cruz, CA) at a
1/1000 dilution.
Transfection, microscopy, and flow cytometry
All COS-7 transfections were conducted using FuGene (Roche, Indianapolis, IN) according to the manufacturer’s instruction. For microscopy, cells
were grown at a density of 50,000 cells/well in a 12-well plate on top of a
poly-L-lysine-coated coverslip. At 24 – 48 h post-transfection, the cells
were washed with 0.1% BSA in PBS and fixed with 1% paraformaldehyde
for 10 min. The cells were mounted on slides with a drop of Fluoromount
G (Fisher Scientific, Suwanee, GA) and allowed to dry for 2–3 h before
fluorescence microscopy (30). For export studies, COS-7 cells were treated
with 20 nM LMB (Sigma-Aldrich, St. Louis, MO) 24 h after transfection,
and washed and fixed 3 h later. Cells were examined using a DM-XRA
microscope (Zeiss, New York, NY) with a ⫻40 objective. Images were
captured using the SPOT camera and imaging system. Quantitation of images from these experiments was performed using Photoshop software
(Adobe Systems, Mountain View, CA) for masking nuclear and cytoplasmic compartments. Three or more cells per field were quantitated, and the
percentage of GFP that was nuclear was plotted in the indicated figures. In
some experiments cells were stained with 4⬘,6-diamido-2-phenylindole hydrochloride (DAPI; 1 ␮g/ml; Sigma-Aldrich) and an Alexa-fluor-conjugated phalloidin stain for cytoplasmic filamentous actin (1 ␮g/ml; Molecular Probes, Eugene, OR). These preparations were examined on a Zeiss
200M inverted microscope with a ⫻63 objective, and the images were
captured using a scientific grade, cooled, charge-coupled device (CoolSnap HQ with ORCA-ER chip; Intelligent Imaging Innovations, Baltimore, MD). To determine the percentages of GFP protein in the cytoplasmic and nuclear compartments, Slidebook software version 4.0 (Intelligent
Imaging Innovations) was used. Up to 11 cells from multiple fields were
masked for their nuclear or cytoplasmic compartments, as delineated by
DAPI or phalloidin staining, respectively, and the percentage of GFP in
these compartments was determined. Statistical analysis of the data was
performed using Student’s t test with the InStat program (GraphPad, San
Diego, CA).
For biochemical subcellular localization experiments, the transfection
reactions were scaled up. Cells (1 ⫻ 105) were plated in a six-well plate.
The GFP cDNA constructions were transfected using 1–2 ␮g of DNA/well
as described above.
All B cell lines were transfected with 40 ␮g of GFP-tagged constructs
by electroporation as described previously (21). Ten micrograms of the red
fluorescence protein-expressing plasmid pDsRed2-C1 (BD Biosciences,
San Diego, CA) was added to all transfections to gate for transfected red
cells during flow cytometry. Cells were stained using unconjugated
HLA-DR (L243 monoclonal supernatant), HLA-DQ (clone SK10; BD Biosciences), and HLA-DP (B7/21; BD Biosciences) antibodies. Cells were
then stained with allophycocyanin-conjugated anti-mouse Abs (Caltag
Laboratories) and analyzed by flow cytometry on an FL4 channel after
gating for cells in the FL2 channel (pDsRed2 positives) using a FACSCalibur (BD Biosciences).
Subcellular fractionation and Western blot analysis
COS-7 cells were harvested 24 h after transfection with the GFP fusion
protein constructions. One quarter of the cells were separated to prepare
unfractionated cellular lysates. The remaining cells were processed for nuclear and cytoplasmic fraction separation as described by Cressman et al.
(30). The cells were centrifuged and suspended in 200 ␮l of buffer A
(containing 10 mM HEPES (pH 7.9), 10 mM KCl, 0.1 mM EDTA, 0.1 mM
EGTA, 1 mM DTT, and 0.5 mM PMSF) and were left on ice for 15 min
before adding 20 ␮l of 10% Nonidet P-40. The cells were vortexed for
20 –30 s and spun for 5 min at 3000 rpm to spin down the nuclei. The
cytoplasm fraction was saved, and the nuclear pellet was washed once with
buffer A containing 1% Nonidet P-40. Nuclei were lysed in 50 ␮l of buffer
B (containing 20 mM HEPES (pH 7.9), 400 mM NaCl, 1 mM EDTA, 1
mM EGTA, 1 mM DTT, and 1 mM PMSF). The nuclear lysates were
rocked at 4°C for 1 h. The lysate was cleared by centrifugation at 14,000
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
process. A loose consensus nuclear export sequence (NES) of hydrophobic residues has been described (LXXXLXXLXL) (27).
This sequence is recognized by the nuclear export receptor, chromosomal region maintenance-1 (CRM-1)/exportin 1, which is a
member of the importin ␤ family of transport receptors (24). Identification and characterization of NESs have been aided by the
antifungal, antibiotic leptomycin B (LMB), which specifically
binds to and inhibits the action of CRM-1 (28).
Because the processes of nuclear import and export can be regulated, the present study was conducted to define the nuclear targeting signals within the RFX complex and to address the link
between complex association and nuclear localization. Little is
known about how complex transcription factors enter the nucleus.
Do the subunits of multisubunit transcription factors translocate
independently and assemble in the nucleus, or do they assemble
before nuclear import and are then transported as a complex? The
results showed that RFX5 contains a primary atypical bipartite
NLS that drives RFX5 to the nucleus. RFXAP contains an NLS
and a novel export signal. The presence of an export signal in one
of the RFX complexes suggests another possible level of MHC class
II regulation. Mutations of the NLS signals have consequences for
RFX function, displaying a greater effect on the expression of MHC
class II isotypes HLA-DQ and HLA-DP compared with that of HLADR. Cotransfection studies using wild-type and mutant RFX subunits
suggest that association of RFX5 and RFXAP may precede nuclear
transport, whereas complete assembly of the complex with RFX-B
occurs in the nucleus.
411
412
rpm for 10 min. The whole cell lysates were prepared using Nonidet P-40
lysis buffer (50 mM Tris (pH 8), 300 mM NaCl, and 1% Nonidet P-40, 1
mM DTT, and 0.5 mM PMSF). The cells lysed with this buffer were sonicated for 30 s and cleared by centrifugation at 14,000 rpm for 10 min. The
protein concentration on all extracts was determined using the Bradford
dye method (Bio-Rad, Hercules, CA), and 10 ␮g of lysate was separated
using SDS-PAGE (9% acrylamide). SDS-PAGE gels were processed following standard Western blotting procedures as previously described (31).
The immunoblots were stained with anti-GFP Abs and HRP-coupled antirabbit secondary Abs (Sigma-Aldrich) and were developed using the ECL
system (Amersham, Arlington Heights. IL). These experiments were performed three times with similar results.
Raji B cell nuclear and cytoplasmic extracts were prepared as described
previously (6). Fifty micrograms of nuclear and cytoplasmic extracts were
analyzed by SDS-PAGE, followed by Western blotting, using rabbit antisera to RFX5 (32), RFX-B, and RFXAP. Western blots were processed as
described above.
Yeast reporter assay for nuclear import
Results
RFX5 and RFXAP have NLS signals, and RFXAP has an export
signal
To initially identify NLS or NES within the RFX subunits, visual
inspection of the primary sequences of the RFX subunits and a
computer search program (http://psort.nibb.ac.jp) were used. The
analysis predicted two classical NLSs for RFX5, one beginning at
aa 144 and the other at aa 469. A single bipartite NLS was predicted for RFXAP beginning at aa 163, but no classical NLS was
identified in RFX-B. No obvious classical NESs were found in any
of the RFX proteins.
To assess the roles of these predicted classical NLSs and to
examine the localization of the RFX complex in vivo, RFX5,
RFXAP, and RFX-B GFP fusion protein expression constructions
were generated, with GFP fused to the 5⬘ end of each of the coding
sequences. This created the three expression vectors: G-RFX5, GRFXAP, and G-RFX-B. To examine the cellular localization of each
individual fusion protein, each of the vectors was transfected into
COS-7 cells. COS-7 cells were chosen over B cells because of their
relatively large and distinct cytoplasmic compartment and the fact that
they have very low levels of RFX subunits, allowing the targeting of
each subunit to be examined independently of the other subunits.
Twenty-four hours after transfection, the cells were examined by fluorescence microscopy (Fig. 1A, top row). Of the three proteins, only
G-RFX5 was localized completely to the nucleus. G-RFXAP appeared to be mostly nuclear, but there was significant fluorescence in
the cytoplasm. In contrast, G-RFX-B was not clearly targeted, as its
distribution in the cytoplasm and nucleus was the same as that of the
control GFP protein.
To determine the percentage of GFP visible in cytoplasmic and
nuclear compartments for the experiments performed in Fig. 1A,
additional slides were prepared with DAPI (nuclear) and phalloidin
staining (cytoplasmic filamentous actin). The RFXAP experiment
(with or without LMB) is shown as an example in Fig. 1B. Using
these stains to highlight nuclear and cytoplasmic compartments, the
GFP signal was quantitated, and the percentage of nuclear GFP was
determined (Fig. 2A). The results showed that RFX5 was ⬎85% nuclear, and that the distributions of RFXAP and RFXB were statistically similar to that of GFP itself (38 – 43% nuclear).
To analyze the subcellular distribution of the RFX proteins, nuclear and cytoplasmic extracts were prepared from the transfected
cells. Immunoblots were performed using anti-GFP Abs (Fig. 2B).
The results showed that RFX5 was completely localized to the
nucleus, RFXAP was evenly distributed between the nucleus and
the cytoplasm, and RFX-B was mostly cytoplasmic in its localization. The GFP control, which was expressed at much higher
levels than the others, was found in both the cytoplasm and nuclear
compartments, although the level in the cytoplasm was greater
than that in the nucleus. A similar analysis was performed on B
cell extracts to determine whether the localization was due to overexpression of each of the RFX proteins or the presence of the GFP
tags. Nuclear and cytoplasmic extracts were prepared from Raji B
cells. Western blot analysis was performed using Abs generated to
RFX5, RFX-B, and RFXAP (Fig. 2C). These results confirm the
microscopic analysis and show that RFX5 is completely nuclear,
whereas RFXAP and RFX-B can be found in both compartments.
Thus, the current expression test system recapitulates what is normally observed in cells where the RFX complex is active.
To determine whether the presence of G-RFXAP and G-RFX-B
in the cytoplasm was due to inefficient import into the nucleus or
to active export from the nucleus, cells were treated with LMB, an
inhibitor of the CRM-1-dependent nuclear export system (28, 35).
Treatment with LMB altered the intracellular distribution of GRFXAP, with the majority of the G-RFXAP being nuclear (Fig. 1).
This was confirmed through quantitation of the images (Fig. 2A).
LMB had no effect on G-RFX5, G-RFX-B, or GFP alone (Figs. 1A
and 2A). These experiments suggest that RFX5 and RFXAP both
contain nuclear localization signals. In addition, RFXAP contains
a CRM-1-dependent export signal. The observed distribution of
G-RFXAP therefore reflects an equilibrium between import and
export of the protein into and out of the nucleus. RFX-B’s distribution was unaffected by LMB, suggesting that its localization is
not influenced by the CRM-1/exportin-1 system.
RFX5 has a distinct NLS at its C terminus
To map the exact location of the NLS within RFX5, N- and C-terminal deletions of RFX5 were created in the GFP expression vector
(Fig. 3A). All the deletion constructs expressed proteins of the expected size (Fig. 3B). Upon transfection into COS-7 cells, the Cterminal deletion, G-RFX51– 409, was completely excluded from the
nucleus, displaying a cytoplasmic and perinuclear distribution, suggesting that the RFX5 NLS is located C-terminal to aa 409 (Fig. 3C).
This implies that the predicted NLS at aa 144 is not a functional NLS.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
The yeast system for assaying nuclear import was performed as previously
described with slight modifications (33). The yeast pNIA vector was used
as the base vector for testing the NLSs. All NLS test constructions were
made by annealing oligonucleotides that created BamHI and PstI overhangs. After digestion with the above enzymes, the fragments were cloned
directly in-frame with the Gal4 activation domain within pNIA. The
RFXAP NLS, RFX5462– 474, RFX5 NLS1, and RFX5 NLS2 vectors were
made using the following oligonucleotide pairs: GATCCAATGCAA
GAAACACCGCAACAAGATGTACAAGGACAAGTATAAAAAGA
AGAAGAGCTGACTGCA and GTCAGCTCTTCTTCTTTTTATACTT
GTCCTTGTACATCTTGTTGCGGTGTTTCTTGCATTG; GATCC
AAGCCAAAAGGAAACGGGGGCGCCCTCGAAAAAAGTCATGA
CTGCA and GTCATGACTTTTTTCGAGGGCGCCCCCGTTTCC
TTTTGGCTTG; GATCCAAGACAAGGGTGTCAAGAGGACATG
ACTGCA and GTCATGTCCTCTTGACACCCTTGTCTTG; and
GATCCAAGCTAAAGCAGCAAAGCAGTGACTGCA and GTCACT
GCTTTGCTGCTTTAGCTTG, respectively. RFX5426 – 453 was made by
annealing GCGCGCGGATCCAAGACAAGGGTGTCAAGAGGACA
GCTGAAGTACCTGTGAGTGAGGCCAGTGGGCAGGCTCCAC and
TTCCAATGCATTGGCTGCAGTCACTGCTTTGCTGCTTTAGCTG
GTGGAGCCTGCCCACTG, extending the overlaps (underline) with the
Klenow fragment of DNA polymerase, digestion with BamHI and PstI, and
cloning the fragment into pNIA vector. The Saccharomyces cerevisiae
strain, L40, which contains the Lex operator integrated upstream of the
HIS3 and lacZ reporter genes was used for all experiments (33). L40 cells
were transformed with each of the pNIA plasmids by standard methods
(34). Single colonies were grown for 72 h, serially diluted (1/9), and
spotted on appropriate media. Spotting was conducted on minimal medium
selection plates deficient for tryptophan (Trp⫺), tryptophan and histidine
(Trp⫺His⫺), or tryptophan and histidine and supplemented with 100 mM
3-amino-1,2,4-triazole (Trp⫺His⫺ 3AT). For ␤-galactosidase assays, cells
were spotted on plates with minimal medium deficient in tryptophan and
supplemented with 5-bromo-4-chloro-3-indolyl-␤-D-galactoside (X-gal).
Plates were examined 48 –72 h later for growth at 30°C and blue color.
NUCLEAR IMPORT OF THE RFX COMPLEX
The Journal of Immunology
Like the full-length G-RFX5 construct, a GFP fusion containing aa
410 – 616 of RFX5 (G-RFX5410 – 616), was sufficient to confer complete nuclear localization of the chimeric protein (Fig. 3C). However,
a slightly smaller construction, G-RFX5453– 616, was no longer efficiently targeted to the nucleus, suggesting that additional signals are
likely to reside between 410 and 452.
FIGURE 2. Subcellular localization of RFX subunits. A, Images from
experiments presented in Fig. 1 were quantitated as described in Materials
and Methods. The amount of GFP fluorescence in at least 10 cells from
each experiment was quantitated and averaged, and the mean percentage of
nuclear GFP was plotted with its SD. ⴱ, p ⬍ 0.0001 vs the no-LMB-treated
sample. B, Lysates prepared from transfected COS-7 cells were separated
into nuclear (N) and cytoplasmic (C) fractions and analyzed by Western
blotting using antiserum to GFP. Total cellular lysates (T) are also shown.
The arrow indicates the migration of full-length protein. C, Lysates prepared from Raji cells were fractionated and analyzed by Western blotting
using antisera directed against each of the subunits as indicated.
RFX5 has two functional NLSs
To map the RFX5 NLS site precisely, a full-length G-RFX5 construct containing a complete deletion of the computer predicted
FIGURE 3. RFX5 has a major bipartite NLS located between aa 427 and 455. A, GFP-RFX5 fusion constructions are schematically shown. The names
of the constructions indicate the amino acids included (with the exception of ⌬461– 476, which indicates the deleted sequences). G, GFP cassette; DBD, DNA
binding domain; pro-rich, proline-rich domain of RFX5. ⴱ, Location of the computer-predicted NLS for RFX5 beginning at aa 469. Alanine substitution
constructions to map the NLS were constructed in the context of the entire RFX5 gene. The amino acid sequence between 427 and 455 is shown with the
corresponding alanine substitutions. Each of these constructs was transfected into COS-7 cells. B, Lysates from these cells were analyzed for the relative
expression of the GFP fusion protein by Western blot analysis using an anti-GFP antiserum. C, Fluorescent microscopic images of the transfections are
shown. These experiments were performed more than three times with identical results.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
FIGURE 1. Nuclear import and export of the RFX subunits. GFP (G) or
GFP-RFX subunit chimeric expression vectors (G-RFX5, G-RFXAP, and
G-RFX-B) were transfected into COS-7 cells. The cells were incubated in
the presence or the absence of LMB and were assayed 24 h post-transfection. A, The cells were viewed by fluorescence microscopy and a sample of
each field is shown. These experiments were performed more than three
times with identical results. B, Additional preparations of GFP-RFXAP
transfections were counterstained with DAPI (blue) and phalloidin (red) to
highlight nuclear and cytoplasmic compartments. In the two pairs of photographs shown, the left panel (green) shows the GFP signal only, whereas
the right panel shows all three signals combined.
413
414
Complementation of MHC class II function by RFX5 deletions
and RFX5 point mutations
An important index of NLS activity is whether the NLS contributes to the function of the protein. RFX5 function and the role of
its NLS can be assayed by rescue of MHC class II function in the
RFX5-deficient B cell line SJO (4). For RFX5 to complement the
RFX5-deficient, BLS B cell line SJO, it must form an active com-
plex with the other RFX subunits, RFX-B and RFXAP. These interactions as well as those with CIITA require the N-terminal two-thirds
of RFX5 (21). In addition, to function, the RFX complex must be
localized to the nucleus. Thus, to test whether the NLS mutations
affect RFX function, SJO cells were transiently transfected with the
GFP RFX5 constructs. Cells were cotransfected with pDsRed2 plasmid, a red-shifted fluorescent protein that allows for gating of all
transfected cells during flow cytometric analysis. The GFP tag on the
RFX5 proteins was not used for gating, because the different constructs showed variable green intensities. To measure MHC class II
expression, the cells were stained for all three MHC class II isotypes
(HLA-DR, -DQ, and -DP) and analyzed by flow cytometry (Fig. 4).
G-RFX5 is fully functional and can rescue all HLA isotypes. The
C-terminal deletions G-RFX51– 409 and G-RFX51– 427 displayed
lower activity for HLA-DR, but were unable to fully rescue HLA-DQ
and -DP expression. This reduction of activity is not due to a failure
to interact with the other RFX subunits, because previous work demonstrated that the N-terminal 409 aa were sufficient for interactions
with RFX-B and RFXAP (21). The activity of the RFX5 chimeras
increased as the NLSs were included in the constructs (aa 1– 455 and
1– 487). The influence of the RFX5 NLS mutations, as illustrated by
G-RFX5NLS1, G-RFX5NLS2, and G-RFX5⌬461– 476, was most evident
when assaying the expression of HLA-DQ and -DP. This is probably
due to the fact that these promoters have a lower affinity for binding
RFX than HLA-DR (11, 36). It should be noted that none of the RFX5
NLS mutations completely eliminated HLA-DR expression. This suggests that interactions between RFX5 and the other RFX subunits in
SJO cells may aid in the nuclear import of RFX5. There was the
concern that the overexpression of the GFP fusion proteins overloaded this system. To ascertain whether this was the case, 20-fold
less G-RFX5 DNA was used in the above assays. Although fewer
MHC class II-positive cells were observed under these conditions, the
patterns of MHC class II expression were similar to the amounts used
above (data not shown), suggesting that the expression system used
did not cause the results observed.
RFXAP has a functional bipartite NLS
To analyze the NLS signal within RFXAP, both basic amino acid
clusters within the bipartite NLS identified by computer analysis
FIGURE 4. Mutation of the RFX5 NLS region reduces MHC class II expression. The indicated GFP-RFX5 fusion expression vectors described in Fig.
3 were cotransfected into the RFX5-deficient cell line SJO with the constitutive expression vector pDsRed2. The pDsRed2-expressing cells, which represent
the transfected pool, were analyzed by flow cytometry. The profiles of the cells stained with Abs against HLA-DR, -DQ, or -DP are indicated by solid
histograms. Open histograms represent the control transfection, in which the empty vector (pEGFP-C2) was used. The control profile was reproduced in
each horizontal set of histograms for comparison. The number in the upper left corner indicates the mean fluorescence intensity of control, transfected cells.
The numbers in the upper right of each histogram indicate the mean fluorescence intensity for cells transfected with GFP-RFX5 constructs. This experiment
was performed twice with similar results.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
469 NLS (G-RFX5⌬461– 476) was generated and analyzed (Fig.
3A). Surprisingly, this deletion had no effect on the nuclear localization of RFX5 (Fig. 3C). Thus, NLS-469, although capable of
targeting some GFP into the nucleus, is not required for nuclear
import of RFX5. This result suggested that a second nuclear targeting signal must be located within aa 410 – 460.
To further map the exact region of RFX5 involved in nuclear
import, a series of deletions lacking the sequences between aa 410
and 487 was constructed and analyzed (Fig. 3A). All the expression vectors produced proteins of the expected size (Fig. 3B). The
smallest construction, G-RFX51– 427, displayed a perinuclear distribution. In contrast, G-RFX51– 455, which lacks the predicted
NLS beginning at aa 469, was primarily localized to the nucleus.
As expected, G-RFX51– 487 was completely nuclear. The sequence
between aa 427 and 455, which resembles a long bipartite NLS,
contains two basic sequence motifs separated by 16 aa (Fig. 3A).
To determine whether these sequences functioned as an NLS, each
of the basic regions was mutated individually to create
G-RFX5NLS1 and G-RFX5NLS2 (Fig. 3A). Both mutants displayed
reduced import into the nucleus as compared with wild-type RFX5
(Fig. 3C). The distribution of G-RFX5NLS1 appeared to be equal
between the cytoplasm and the nucleus, with no cell displaying
complete nuclear localization. In ⬃10% of the cells containing
G-RFX5NLS1, the distribution was mostly cytoplasmic. In contrast,
the distribution of the G-RFX5NLS2 mutant was mostly nuclear,
with some cytoplasmic signal. These experiments therefore suggest that the major NLS for RFX5 is located in the region defined
by aa 428 – 455, and that the N-terminal basic amino acid cluster
(aa 428 – 432) within this bipartite NLS plays the major role. The
computer predicted NLS at aa 469 could play an additive/supplemental role, because it was sufficient to confer nuclear import of
GFP in the chimeric G-RFX5453– 616 protein.
NUCLEAR IMPORT OF THE RFX COMPLEX
The Journal of Immunology
415
export from the nucleus. In contrast, LMB increased the nuclear
pool of G-APNLS1 from 35 to 52% ( p ⫽ 0.001). Given these results, it is likely that the second set of lysine residues in the cluster
is the functional import sequence for RFXAP.
RFXAP has an export signal in its N terminus
were individually mutated to alanine residues (Fig. 5A). The localization of these mutant proteins was examined as described for
RFX5 (Fig. 5B). Mutation of the first basic amino acid cluster,
G-APNLS1, caused no significant defect in nuclear localization of
RFXAP. Mutation of the basic residues in the C-terminal end of
the bipartite NLS (G-APNLS2) had a more drastic effect on nuclear
localization of RFXAP. Quantitative analysis of the data showed a
decrease from 42 to 33% in nuclear content of the G-APNLS2 mutant compared with the wild type ( p ⫽ 0.0119; Fig. 5D). The
difference between G-APNLS1 and wild type was only marginally
significant ( p ⫽ 0.0509). Importantly, LMB had no effect on the
distribution of G-APNLS2 (Fig. 5D), implying that G-APNLS2 is
poorly imported due to the lack of an NLS rather than to active
FIGURE 6. RFXAP contains a nuclear export signal. A and C, GFPRFXAP deletions are schematically
shown. B, GFP-RFXAP deletion constructions shown in A were transfected
into COS-7 cells and cultured in the absence or the presence of LMB. The cells
were examined by fluorescent microscopy. D, Fluorescent microscopy of
COS-7 cells transfected with the deletion
series shown in C.
RFXAP NLS mutants affect MHC class II expression
To test the functionality of the RFXAP NLS mutants, GFP-tagged
RFXAP constructs were transfected into the RFXAP-deficient
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
FIGURE 5. The C-terminal portion of the RFXAP bipartite NLS is required for nuclear localization. A, The sequence of the predicted NLS for
RFXAP along with the locations of two series of alanine substitutions are
shown. Each of the alanine substitution mutants was created in the context
of the full-length RFXAP cDNA. B, Fluorescent microscopy of the wildtype and mutant GFP-RFXAP chimeric vectors. C, Lysates from the transfected COS-7 cells were examined for expression of the GFP-RFXAP constructs from this figure and Fig. 6. D, Transfections of the above constructs
in cells untreated or treated with LMB were counterstained, quantitated,
and plotted as described in Fig. 1B. #, p ⬍ 0.0001 for comparisons between
the indicated sample and its no LMB control (by Student’s t test). ⴱ, p ⬍
0.0001, comparison between the G-AP and G-APNLS2 samples, both
treated with LMB (by Student’s t test).
The experimental evidence presented in Fig. 1 suggested that
RFXAP has a CRM-1-dependent nuclear export signal. To determine the location of the export signal, N- and C-terminal deletions
of RFXAP were generated and cloned into the GFP expression
vector. The constructs were transfected into COS-7 cells. Localization of the various fusion proteins was then examined in the
absence or the presence of the export inhibitor, LMB. Even in the
absence of LMB, the fusion protein lacking the first 121 aa of
RFXAP, G-AP122–272, was completely nuclear, indicating that the
export signal resides in the N terminus of RFXAP (Fig. 6, A and
B). The deletion G-AP179 –272, which lacks the NLS identified
above, also appears to lack the export signal. Hence, it was not
targeted specifically to the nucleus (as in the case of G-APNLS2)
and was not responsive to LMB. Its even distribution could be due
to diffusion, because the chimeric molecule is small. G-AP1–245
lacks the interaction domain with RFX5 and RFX-B (21), but behaves like wild-type RFXAP with regard to nuclear localization
and LMB sensitivity. Taken together, these data suggest that
RFXAP contains a nuclear export signal upstream of aa 122.
The N-terminal sequence of RFXAP was examined visually for
the presence of classical nuclear export signals such as LXXX
LXXLXL (27, 37), but no such sequences were found. To map the
export signal further, a high resolution deletion series of the Nterminal sequences between aa 1 and 85 of RFXAP was created
(Fig. 6, C and D). The micrographs suggest that the export signal
is located between aa 41 and 85, as nuclear import does not increase with further deletion of the N-terminal sequences. LMB
treatment of G-AP41–272, G-AP85–272, or G-AP122–272 transfectants caused no further change in nuclear localization (data not
shown), indicating that the nuclear export signal was not present in
these constructions. The N-terminal sequence within this region of
RFXAP is highly hydrophobic, as are conventional NESs, and may
form a hydrophobic pocket that is recognized by the export
receptor.
416
NUCLEAR IMPORT OF THE RFX COMPLEX
FIGURE 7. NLS mutations in RFXAP show differential MHC class II expression profiles. The 6.1.6 B
lymphoblastoid cell line, which is deficient for
RFXAP (29), was cotransfected with the indicated
GFP-RFXAP constructions (Figs. 5 and 6) and
pDsRed2. Cells were stained with anti-HLA-DR, -DQ,
or -DP Abs and analyzed as described in Fig. 4.
RFX5 and RFXAP NLSs are sufficient for nuclear transport in a
heterologous system
The data indicate that the NLSs identified are necessary for nuclear
import of RFX5 and RFXAP. To address whether these sequences
are also sufficient to target a heterologous protein into the nucleus,
the yeast NLS reporter system developed by Rhee et al. (33) was
used. In this assay the reporter genes HIS3 and lacZ are integrated
into the yeast genome under the control of a LexA operator (Fig.
8A). This reporter strain (L40) is then transformed with a plasmid
(pNIA) that encodes the LexA DNA binding domain fused to the
Gal4p activation domain. For this fusion protein to activate transcription of the reporter genes, it must be targeted to the nucleus.
However, the fusion protein does not contain a nuclear targeting
signal. To assess the function of a putative NLS, that sequence is
cloned in-frame with the LexA/Gal4p fusion protein. If the NLS is
functional in vivo, the fusion protein will be targeted to the nucleus, and the reporter genes, HIS3 and lacZ, will be expressed. In
contrast, when non-NLS control sequences were used, HIS3 and
lacZ were not expressed (33). Controls used for this experiment
include the canonical SV40 NLS, which serves as a positive control for import into the nucleus, and the fusion protein alone lacking an NLS, which serves as a negative control.
To examine the function of the RFX5 and RFXAP NLSs, the
short NLSs were cloned in-frame with the modified LexA DNA
binding domain and Gal4 activation domain in the yeast pNIA
shuttle vector and transformed into yeast cells. The ability of the
transformed yeast cells to grow on plates lacking histidine (His⫺)
and to turn blue on X-gal plates was determined. Growth on His⫺
plates indicates that the HIS3 reporter gene is expressed, and accumulation of the blue pigment on X-gal plates indicates that the
lacZ reporter gene is expressed. Cells transformed with each of the
test plasmids as well as the control plasmids were grown to saturation in liquid culture, serially diluted, and spotted onto control
(Trp⫺), His⫺ 3AT⫹ Trp⫺, and X-gal Trp⫺ plates. The control
plate shows an equal number of cells spotted for each sample (Fig.
8B). On the His⫺ plate, cells expressing a fusion protein that contains the wild-type or any of the test NLSs can grow. No growth
was observed for the negative control plasmid (pNIA alone),
which lacks an NLS. The same NLS-containing constructs turned
blue on the X-gal plate. Thus, RFX5462– 474, RFX5-NLS1, RFX5NLS2, and the combined RFX5 NLS426 – 453, which includes both
NLS1 and NLS2, were sufficient to transport LexA-Gal4 into nucleus and activate expression of the reporter genes. The NLS of
RFXAP can also target the LexA-Gal4 fusion to the nucleus. These
results suggest that each of the NLSs identified within RFX5 and
RFXAP can function independently.
RFX5 is the major importer of the complex, whereas RFXAP
plays a secondary role
The fact that RFX5, RFXAP, and RFX-B must associate to perform their functions raises the issue of where they associate (before or after nuclear import) and how RFX-B, which does not
contain a classical NLS, is targeted to the nucleus. To address
these issues, the ability of an RFX subunit to transport a GFPtagged, NLS-deficient, RFX subunit was analyzed after cotransfection into COS-7 cells. In the first series the ability of RFXAP to
be brought into the nucleus by the other factors was tested (Fig.
9A). In this study G-APNLS2, which is deficient for nuclear targeting, was cotransfected with wild-type, His-tagged RFX5; HisRFX-B; or both expression vectors. In the presence of wild-type
RFX5, most of the G-APNLS2 mutant protein was targeted to the
nucleus (Fig. 9A). Visually, RFX-B had a minimal effect on
the transport of G-APNLS2, and RFX5 and RFX-B together had the
same effect as RFX5 itself. Quantitation of the images showed that
indeed RFX5 expression resulted in increased nuclear localization,
whereas RFX-B had no effect (Fig. 9C). Because it was possible
that the ability to measure the nuclear accumulation of G-APNLS2
by RFX5 or RFX-B was inhibited by the RFXAP export signal, the
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
B cell line 6.1.6 (5, 29). 6.1.6 cells were cotransfected with
pDsRed2; stained with HLA-DR, -DQ, and -DP Abs; and analyzed
by flow cytometry (Fig. 7). G-AP122–272, which lacks the putative
NES domain, activated all three class II isotypes in a manner similar to that of the full-length RFXAP. G-AP179 –272, which lacks
both its NLS and NES, retained half of its HLA-DR activity, but
lost all of its HLA-DQ- and -DP-activating potential. G-AP1–245
lacks a critical region for interacting with RFX5 and RFX-B (21),
and although it localizes to the nucleus, it is unable to complement
the defect in 6.1.6 cells. The RFXAP NLS mutants showed some
effect on class II expression. The G-APNLS2 mutant lost 40% of
HLA-DR and 90% of HLA-DP and was nonfunctional for HLADQ. G-APNLS1 displayed a milder effect and was near wild type
for HLA-DR. However, both HLA-DQ and -DP expressions were
greatly affected (75 and 35% reductions, respectively). As was the
case for RFX5 mutants, none of the RFXAP NLS mutants abrogated HLA-DR expression completely, suggesting that RFXAP
import in B cells may be aided by its interaction with the other
RFX subunits, which are present at wild-type levels in the 6.1.6
cells. It should be noted that the expression of HLA-DQ molecules
is very sensitive to mutations in RFXAP (36). Thus, the changes
observed in HLA-DQ expression probably represent a combination of decreased RFXAP import as well as potential activity
changes due to the mutations themselves.
The Journal of Immunology
affect of LMB on these experiments was tested. However, no discernable difference was observed (data not shown), suggesting that
export did not play a dominant role in these experiments. As a
negative control, two previously described deletions of RFX5 were
also cotransfected with G-APNLS2. The RFX5 deletion HisRFX5261– 616 (previously referred to as RFX5⌬2) contains the
RFX5 NLS, but lacks sequences necessary for interactions with
RFXAP and RFX-B, whereas His-RFX51– 409 (previously referred
to as RFX5⌬6) can interact with RFXAP and RFX-B (21), but is
missing the NLS. Both mutant-RFX5 constructs failed to transport
G-APNLS2 into the nucleus.
Biochemical analysis of the subcellular localization of RFXAPNLS2
after cotransfection with RFX5, RFX-B, or both was also performed
as described above (Fig. 9B). The results showed that the transfection
FIGURE 9. RFX5 can import an NLS-defective
RFXAP into the nucleus. Cotransfections of the indicated plasmids were performed in COS-7 cells, followed by fluorescent microscopy. A, All cotransfectants contained the G-APNLS2 mutant, which does not
localize to the nucleus on its own. RFX5261– 616 cannot interact with RFXAP, and RFX51– 409 lacks the
RFX5 NLS. B, COS-7 cells cotransfected with GAPNLS2 and empty vector, RFX5, RFX-B, or both
were analyzed for the subcellular distribution of the
GFP-RFXAP protein by immunoblot after fractionation of cellular lysates into cytoplasmic (C) and nuclear (N) fractions. C, Experiments presented in A
were quantitated for the amount of GFP in the nucleus as described above. ⴱ, p ⬍ 0.011 vs control (by
Student’s t test). D, All cotransfectants contain GRFX51– 427, which does not enter the nucleus. Transfection with wild-type RFXAP, but not RFX-B, displays weak nuclear localization. E, Cotransfectants
with GFP-RFX-B included empty vector, RFX5, or
RFXAP as indicated.
of RFX5 led to a significant increase in the amount of the RFXAP
NLS mutant found in the nucleus. RFX-B by itself had a very minor
effect on RFXAPNLS2 import. Together, these results suggest that
RFX5 can mediate the nuclear accumulation of RFXAP. Moreover,
the results suggest that RFX5 and RFXAP can interact in the cytoplasm and be imported as a single complex.
To determine whether RFXAP can transport RFX5 into the nucleus, the G-RFX51– 427 mutant, which lacks the NLS, was used in
a similar set of experiments. Cotransfections were performed using
His-RFXAP, His-RFX-B, or both. Only minor differences between
the control and the cotransfections containing RFXAP were observed (Fig. 9D). The addition of RFX-B did not substantially
improve RFXAP’s ability to transport an NLS-defective RFX5
into the nucleus, indicating the requirement for a functional NLS
within RFX5.
To determine whether RFX5 or RFXAP mediates nuclear import of RFX-B, GFP-RFX-B was cotransfected with His-RFX5,
His-RFXAP, or both. The results of this experiment showed that
RFX-B was not efficiently transported into the nucleus by either
subunit (Fig. 9E). Because RFX-B can interact with CIITA (20),
and CIITA is in itself transported into the nucleus, the possibility
that CIITA may transport RFX-B into the nucleus was tested.
However, no combination of CIITA with RFX-B and the other
RFX subunits resulted in the accumulation of RFX-B in the nucleus over the control (data not shown), suggesting an independent
mode of transport.
Discussion
The RFX complex, comprised of RFX5, RFXAP, and RFX-B/
ANK, binds the MHC class II promoter and plays an essential role
in MHC class II expression. The absence of any one of the three
subunits completely abrogates class II expression (reviewed in
Refs. 8 and 9). In the present study the link between nuclear localization and assembly of the RFX complex was explored. Using
a deletion and mutagenesis scheme, nuclear localization signals for
RFX5 and RFXAP were identified and mapped. In contrast, a specific NLS in RFX-B was not identified. Whereas RFX5 was found
to be a nuclear protein that was imported efficiently into the nucleus, RFXAP had the characteristics of a nuclear/cytoplasmic
shuttling protein, with an LMB-sensitive nuclear export signal.
The identified nuclear localization signals for both proteins can
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
FIGURE 8. The NLSs identified within RFX5 and RFXAP function in
a yeast heterologous system. A, The pNIA NLS test vector and yeast reporter genes are schematically shown (33). Test NLSs were cloned into the
multiple cloning site (MCS) indicated in the vector. The SV40 NLS used
as a positive control was cloned into the site indicated by the arrowhead.
B, L40 yeast cells were transformed with the indicated test constructs.
Transformants were serially diluted (1/9) and spotted onto control plates
lacking Trp (Trp⫺), Trp⫺ and His⫺ plates containing 3-aminotriazole
(Trp⫺ His⫺ 3AT⫹), or Trp⫺ plates containing X-gal (Trp⫺ X-gal⫹).
417
418
function independently, because they were sufficient to function in
a heterologous yeast system. Importantly, the data showed that
nuclear localization is a critical event for MHC class II expression,
as mutagenesis of the nuclear localization signals in either RFX5
or RFXAP altered their trans-activation function. The results also
suggest that RFX5 and RFXAP can associate before nuclear
localization.
NLS signals within RFX5 and RFXAP
RFXAP has an export signal
Although the relative levels of the RFX proteins may vary between
cell types, they are thought to be constitutively transcribed (32).
However, their regulation at the post-translational level or through
changes in intracellular localization has not been studied. The presence of an LMB-sensitive export signal on the N terminus of
RFXAP suggests one possible means of RFX regulation or modulation. The export signal of RFXAP does not conform to other
export signals described to date, with no obvious homology to
motifs such as LXXXLXXLXL (27). RFXAP’s export signal is,
however, located in a very hydrophobic region of the protein. It is
possible that the RFXAP export signal is novel and represents a
new class of nuclear export signals. Alternatively, this region may
bind to another protein with a CRM-1-dependent export signal.
The N terminus of RFXAP did not function effectively as an export
signal in the heterologous yeast system or when fused to the RFX5
import signal (data not shown). Therefore, the export signal in
RFXAP is either weak or may function only in the context of the
entire RFXAP protein.
Consequence of nuclear localization on RFX function
Many of the studies describing nuclear localization signals of proteins are in COS-7 cells, in which the protein can be overexpressed
and studied independently from other factors. However, it was
important to determine the roles of the identified NLS in B cells
where these proteins normally function. The BLS-derived cell line
SJO (RFX5 deficient) and the in vitro BLS-like mutant cell line
6.1.6 (RFXAP deficient) provided the opportunity to study the
roles of the RFX5 and RFXAP NLS in B lymphocytes, respectively. Complete deletion of the entire NLS in RFX5 significantly
reduced HLA-DQ and HLA-DP expression, whereas the effect on
HLA-DR expression was less pronounced. None of the NLS mutants completely abolished HLA-DR expression. One explanation
for this is that all the RFX5 NLS mutants contained domains for
interaction with RFXAP, RFX-B, and CIITA. Such interactions
could allow nuclear transport of small amounts of RFX5 into the
nucleus. Because the MHC class II promoters display differential
binding of the RFX complex (11, 36) it is likely that sufficient RFX
was present to activate the DR promoters, but not those for the DQ
or DP genes.
A significant difference in activity between G-RFX51– 409 and
G-RFX51– 427 was observed, although both lack a functional NLS.
This could be due to weaker association with CIITA in the GRFX51– 409 deletion as the interacting region of RFX5 with CIITA
is just proximal to this deletion (21). All RFX5 NLS point mutants
showed reduced HLA-DQ and HLA-DP expression, suggesting
the additive roles of the various NLSs, including the NLS beginning at aa 469. Point mutations within the RFXAP NLS also
showed a greater effect on HLA-DQ and HLA-DP expression than
they did on HLA-DR expression. As with RFX5 NLS mutations,
wild-type RFX5 present in the RFXAP-deficient cell lines could
transport some of the RFXAP-NLS mutant protein into the nucleus. Limiting amounts of this factor would be predicted to function at the stronger HLA-DR promoters over the weaker HLA-DQ
or -DP gene promoters.
RFX5 and RFXAP form a complex before translocation
An interesting aspect of how multicomponent transcription factors
and transcription factor complexes function is whether they assemble before or after nuclear import. In the RFX system examined in this study, a hierarchy in nuclear transport efficiency was
observed among the subunits (Fig. 10). RFX5 had a strong NLS,
as it transported and retained an RFXAP NLS mutant in the nucleus. This suggests that RFX5 and RFXAP assemble in the cytoplasm. Similarly, RFXAP played a secondary role in the transport, as it was able to transport a minor portion of an RFX5 NLS
mutant into the nucleus. Thus, at least these two subunits of the
RFX complex can interact before nuclear import. However,
RFX-B transport appears to be separated from these events, as
FIGURE 10. Model of RFX subunit nuclear translocation. The schematic shows that RFXAP and RFX5 contain classically definable nuclear
import sequences and can each enter the nucleus on their own. RFXAP is
also exported, a process that is CRM-1 dependent and inhibited by LMB.
RFX-B can be imported into the nucleus, but it does not contain a classical
NLS. Ultimately, the molecules assemble on an MHC class II promoter
with the other known MHC class II-specific transcription factors.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
The major RFX5 NLS was found to be an atypical bipartite NLS
(aa 427– 455), in which the N-terminal lysines were more critical
for nuclear import. Bipartite NLSs typically have a 6- to 12-residue
spacer. Thus, with 16 residues separating the charged residues, the
RFX5 NLS is atypical. Another NLS was found starting at aa 469
that is very similar to those observed in several other proteins (24).
This sequence functions as an excellent NLS in the heterologous
yeast system (Fig. 8) and can direct GFP protein into the nucleus.
However, complete deletion of this sequence in full-length RFX5
did not affect nuclear localization. This suggests that this signal
plays only a minor or an additive role, if any, in RFX5 import. It
is possible that in the context of the full-length RFX5 subunit, this
region is not on an exposed surface that can be recognized by the
importin ␣ NLS receptor. Another member of the RFX family,
RFX-1, also contains a strong NLS in its extreme C terminus and
a weak NLS in its DNA binding domain, which may play an additive role (38). However, the extreme C terminus of RFX-1 has no
homology to other RFX proteins, including RFX5.
RFXAP contains a single NLS, which was sufficient for its
transport into the nucleus. Previously, Peretti et al. (36) had suggested that mutations of the lysines (to asparagines) within this
NLS did not affect RFXAP activity, leading to the conclusion that
this region was not important for RFXAP function. It is likely that
the mutations used in that study were less severe than those used
in this study. In this study alanine substitution of three of the four
lysines in the C-terminal portion of the bipartite NLS resulted in
reduced RFXAP nuclear localization. Additionally, the activity of
the mutant (G-APNLS2) was significantly reduced when analyzed
for HLA-DP and -DQ complementation in 6.1.6 cells.
NUCLEAR IMPORT OF THE RFX COMPLEX
The Journal of Immunology
Acknowledgments
We gratefully acknowledge Drs. Valerie Boss, Brian Ciliax, Dan Kalman,
and Patrick Reeves for their help with the use of their fluorescent microscopes for this study. We also thank Dr. Paul Wade and the members of our
laboratory for comments and discussions on this work.
References
1. Reith, W., and B. Mach. 2001. The bare lymphocyte syndrome and the regulation
of MHC expression. Annu. Rev. Immunol. 19:331.
2. DeSandro, A., U. M. Nagarajan, and J. M. Boss. 1999. The bare lymphocyte
syndrome: molecular clues to the transcriptional regulation of major histocompatibility complex class II genes. Am. J. Hum. Genet. 65:279.
3. Steimle, V., L. A. Otten, M. Zufferey, and B. Mach. 1993. Complementation
cloning of an MHC class II transactivator mutated in hereditary MHC class II
deficiency (or bare lymphocyte syndrome). Cell 75:135.
4. Steimle, V., B. Durand, B. Emmanuele, M. Zufferey, M. R. Hadam, B. Mach, and
W. Reith. 1995. A novel DNA-binding regulatory factor is mutated in primary
MHC class II deficiency (bare lymphocyte syndrome). Genes Dev. 9:1021.
5. Durand, B., P. Sperisen, P. Emery, E. Barras, M. Zufferey, B. Mach, and
W. Reith. 1997. RFXAP, a novel subunit of the RFX DNA binding complex, is
mutated in MHC class II deficiency. EMBO J. 16:1045.
6. Nagarajan, U. M., P. Louis-Plence, A. DeSandro, R. Nilsen, A. Bushey, and
J. M. Boss. 1999. RFX-B is the gene responsible for the most common cause of
the bare lymphocyte syndrome, a MHC class II immunodeficiency. Immunity
10:153.
7. Masternak, K., E. Barras, M. Zufferey, B. Conrad, G. Corthals, R. Aebersold,
J.-C. Sanchez, D. F. Hochstrasser, B. Mach, and W. Reith. 1998. A gene encoding a novel RFX-associated transactivator is mutated in the majority of MHC
class II deficiency patients. Nat. Genet. 20:273.
8. Boss, J. M., and P. E. Jensen. 2003. Transcriptional regulation of the MHC class
II antigen presentation pathway. Curr. Opin. Immunol. 15:105.
9. Ting, J. P., and J. Trowsdale. 2002. Genetic control of MHC class II expression.
Cell 109(Suppl.):S21.
10. Reith, W., M. Kobr, P. Emery, B. Durand, C.-A. Siegrist, and B. Mach. 1994.
Cooperative binding between factors RFX and X2BP to the X and X2 boxes of
MHC class II promoters. J. Biol. Chem. 269:20020.
11. Louis-Plence, P., C. S. Moreno, and J. M. Boss. 1997. Formation of a regulatory
factor X/X2 box-binding protein/nuclear factor-Y multiprotein complex on the
conserved regulatory regions of HLA class II genes. J. Immunol. 159:3899.
12. Reith, W., C. A. Siegrist, B. Durand, E. Barras, and B. Mach. 1994. Function of
major histocompatibility complex class II promoters requires cooperative binding
between factors RFX and NF-Y. Proc. Natl. Acad. Sci. USA 91:554.
13. Moreno, C. S., G. W. Beresford, P. Louis-Plence, A. C. Morris, and J. M. Boss.
1999. CREB regulates MHC class II expression in a CIITA-dependent manner.
Immunity 10:143.
14. Riley, J. L., S. D. Westerheide, J. A. Price, J. A. Brown, and J. M. Boss. 1995.
Activation of class II MHC genes requires both the X box region and the class II
transactivator (CIITA). Immunity 2:533.
15. Mahanta, S. K., T. Scholl, F.-C. Yang, and J. L. Strominger. 1997. Transactivation by CIITA, the type II bare lymphocyte syndrome-associated factor, requires
participation of multiple regions of the TATA box binding protein. Proc. Natl.
Acad. Sci. USA 94:6324.
16. Fontes, J. D., B. Jiang, and B. M. Peterlin. 1997. The class II trans-activator
CIITA interacts with the TBP-associated factor TAFII32. Nucleic Acids Res.
25:2522.
17. Beresford, G. W., and J. M. Boss. 2001. CIITA coordinates multiple histone
acetylation modifications at the HLA-DRA promoter. Nat. Immunol. 2:652.
18. Kretsovali, A., T. Agalioti, C. Spilianakis, E. Tzortzakaki, and M. Merika. 1998.
Involvement of CREB binding protein in expression of major histocompatibility
complex class II genes via interaction with the class II transactivator. Mol. Cell.
Biol. 18:6777.
19. Fontes, J. D., S. Kanazawa, D. Jean, and B. M. Peterlin. 1999. Interactions between the class II transactivator and CREB binding protein increase transcription
of major histocompatibility complex class II genes. Mol. Cell. Biol. 19:941.
20. Zhu, X. S., M. W. Linhoff, G. Li, K. C. Chin, S. N. Maity, and J. P. Ting. 2000.
Transcriptional scaffold: CIITA interacts with NF-Y, RFX, and CREB to cause
stereospecific regulation of the class II major histocompatibility complex promoter. Mol. Cell. Biol. 20:6051.
21. DeSandro, A., U. M. Nagarajan, and J. M. Boss. 2000. Associations and interactions among the bare lymphocyte syndrome factors. Mol. Cell. Biol. 20:6587.
22. Nekrep, N., N. Jabrane-Ferrat, and B. M. Peterlin. 2000. Mutations in the bare
lymphocyte syndrome define critical steps in the assembly of the regulatory factor
X complex. Mol. Cell. Biol. 20:4455.
23. Jans, D. A., and S. Hubner. 1996. Regulation of protein transport to the nucleus:
central role of phosphorylation. Physiol. Rev. 76:651.
24. Gorlich, D., and U. Kutay. 1999. Transport between the cell nucleus and the
cytoplasm. Annu. Rev. Cell. Dev. Biol. 15:607.
25. Kalderon, D., B. L. Roberts, W. D. Richardson, and A. E. Smith. 1984. A short
amino acid sequence able to specify nuclear location. Cell 39:499.
26. Robbins, J., S. M. Dilworth, R. A. Laskey, and C. Dingwall. 1991. Two interdependent basic domains in nucleoplasmin nuclear targeting sequence: identification of a class of bipartite nuclear targeting sequence. Cell 64:615.
27. Nigg, E. A. 1997. Nucleocytoplasmic transport: signals, mechanisms and regulation. Nature 386:779.
28. Kudo, N., N. Matsumori, H. Taoka, D. Fujiwara, E. P. Schreiner, B. Wolff,
M. Yoshida, and S. Horinouchi. 1999. Leptomycin B inactivates CRM1/exportin
1 by covalent modification at a cysteine residue in the central conserved region.
Proc. Natl. Acad. Sci. USA 96:9112.
29. Gladstone, P., and D. Pious. 1978. Stable variants affecting B cell alloantigens in
human lymphoid cells. Nature 271:459.
30. Cressman, D. E., K. C. Chin, D. J. Taxman, and J. P. Ting. 1999. A defect in the
nuclear translocation of CIITA causes a form of type II bare lymphocyte syndrome. Immunity 10:163.
31. Morris, A. C., W. E. Spangler, and J. M. Boss. 2000. Methylation of class II
trans-activator promoter. IV. A novel mechansim of MHC class II gene control.
J. Immunol. 164:4143.
32. Moreno, C. S., E. M. Rogers, J. A. Brown, and J. M. Boss. 1997. RFX, a bare
lymphocyte syndrome transcription factor, is a multimeric phosphoprotein complex. J. Immunol. 158:5841.
33. Rhee, Y., F. Gurel, Y. Gafni, C. Dingwall, and V. Citovsky. 2000. A genetic
system for detection of protein nuclear import and export. Nat. Biotechnol.
18:433.
34. Adams, A., D. E. Gottschling, C. A. Kaiser, and T. Stearns. 1997. Methods in
Yeast Genetics. Cold Spring Harbor Laboratory Press, Plainview.
35. Fukuda, M., S. Asano, T. Nakamura, M. Adachi, M. Yoshida, M. Yanagida, and
E. Nishida. 1997. CRM1 is responsible for intracellular transport mediated by the
nuclear export signal. Nature 390:308.
36. Peretti, M., J. Villard, E. Barras, M. Zufferey, and W. Reith. 2001. Expression of
the three human major histocompatibility complex class II isotypes exhibits a
differential dependence on the transcription factor RFXAP. Mol. Cell. Biol.
21:5699.
37. Neufeld, K. L., D. A. Nix, H. Bogerd, Y. Kang, M. C. Beckerle, B. R. Cullen, and
R. L. White. 2000. Adenomatous polyposis coli protein contains two nuclear
export signals and shuttles between the nucleus and cytoplasm. Proc. Natl. Acad.
Sci. USA 97:12085.
38. Katan-Khaykovich, Y., and Y. Shaul. 2001. Nuclear import and DNA-binding
activity of RFX1: evidence for an autoinhibitory mechanism. Eur J. Biochem.
268:3108.
39. Fan, P., F. Dong, Y. Huang, and G. Zhong. 2002. Chlamydia pneumoniae secretion of a protease-like activity factor for degrading host cell transcription
factors required for major histocompatibility complex antigen expression. Infect.
Immun. 70:345.
40. Zhong, G., P. Fan, H. Ji, F. Dong, and Y. Huang. 2001. Identification of a chlamydial protease-like activity factor responsible for the degradation of host transcription factors. J. Exp. Med. 193:935.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
neither RFX5 nor RFXAP affected its nuclear import. Although
the possibility existed that the GFP-RFX-B fusion protein was not
functional, this was ruled out by the ability of the fusion protein to
complement an RFX-B-deficient B cell line (data not shown).
How, then, does RFX-B enter the nucleus? One possibility was
through binding to CIITA. Experiments to prove this point failed,
suggesting that although RFX-B and CIITA associate (20), this
association does not take place in the cytoplasm. RFX-B contains
several interacting surfaces, including four ankryin repeats, any of
which may function to piggyback RFX-B onto another molecule.
Thus, it is possible that RFX-B enters the nucleus in association
with another protein that contains an NLS. Likewise, the localization of RFX-B may consist of an export process as well; however,
this process would be independent of CRM-1, because RFX-B’s
localization was not affected by LMB in our study. Further studies
are required to characterize RFX-B translocation across the nuclear membrane.
The presence of import and export signals in RFXAP suggests
an interesting possibility for regulation of the RFX complex. As
RFXAP and RFX5 form a complex, these signals provide a mechanism for differential regulation of HLA isotypes. Microorganisms
such as Chlamydia have been recently shown to target RFX5 for
degradation, suggesting an important mode of down-regulating
MHC class II expression (39, 40). Likewise, cytokines such as
TGF-␤ and IL-10 that down-regulate MHC class II expression (9)
may use the RFX complex as a target. Modulation of the nuclearcytoplasmic shuttling of these factors provides an instant mechanism for effectors to down-regulate class II expression.
419