| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
B or PU.1 to Mediate IL-4-Induced Activation of IgE Germline Gene Transcription
Department of Immunology, Novartis Research Institute, Vienna, Austria
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
B/rel. The authenticity of PU.1 was demonstrated by
specific competition and supershifts in EMSA experiments. In addition,
in vitro translated PU.1 could interact with an oligonucleotide derived
from the IgE germline promoter containing the PU.1 binding site and
migrated with the same mobility compared with the complex formed with
nuclear extracts. Transient transfection experiments using IgE germline
promoter reporter gene constructs demonstrated that mutations affecting
DNA binding of PU.1 or NFB/rel had no or little
effect on IL-4 inducibility of these plasmids. However, point mutations
that abolished binding of both factors abrogated cytokine inducibility.
No strict spacing of the STAT6 and the composite PU.1/NF-B elements
is required for IL-4 induction. IL-4-induced STAT6 DNA binding was
retained in PU.1-/NFB/rel-
double mutants. The data demonstrate that cooperation of STAT6 with at
least PU.1 or NFB/rel is necessary for IL-4-induced
activation of IgE germline gene transcription. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
( switch) region into the C locus. It
is generally believed that the IL-4-induced transcriptional activity
directs class switching to IgE by specifically making the S region
accessible for the subsequent DNA recombination step. The second step
recombines the Ig heavy chain locus between S and another S region.
The intervening DNA is looped out and deleted. In the recombined genome
the C locus is located in close proximity to the
VHDJH gene segments,
allowing for the production and secretion of functional IgE. DNA
recombination is induced by direct B/T cell contact. One of the
structures involved in this cell communication is the CD40-CD40
ligand pair.
The molecular analysis of cis-acting regulatory DNA elements
and their corresponding binding factors has identified two distinct
IL-4 response elements in the human IgE-GLP sequence
(4, 5, 6) as well as in the murine counterpart (7, 8). These motifs are bound by transcription factors upon IL-4
treatment. The DNA sequence of one of the two motifs as well as the
characteristics of its binding factor has identified this protein as
STAT6 (9, 10). In addition to STAT6, other transcription
factors were identified that interact with their recognition sequence
in a constitutive, IL-4-independent fashion. The B cell-specific
transcription factor BSAP binds downstream of STAT6 close to a major
start site of transcription. Conflicting results were reported on its
involvement in the IL-4 response (11, 12, 13). Upstream of the
STAT6 site the importance of a cis-acting element binding
C/EBP family members was identified (4, 14, 15). Recently,
two binding sites for members of the NFB/rel
transcription factor family were identified (14, 16).
Factor binding to the proximal site (NF-B2) was shown to be
necessary but not sufficient for cytokine induction of IgE-GLP
activity, suggesting a functional cooperation between STAT6 and NF-B
proteins. In contrast, reporter plasmids carrying mutations in the
distal NF-B1 binding site which abrogated B protein interaction
responded to the cytokine stimulus as well as wild-type promoter
constructs (16).
This study describes the interaction of the transcription factor PU.1
with a binding site that overlaps with the distal NF-B1 binding
sequence. Functional studies suggest that loss of PU.1 or NF-B DNA
binding has no adverse effect on IgE-GLP activation upon IL-4
stimulation. However, mutations affecting binding of both factors
simultaneously led to abrogation of cytokine responsiveness, suggesting
the cooperation of at least one of these factors with STAT6 to activate
IgE germline gene transcription.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The human EBV-negative Burkitt’s lymphoma cell line DG75 was kept at 37°C with 5% CO2 in IMDM supplemented with 10% heat-inactivated FCS (Life Technologies, Grand Island, NY). Purified human rIL-4 was obtained from Novartis (Basel, Switzerland) with a specific activity of 0.5 U/ng.
Construction of plasmids
The promoterless plasmid pGL2 basic (Promega, Madison, WI) was
used for the preparation of all the constructs described. A fragment of
the human IgE germline promoter from position -148 to +90 (construct
LUCwt) was cloned in front of the luciferase gene as described
(4). Site-directed mutations in the NF-B1 site were
generated as reported earlier (17) using
oligonucleotides described below. Introduction of a 6-bp-long
nucleotide stretch between the STAT6 and the PU.1/NF-B sequences
(plasmid LUC sp) was achieved by the same technique (17)
using the oligonucleotides 5'-CCAAGAACAGAGAGAGAGAGAAAAGGGAACTTCCAGG-3'
and 5'-TCTCTCCTGTTCTTGGGAAGTCGATTGAGCAACAGCG-3'. The STAT6
expression vector was cloned by insertion of a
XbaI/KpnI fragment containing the complete human
STAT6 cDNA into the pcDNA3.1 vector digested with
NheI/KpnI.
Plasmids were analyzed by digestion with restriction endonucleases and DNA sequencing. Constructs used for transfection into DG75 cells were purified by cesium chloride density gradients.
Transient transfection and assays
Before transfection, DG75 cells were cultured in fresh medium for 24 h at 37°C, then harvested and washed twice in cold RPMI 1640 medium. Twenty micrograms of supercoiled plasmid DNA together with 1 µg of pRL-Tk (Promega) as internal control for normalization of differences in transfection efficiency were mixed with 107 cells in 300 µl cold RPMI 1640 medium. The cells were exposed to a single pulse at 1500 µF and 240 V using a Bio-Rad Gene Pulser (Bio-Rad, Richmond, CA). Immediately after transfection, 700 µl of warm culture medium was added and cells were diluted in 3 ml of complete culture medium. Aliquots were cultured for 8 h in the presence or absence of 25 U/ml human IL-4. Then luciferase assays were conducted in triplicates according to the instructions from the manufacturer using the Dual-Luciferase Reporter Assay System (Promega). The measured light units of the reporter plasmids were normalized to the activity of the pRL-Tk plasmid to correct for differences in transfection efficiency.
Preparation of nuclear extracts and EMSA
Nuclear extracts from DG75 cells were prepared as described
previously (18). Synthetic oligonucleotides were annealed
and end-labeled using [
-32P]dCTP (Amersham;
3000 Ci/mmol) and Klenow polymerase (Bethesda Research Laboratories,
Gaithersburg, MD). Binding reaction with 1 or 5 µg of nuclear
extracts was done according to the protocol described recently
(6). For competition experiments, extracts were
pre-incubated with a 100-fold molar excess, unless stated otherwise, of
competitor double-stranded oligonucleotides for 30 min on ice before
the radiolabeled probe was added. In supershift experiments, extracts
were incubated with 1–2 µg of Abs for 45 min on ice before adding
the labeled double-stranded oligonucleotide. Protein-DNA complexes were
resolved in a native 4% polyacrylamide gel in 1x TBE at 200 V for
1.5 h. Dried gels were exposed to an x-ray film at -70°C (Kodak
Xomat MR, Rochester, NY). The anti-PU.1 antiserum was a kind gift
of Dr. Richard Maki (La Jolla, CA). NF-B p50 (sc-114X), bcl-6
(sc-858X), STAT6 (sc-981), ets1/2 (sc-112X), and C/EBP (sc-746X) Abs
were purchased from Santa Cruz Biotechnology (Santa Cruz, CA).
Additional PU.1 Abs were obtained from PharMingen (San Diego, CA;
15501A) and from Santa Cruz Biotechnology (SC-352X). The following
oligonucleotides were used: 80/52wt,
5'-GAGAGAAAAGGGAACTTCCAGGGCGGCCC-3'; 80/52 m64-62,
5'-GAGAGAAAAGGGAACTAGTAGGGCGGCCC-3'; 80/52 m64,
5'-GAGAGAAAAGGGAACTACCAGGGCGGCCC-3'; 80/52 m64-63,
5'-GAGAGAAAAGGGAACTAGCAGGGCGGCCC-3'; 80/52 m70-69,
5'-GAGAGAAAAGTTAACTTCCAGGGCGGCCC-3'; 80/52 m70A,
5'-GAGAGAAAAGAGAACTTCCAGGGCGGCCC-3'; 80/52 m70T,
5'-GAGAGAAAAGTGAACTTCCAGGGCGGCCC-3'; 80/52 m68,
5'-GAGAGAAAAGGGCACTTCCAGGGCGGCCC-3'; 80/52 m71,
5'-GAGAGAAAACGGAACTTCCAGGGCGGCCC-3'; 80/52 m72,
5'-GAGAGAAACGGGAACTTCCAGGGCGGCCC-3'; 80/52 m73,
5'-GAGAGAACAGGGAACTTCCAGGGCGGCCC-3'; 80/52 m74-73,
5'-GAGAGATCAGGGAACTTCCAGGGCGGCCC-3'; 80/52 m77-75,
5'-GAGTCTAAAGGGAACTTCCAGGGCGGCCC-3'; 106/55,
5'-GTTGCTCAATCGACTTCCCAAGAACAGAGAGAAAAGGGAACTTCCAGGGCGG-3'; 106/55
m73,
5'-GTTGCTCAATCGACTTCCCAAGAACAGAGAGAACAGGGAACTTCCAGGGCGG-3';
106/55 m70-69,
5'-GTTGCTCAATCGACTTCCCAAGAACAGAGAGAAAAGTTAACTTCCAGGGCGG-3'; 106/55
m64-63,
5'-GTTGCTCAATCGACTTCCCAAGAACAGAGAGAAAAGGGAACTAGCAGGGCGG-3'. The
sequence of the 20/3 double-stranded oligonucleotide containing the
IgE-GLP BSAP site is described in Ref. 11 . Competitor
oligonucleotides containing PU.1 sites from the SV40 genome
(19) or from the Ig 3' enhancer (20) are:
SV40, 5'-GACCTCTGAAAGAGGAACTTGGTTAG-3', and 3',
5'-GATCCCTTTGAGGAACTGAAAACAGAA-3'. The sequence of the polyoma
virus enhancer c-ets1 binding site (21) oligonucleotide
was the following: 5'-GTCAGTTAAGCAGGAAGTGACT-3'. The competitor
oligonucleotide containing a NF-AT recognition sequence derived from
the TNF- promoter (22) had the sequence
5'-AGCTCATGGGTTTCTCCACCAA-3'.
In vitro transcription/translation
The human PU.1 cDNA (a kind gift of Dr. Richard Maki) was cloned into pcDNA3.1. The resulting PU.1 expression plasmid was used as a template for coupled in vitro transcription/translation reactions using the TnT system (Promega).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
B1 site is involved in
IL-4 responsiveness of the IgE-GLP
Our previous results have identified two binding sites for NF-B
protein factors (NF-B1, NF-B2) in the human IgE-GLP sequence that
are functionally distinct. NF-B proteins binding to the NF-B2
site cooperated with STAT6 in the IL-4 regulation of the IgE germline
gene. By contrast, factor binding to the NF-B1 site was not involved
in cytokine inducibility (16). To analyze the function of
the NF-B1 motif in more detail a substitution mutation was
introduced in the purine-rich region of the NF-B1 site (Fig. 1
A). Transient transfection
experiments into DG75 cells showed that this reporter construct
(LUCm70-69) could not respond to the cytokine stimulus anymore (Fig. 1B). The constitutive promoter activity was moderately lower
than that of the wild-type plasmid. In contrast, a plasmid carrying
mutations in the pyrimidine-rich region of the NF-B1 site
(LUCm64-62) showed increased IL-4 inducibility as well as elevated
constitutive promoter activity as reported earlier (16).
To evaluate the effect of these two different mutations on factor
binding to the NF-B1 site, EMSA were performed. Incubation of the
radiolabeled double-stranded oligonucleotide probe 80/52 with nuclear
extracts prepared from DG75 cells resulted in two retarded
nucleoprotein complexes, the upper one containing NF-B family
members p50 and p65 (Fig. 1C). The faster migrating complex
B is unrelated to NF-B (16). Competition with excess of
the mutant double-stranded oligonucleotides 80/52 m64-62 and 80/52
m70-69 demonstrated that both were unable to compete with NF-B
binding to the wild-type probe, indicating that both types of mutants
were defective to interact with NF-B proteins. In contrast to the
80/52 m64-62 oligonucleotide, the 80/52 m70-69 competitor could not
compete for complex B binding either (Fig. 1C). This showed
that the two point mutations in the purine-rich region of the NF-B1
site affected not only binding of p50/p65 but also DNA interaction of
an unknown factor in complex B. These results raised the possibility
that the failure of the LUCm70-69 reporter construct to respond to IL-4
stimulation was connected to the inability for complex B formation.
Therefore, experiments were conducted to identify the unknown protein
contained in complex B.
|
B1
site
The two guanosine residues mutated in the m70-69 construct are
part of the 5'-GGAA-3' core recognition structure important for DNA
binding of the ets-transcription factor family (23). To
test the possibility that ets factors are contained in complex B, the
80/52 wild-type binding pattern in EMSA was competed with an excess of
known binding sites for the family members c-ets-1 and PU.1. In
addition, a double-stranded oligonucleotide containing a NF-AT
recognition sequence was used. Fig. 2A shows that the three
competitors containing c-ets-1 or PU.1 binding sites competed
specifically for complex B formation but not for NF-B binding. The
NF-AT competitor was unable to compete for either complex. This
suggested that members of the ets transcription factor family could
give rise to complex B. To substantiate this finding nuclear extracts
were preincubated with Abs directed against PU.1 or other ets family
members before the radioactive probe was added (Fig. 2B).
Preincubation with the anti-PU.1 Ab led to disappearance of complex
B. This effect seemed to be specific because the ets1/2 Ab that broadly
cross-reacts with a number of different ets-family members had no
effect similar to the bcl-6 Ab which was used as control. Similar
results were obtained when the NF-B binding-defective 80/52 m64-62
double-stranded oligonucleotide was used as EMSA probe (Fig. 2B). As expected, no NF-B band was observed and only
complex B was detected. Similar to the wild-type probe, anti-PU.1
incubation resulted in diminished complex B formation. In addition, the
generation of a supershifted complex was observed. Again, no effects
were seen with the anti-ets1/2 and the bcl-6 Abs. These data
strongly suggested that PU.1 was recognizing a sequence motif between
position -80 and -52 in the human IgE-GLP. To confirm this assumption
in vitro-translated PU.1 protein was incubated with the radiolabeled
80/52 probe or with an authentic PU.1 site identified in the SV40
genome (19) (Fig. 2C). In vitro translated PU.1
comigrated with a complex formed upon incubation of the SV40-PU.1 site
with DG75 extracts. Moreover, authentic PU.1 formed a nucleoprotein
complex with the 80/52 double-stranded oligonucleotide. This complex
migrated with the same mobility as complex B in DG75 extracts. It
should be noted that the PU.1 supershifts almost completely abrogated
complex B formation under the experimental conditions used (see also
Fig. 5B). The remaining signal may be explained by binding
of additional, PU.1 unrelated, factors collectively giving rise to
complex B formation. However, we consider this alternative explanation
unlikely because in vitro translated PU.1 comigrated with complex
B.
|
|
PU.1 recognizes the ets core motif and upstream sequences
The nucleotides involved in DNA binding of PU.1 to its site in the
IgE-GLP were mapped by EMSA using a series of substitution mutations
between position -77 and -62 in the context of the 80/52
double-stranded oligonucleotide (Fig. 3).
Mutations located in the 5'-GGAA-3' PU.1 core motif abrogated PU.1
binding (m70-69, m70A, m70T, m68). These mutations also completely
prevented NF-B complex formation with the exception of the m68
mutant which bound NF-B factors but much weaker compared with the
wild-type probe. Interestingly, mutations upstream of the 5'-GGAA-3'
core motif and the NF-B decanucleotide sequence (m73, m74-73, and
m77-75) were also defective in PU.1 complex formation (Fig. 3). NF-B
binding was unaffected in these mutants. A sequence comparison of the
human and mouse IgE-GLP sequence showed that positions -76 to -78 in
the mouse sequence upstream of the 5'-GGAA-3' core motif differ from
the human counterpart (see Fig. 1
). To assess whether these differences
had an influence on PU.1 binding to the murine promoter sequence, EMSA
experiments were performed with the murine 80/52 homologue as a probe.
These experiments revealed that both NF-B and PU.1 could interact
with the mouse oligonucleotide very similar to the human 80/52 probe
(data not shown). The same results were obtained using a murine B cell
extract. These data showed that the 3-bp difference in the mouse
sequence still allowed for PU.1 binding despite the similar location of
the differences to the upstream important region defined by the m73,
m74-73, and m77–75 mutations.
|
B sequence (m64-62,
m64, m64-63) had no effect on PU.1 binding but prevented NF-B
complexes in m64-62 and m64-63 as expected. In these mutants, the PU.1
complex was much stronger compared with the wild-type probe. This may
be explained by a weaker specific activity of the 80/52wt probe
compared with the mutants. More likely, the absence of NF-B proteins
on the mutant sites allowed for better access of PU.1 (compare also
Fig. 5B). It is also possible that the mutations created a
PU.1 site with a higher binding affinity.
The single nucleotide exchange in m64 resulted in significantly weaker
NF-B binding compared with the wild-type sequence. Substitutions at
position -71 (m71) or -72 (m72) had no or little effect on PU.1
interaction. The m71 mutant probe could only weakly interact with
NF-B factors, whereas wild-type binding was observed with the m72
mutant, consistent with the known sequence requirements for NF-B.
These data demonstrated that PU.1 recognized not only the previously
described 5'-GGAA-3' (position -70 to -67) ets family motif but also
sequences upstream between position -77 and -73. Nucleotide positions
-71 and -72 or sequences downstream of the ets core motif appeared
not to be involved in PU.1 interaction. As control, a mutant containing
both the m77-75 and the m64-62 substitutions was tested. This probe was
unable to form a nucleoprotein complex similar to the m70-69, m70A, and
m70T mutations as expected (data not shown). No PU.1 binding to the
previously described NF-B2 site (16) containing an ets
factor core motif at position -35 to -38 was detectable (data not
shown).
Cooperation of PU.1 or NF-B with STAT6 in the IL-4 induction of
IgE-GLP activity.
To address the functional consequence of these mutants on the
IL-4-induced activation of the human IgE-GLP, mutant reporter gene
plasmids were constructed and tested for IL-4 inducibility in transient
transfection experiments. To increase the level of IL-4 inducibility in
these experiments a titrated amount of STAT6 expression vector was
co-transfected (24). It was observed earlier that, similar
to the observation made by Lu et al. (24), increasing
amounts of cotransfected STAT6 expression vector increased the IL-4
inducibility of the wild-type LUCwt IgE-GLP reporter DNA (Fig. 4A). Consistent with the data
shown in Fig. 1B, constructs in which only NF-B binding
was abolished (m64-62, m64-63) were consistently more responsive to the
cytokine stimulus than the wild-type plasmid (Fig. 4B).
Interestingly, these plasmids could bind PU.1 more effectively in the
gel shift assays (see Fig. 3). This raises the possibility that the
extent of PU.1 binding correlates directly with cytokine
inducibility.
|
B was abrogated (m70-69, m70A, m70T)
did not respond with increased luciferase expression upon IL-4
stimulation. In general, mutant plasmids in which both or either one of
the two factors could interact with DNA at levels comparable to the
wild-type sequence were cytokine inducible very similar to the control
plasmid. Only reporter constructs in which both factors were absent or
strongly reduced had impaired IL-4 inducibility. These data
demonstrated that loss of DNA binding of one of the two factors was
dispensable for IL-4 inducibility of the human IgE-GLP. However, the
presence of either PU.1 or NF-B was necessary to mediate activation
of the human IgE-GLP upon STAT6-DNA interaction. Thus, the composite
PU.1/NF-B site is necessary but not sufficient for IL-4 induction of
the IgE germline gene.
STAT6 can bind its target sequence in the absence of PU.1 and
NF-B
The composite PU.1/NF-B binding site is located in close
proximity to the STAT6 responsive element (Fig. 1A)
(14, 16). Because IL-4-induced STAT6 binding is a
prerequisite for IgE-GLP activation the question was raised whether the
PU.1/NF-B defective phenotype was due to disturbed STAT6 DNA
interaction. To address this question, EMSA experiments were conducted
using a double-stranded oligonucleotide probe spanning position -106
to -55 of the human IgE-GLP sequence. This probe contains the STAT6,
the composite PU.1/NF-B site, and a C/EBP binding site that was
shown to be involved in the IL-4 response of the IgE-GLP (4, 14, 15). Incubation with the 106/55wt probe with uninduced nuclear
extracts prepared from DG75 cells resulted in two specific retarded
complexes similar to the ones observed with the 80/52wt probe.
Incubation with IL-4 induced extracts resulted in two weak additional
nucleoprotein complexes which migrated above the major predominant
constitutive band (Fig. 5A).
Preincubation with anti-STAT6 Ab specifically abolished both of the
IL-4-induced complexes suggesting that STAT6 was involved in this
interaction. The anti-p50 Ab led to disappearance of the slowest
migrating IL-4-induced band and to a reduction of the major
constitutive complex. This result raised the possibility that the
predominant constitutive band contained NF-B p50 and that the
largest IL-4-induced complex contained both STAT6 and p50. When the
anti-PU.1 antiserum was preincubated with the extracts the fast
migrating constitutive band disappeared and a supershifted complex was
observed slightly above the predominant band. These data demonstrated
that the smallest complex contained PU.1. Anti-C/EBP Ab incubation had
no effect. These data were strengthened by the results obtained
when the 106/55 m64-63 NF-B-defective double-stranded
oligonucleotide was used as radiolabeled probe (Fig. 5B).
Incubation with uninduced extracts gave rise to PU.1 complex formation
but not NF-B, similar to the situation observed with the 80/52
NF-B mutations. Incubation with IL-4-induced extracts generated in
addition a larger complex that comigrated with the lower cytokine
induced band observed with the wild-type probe. Similar to the 106/55wt
probe this complex could be inhibited with anti-STAT6 Ab and a
slower migrating supershift was formed. This suggested that this
complex contains STAT6 but not NF-B family members. This was
confirmed by anti-p50 Ab supershift that did not change the binding
pattern. In contrast, preincubation with the anti-PU.1-specific mAb
resulted in almost complete disappearance of the PU.1 band and the
generation of a supershifted complex just below the STAT6 band. These
results were confirmed in competition experiments (Fig. 5C).
Competition with increasing amounts of unlabeled 80/52wt
double-stranded oligonucleotide led to loss of all complexes except the
band which contained STAT6 only. The 80/52 m64-62 NF-B-defective
double-stranded oligonucleotide competed specifically for PU.1 complex
formation, whereas the 80/52 m70-69
NF-B-/PU.1- competitor
was completely unable to interfere with factor binding to the 106/55wt
probe. The 20/3 control competitor gave identical results. These data
demonstrated that NF-B p50/p65 and PU.1 could constitutively
interact with the 106/55wt sequence similar to the shorter 80/52 probe.
Upon IL-4 stimulation two additional complexes appeared, one containing
STAT6 only, the other one being composed of STAT6 and NF-B. These
data support published results in the murine system
(25).
To directly correlate the effect of individual NF-B and PU.1
mutations with the protein interaction pattern observed on the 106/55wt
DNA, selected point mutations were introduced into the 106/55 sequence
and used as EMSA probes with uninduced or IL-4-treated DG75 extracts
(Fig. 6). The results obtained with
uninduced extracts mirrored exactly the data generated with the
mutations in the 80/52 sequence background (compare with Fig. 3). Probe
106/55 m64-63 formed a PU.1 nucleoprotein complex but was unable to
bind NF-B factors demonstrating that PU.1 could bind in the absence
of p50/p65. It has been shown that DNA bound PU.1 enables interaction
of the transcriptional activator Pip (NF-EM5) with a sequence next to
the PU.1 binding site in the Ig-3' enhancer (19, 26, 27). Despite the presence of Pip in the DG75 extracts as shown
by EMSA using the Ig-3' composite PU.1/Pip double-stranded
oligonucleotide as probe (data not shown) only the PU.1 complex was
observed with the m64-63 NF-B-defective probe. In addition,
competition with the PU.1/Pip oligonucleotide abrogated specifically
only PU.1 binding (Fig. 2A). This suggested that Pip could
not interact with the DNA region surrounding the PU.1 site. The 106/55
m73 EMSA probe was defective to interact with PU.1 but had wild-type
NF-B binding capacity. The 106/55 m70-69 probe could not form any
nucleoprotein complex. With IL-4-treated extracts the m64-63 mutation
abolished NF-B and STAT6/NF-B complex formation but not STAT6
binding by itself. Similarly, abrogation of PU.1 (m73) did not change
the interaction pattern of the other three complexes. STAT6 was able to
bind its target sequence even in the absence of PU.1 and NF-B
(m70-69), demonstrating that the absence of one or both of these
closely spaced factors does not qualitatively alter the interaction of
STAT6 with DNA. The lower intensity of the STAT6 band observed with the
m70-69 probe was not reproducible. The results show that the failure of
the PU.1-/NF-B-
reporter constructs to respond to IL-4 stimulation was not due to
defective STAT6 binding.
|
B for
IL-4 induction
The composite PU.1/NF-B binding site is separated from the
STAT6 recognition sequence by exactly two helix turns (Fig. 1A). To determine, whether the apparent cooperation of these
factors is dependent on their relative positioning a 6-nt-long
5'-AGAGAG-3' spacer was introduced into the LUCwt reporter construct
between positions -80 and -79, a position outside of the mapped
sequence recognition requirements of STAT6 and PU.1/NF-B. In the
resulting plasmid (LUCspwt) the STAT6 site is two and a half helix
turns separated from the PU.1/NF-B sequence. Transient transfection
experiments in DG75 cells were performed to monitor the effect of this
change on IL-4 responsiveness. The results shown in Fig. 7 demonstrate that the degree of IL-4
inducibility of the LUCspwt plasmid was higher than that of the
wild-type LUCwt construct in all three experiments. In addition, no
effect on the constitutive promoter activity was observed. The 6-bp
insertion repositions the PU.1 site to the original NF-B location
relative to the STAT6 site. To rule out the possibility that the new
PU.1 location masked a possible effect of the inserted linker a plasmid
was generated that carried the m77-75 (PU.1 defective) nucleotide
substitutions in addition to the 6-bp spacer (LUCspm77-75). Although
the constitutive promoter activity was lower than that of the LUCwt
construct it responded comparably to the IL-4 stimulation. As control,
a NF-B-defective m64-63 mutation was tested in the presence of the
spacer (LUCspm64-63). Also this construct was indistinguishable from
the wild-type plasmid. The reason for the hyperinducible phenotype of
the LUCspwt plasmid is unclear at present although the IL-4 induction
indices of the mutant plasmids suggest an involvement of NF-B and
PU.1. From these data one can conclude that there is no strict spacing
requirement of the STAT6 and PU.1/NF-B binding sequences for their
functional interaction.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
B and C/EBP. In particular, a recognition site for members of the
NF-B/rel factor family, NF-B2, was described to act in
concert with STAT6 (16). A second binding site for B
proteins (NF-B1) appeared not to have this function in the human
promoter in contrast to a recent study in the mouse system
(25). The present study demonstrates that the NF-B1
site overlaps with a binding site for the transcription factor PU.1.
Abrogation of DNA binding of both PU.1 as well as NF-B resulted in
loss of IL-4 inducibility of IgE germline promoter constructs. However,
plasmids in which only one of the factors was missing responded to the
cytokine stimulus similar to the wild-type sequence. These data may
explain the divergent results regarding the function of the NF-B1
element obtained in the two species (16, 25). The sequence
comprising the NF-B1 site is highly conserved between mouse and
human, suggesting that PU.1 may also bind to the mouse promoter and
exert a similar function. In fact, we could demonstrate interaction of
PU.1 with the murine IgE-GLP sequence (data not shown). In such a
scenario the linker scanning mutant reporter construct used for the
identification of NF-B1 function in the mouse system
(25) most likely affected binding of both B factors and
PU.1, resulting in a noninducible phenotype similar to our
data.
Reporter constructs in which interaction of PU.1 or NF-B with DNA
was abrogated by substitution mutations responded to IL-4 stimulation
indistinguishable to the wild-type promoter. However, point mutations
that affected binding of both factors simultaneously led to a cytokine
unresponsive phenotype even in the presence of the functionally
important C/EBP site. This result suggested that PU.1 as well as
NF-B factors could cooperate functionally with STAT6 to mediate IL-4
induction of the IgE germline promoter. Our previous data have shown
that two copies of the C/EBP-STAT6 portion of the IgE-GLP were not
sufficient to confer IL-4 responsiveness to a heterologous minimal
promoter. Only the additional presence of two copies of the
PU.1-NF-B unit led to activation by the cytokine (16).
We conclude from these data that the PU.1-NF-B1 unit plays a key
role in defining a minimal IL-4 response element together with STAT6,
whereas C/EBPß exerts an amplifying function (4).
It is well established that PU.1 interacts with a 5'-GGAA-3' core
sequence motif (23). A panel of substitution mutants in
and around the NF-B1 motif confirmed the importance of this
sequence, but nucleotides upstream of the core sequence were also
necessary for PU.1 DNA interaction. Similar results have been described
for the PU.1 site in the CD11c promoter (33).
Interestingly, no binding of PU.1 was detectable on the NF-B2 motif
despite the presence of a 5'-GGAA-3' core sequence. This underscores
the importance of surrounding nucleotides in determining specificity of
factor binding. In B lymphocytes, PU.1 regulates the expression of a
number of genes by virtue of a binding site in the respective promoter
sequence, such as CD20, CD72, both the - and the 
-light chain
genes; mb-1, IgH µB and µA enhancer elements; IgV19 light chain;
IgJ and PU.1 itself (23, 34, 35). Similar to the IgE
germline gene most of these genes have no "TATA" box. In some
cases, the PU.1 site is located close to a transcription initiation
site (33, 36). Interestingly, an ets protein was proposed
to be involved in a LPS-inducible nucleoprotein complex on a major
transcription initiation site of the murine IgG1 germline promoter
(37). This suggests that the relative position of the PU.1
binding site is important for its regulatory function as a "bridging
factor" between upstream regulatory elements and the basal
transcription machinery (36). In the case of the IgE
germline promoter the PU.1/NF-B site maps significantly upstream of
the transcription start sites (38). In addition, PU.1
promoter mutants had wild-type basal promoter activity as well as IL-4
inducibility. This suggests that the regulatory role of PU.1 may be
different for this gene.
There are a number of reports describing interactions of PU.1 with
other transcription factors. For example, PU.1 interacts with Pip
(NF-EM5, LSIRF, ICSAT, or IRF4) to transactivate the Ig and
enhancer elements (19, 26). Our present data do not
support PU.1 dependent interaction of Pip with the IgE germline
promoter sequence (data not shown). PU.1 has also been shown to
associate in vitro with Rb and TFIID (39), Jun family
members (40), NF-IL6ß, HMG proteins, MKP1, and hsp90
(41). At present, it is unclear whether PU.1 physically
contacts NF-B proteins and/or STAT6 or C/EBPß to exert its
function. A recent study described physical interaction of STAT6 with
NF-B p50 and p65 (42). Given the fact that the NF-B2
sequence also is important for mediating cytokine inducibility
(16), it is possible that PU.1 is involved in the
formation of a higher order nucleoprotein complex containing STAT6,
C/EBPß, and NF-B proteins binding at the two B motifs and PU.1.
In such a structure PU.1 may stabilize the interaction of STAT6 with
NF-B factors in the complex by simultaneous interaction with these
proteins. It is also possible that PU.1 further contributes to the
stabilizing effect of C/EBPß on DNA bound STAT6 (15).
Alternatively, PU.1 may contribute to the formation of a
transcriptional competent complex by DNA bending, a feature also known
for NF-B factors (43, 44). Besides a role in higher
order complex assembly, a second role of PU.1 can be envisioned. Once a
transcriptional competent complex has been formed upon IL-4 induction
which by itself induces transcription, the transacting potential of
PU.1 and NF-B may cooperate in the synergistic up-regulation
mediated through the CD40 receptor (25, 45, 46).
Expression of PU.1 is restricted to B cells and monocytes. Thus, PU.1
may contribute to cell type specific expression of the IgE germline
gene together with the B cell-specific factor BSAP
(11, 12, 13).
Interestingly, the architecture of the HIV-1 long-terminal repeat
promoter/enhancer is reminiscent to the one of the IgE germline
promoter (47). There, two tandemly arranged NF-B
binding structures overlap with PU.1 recognition sequences very similar
to the situation described in this report. Both type of factors are
required for inducible expression of viral genes, suggesting a
functional cooperation not only of the binding proteins but also of the
two NF-B binding sites again suggesting the existence of a complex
nucleoprotein structure. In this example as well as in others
(27), PU.1 function was dependent on serine
phosphorylation at position 148 (26, 27). It remains to be
seen whether this posttranslational modification is also required for
PU.1 function in IL-4 induced IgE germline promoter activation.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Abbreviations used in this paper: IgE-GLP, IgE germline promoter; ets, E26-transformation specific; C/EBP, CCAAT enhancer binding protein.
Received for publication October 14, 1998. Accepted for publication August 2, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
germline transcript: identification of a novel IL-4 responsive element. J. Immunol. 150:5408.[Abstract]
and Fc receptor II (CD23b) genes activated via binding of an interleukin-4-induced transcription factor to a novel responsive element. Eur. J. Immunol. 23:3066.[Medline]
transcripts. Mol. Cell. Biol. 11:5551.
activation site-like sequence upstream of the I exon in a human B cell line: evidence for the involvement of Janus kinase 3 and interleukin-4 Stat. J. Clin. Invest. 96:907.
transcription. J. Immunol. 152:2904.[Abstract]
germline promoter. J. Immunol. 158:5874.[Abstract]
promoter: regulation by NF-IL-4, a C/EBP family member and NF-B/p50. J. Exp. Med. 181:181. promoter mediated by Stat6 and C/EBPß. J. Immunol. 161:1822.B/rel in the IL-4-induced up-regulation of the human IgE germline promoter. J. Immunol. 159:3330.[Abstract]
3' enhancer activity. Mol. Cell. Biol. 12:368.2-4 enhancer. Mol. Cell. Biol. 13:6452. promoter in the murine dendritic cell line 18 and the murine mast cell line CPII is differently regulated. J. Immunol. 157:2645.[Abstract]
germline transcripts in B cells via activation of NF-B. J. Immunol. 158:4769.[Abstract]
promoter. Int. Immunol. 7:1529. responsiveness and basal expression of the myeloid human FcR1b gene is mediated by a functional PU.1 site and a transcription initiator consensus. J. Exp. Med. 179:1985. transcripts in human B cells induced by interleukin 4 to switch to IgE production. J. Exp. Med. 172:463.
, CRP3). J. Immunol. 155:4330.[Abstract]
B: Direct association and synergistic activation of interleukin-4-induced transcription. Mol. Cell. Biol. 18:3395.B transcription factor induces DNA bending which is modulated by its 65-kD subunit. Nucleic Acids Res. 18:6497.B. J. Immunol. 153:4357.[Abstract]
RNA synthesis in human B cells by tumor necrosis factor-, anti-CD40 monoclonal antibodies or transforming growth factor-ß correlates with levels of IgE production. Int. Immunol. 4:397.B mediate lipopolysaccharide-induced HIV-1 long terminal repeat transcription in macrophages. J. Immunol. 161:268.This article has been cited by other articles:
|
|
 |
|
  M. Schedel, L. A. Pinto, B. Schaub, P. Rosenstiel, D. Cherkasov, L. Cameron, N. Klopp, T. Illig, C. Vogelberg, S. K. Weiland, et al. IRF-1 Gene Variations Influence IgE Regulation and Atopy Am. J. Respir. Crit. Care Med., March 15, 2008; 177(6): 613 - 621. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. T. Thieu, E. T. Nguyen, B. P. McCarthy, H. A. Bruns, R. Kapur, C.-H. Chang, and M. H. Kaplan IL-4-stimulated NF-{kappa}B activity is required for Stat6 DNA binding J. Leukoc. Biol., August 1, 2007; 82(2): 370 - 379. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Novak, A. Muller, N. Harrer, C. Gunther, J. M. Carballido, and M. Woisetschlager CCL23 Expression Is Induced by IL-4 in a STAT6-Dependent Fashion J. Immunol., April 1, 2007; 178(7): 4335 - 4341. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Valineva, J. Yang, and O. Silvennoinen Characterization of RNA helicase A as component of STAT6-dependent enhanceosome Nucleic Acids Res., September 1, 2006; 34(14): 3938 - 3946. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Polli, A. Dakic, A. Light, L. Wu, D. M. Tarlinton, and S. L. Nutt The development of functional B lymphocytes in conditional PU.1 knock-out mice Blood, September 15, 2005; 106(6): 2083 - 2090. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Valineva, J. Yang, R. Palovuori, and O. Silvennoinen The Transcriptional Co-activator Protein p100 Recruits Histone Acetyltransferase Activity to STAT6 and Mediates Interaction between the CREB-binding Protein and STAT6 J. Biol. Chem., April 15, 2005; 280(15): 14989 - 14996. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. T. Berton, L. A. Linehan, K. R. Wick, and W. A. Dunnick NF-{kappa}B elements associated with the Stat6 site in the germline {gamma}1 immunoglobulin promoter are not necessary for the transcriptional response to CD40 ligand Int. Immunol., December 1, 2004; 16(12): 1741 - 1749. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Weidinger, N Klopp, S Wagenpfeil, L Rummler, M Schedel, M Kabesch, T Schafer, U Darsow, T Jakob, H Behrendt, et al. Association of a STAT 6 haplotype with elevated serum IgE levels in a population based cohort of white adults J. Med. Genet., September 1, 2004; 41(9): 658 - 663. [Abstract] [Full Text] [PDF]
|
|
, Promoter activity in the
absence of IL-4; 
, promoter activity in the presence of the
cytokine. The numbers give the ratio of reporter gene expression in
IL-4-treated vs uninduced cells (IL-4 induction index). Shown is one
representative out of five independent experiments. Error bars indicate
the SD of triplicate measurements. C, EMSA using the
double-stranded oligonucleotide 80/52 as radiolabeled probe. Competitor
oligonucleotides were used in 100-fold molar excess.
