| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Graduate Group of Molecular Biology and
Laboratories of Biochemistry, Department of Animal Biology, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, PA 19104
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
light chain gene is controlled in part
by the 3' enhancer. Two of the proteins that bind to the 3'
enhancer, PU.1 and Pip, show tissue-restricted expression and may be
responsible for the tissue specificity of 3' enhancer activity. PU.1
alone can bind to DNA; however, Pip cannot bind to its 3' enhancer site
in electrophoretic mobility shift assays, unless recruited by PU.1.
Previously, we showed that the PU.1 PEST domain (rich in the amino
acids proline, glutamate, serine, and threonine; sequences 118–160) is
necessary for Pip recruitment to DNA. Here we used detailed mutagenic
analyzes of PU.1 to more precisely identify sequences required for Pip
recruitment by electrophoretic mobility shift assay. We found that
mutation of three segments within the PU.1 PEST domain (118–125,
133–139, and 141–147) modulated the efficiency of Pip recruitment,
while mutation of sequences between residues 88–118 and 154–168 had
no effect. Interestingly, we found that the PU.1 ETS domain (residues
170 to 255) is both necessary and sufficient for Pip interaction in
solution and that other ETS domain proteins can physically interact
with Pip as well. Our results suggest that Pip recruitment to DNA by
PU.1 occurs via a two-step mechanism. First, a physical interaction
that is not sufficient to recruit Pip occurs via the PU.1 ETS domain.
Second, a conformational change in the PU.1 PEST domain, apparently
mediated by serine phosphorylation, induces a conformational change in
Pip enabling it to bind to DNA. We also show that the PU.1 PEST domain
does not target PU.1 for rapid turnover. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
light chain
gene is controlled by three DNA elements: the promoter, the intron
enhancer (E), and a second enhancer located 8.5 kb downstream of the
C exon, the 3' enhancer (E3'). The intron enhancer is activated
primarily by transcription factor NF-B (1). Activity of the 3'
enhancer, however, does not require NF-B and is therefore controlled
by a distinct mechanism (2, 3). A number of proteins are known to bind
to the 3' enhancer and control its activity. These proteins include
Sp1, c-fos, c-jun, ATF1, CREM, PU.1, NF-EM5 (Pip), E2A, and YY1 (4, 5, 6, 7, 8).
Of these enhancer-binding proteins, PU.1 and Pip are tissue
restricted.
PU.1, the cellular homologue of the viral oncogene Spi-1, is a
272-amino acid-long, erythroid-, macrophage-, and B cell-specific
protein related to the ets family of transcription factors
(9, 10). PU.1 can control expression of a number of genes, including
Ig, Ig
, Ig J chain, c-fms, IL-1ß, and
granulocyte-macrophage CSF receptor 
(11, 12, 13, 14, 15, 16, 17). PU.1 overexpression
can immortalize erythroblasts and appears to play an important role in
the genesis of erythroleukemia (10, 18, 19, 20, 21). PU.1 is a crucial factor
in hemopoetic development because homozygous knockout of the
PU.1 gene results in a loss of B cells, T cells,
granulocytes, and monocytes (22). In addition, oligonucleotides that
contain the PU.1 DNA-binding site can inhibit hemopoetic colony
formation in vitro (23). In the macrophage lineage, PU.1 appears to be
necessary for terminal development (24).
Several structural features of the PU.1 protein have been identified.
The amino terminus of the protein contains the transcriptional
activation domain, which includes both acidic and glutamine-rich
domains (25). A PEST3 domain
(rich in the amino acids proline, glutamate, serine, and threonine) is
located between amino acids 120 and 160, and the DNA-binding ETS domain
is located between amino acids 161 and 260 (9). Although PU.1 is
capable of binding to its cognate DNA site by itself, we previously
identified a nuclear factor, NF-EM5, that is incapable of binding to
the Ig 3' enhancer unless it is recruited by PU.1 (6, 17). The PU.1
and NF-EM5 DNA-binding sites lie adjacent to one another within the 3'
enhancer. A similar arrangement of PU.1 and NF-EM5 binding sites occurs
within the Ig2–4 enhancer (17). PU.1 sequences
necessary for recruitment of NF-EM5 to DNA when assayed by
electrophoretic mobility shift assay (EMSA) include amino acids
118–160 (the PEST domain; 6 . In addition, recruitment is
critically dependent upon phosphorylation of serine 148 in the PU.1
PEST domain (16).
Recently, a protein with properties strongly matching those of NF-EM5
was cloned (26, 27, 28). This protein, named Pip (variously called LSIRF,
ICSAT, or IRF4), is related to the IFN-regulatory factors (IRFs) and is
a lymphoid-restricted transcription factor. Pip function is critical
for proper immune function because disruption of the Pip gene by
homologous recombination causes a deficiency in B cell Ag response and
in cytotoxic and antitumor T cell activity (29). Interestingly, Pip can
bind, in the absence of PU.1, to the IFN-stimulated response element
(ISRE (27, 28)). However, full length Pip is unable to bind to its
DNA-binding site in the 3' and enhancers when assayed by EMSA
(26). As expected, incubation in the presence of PU.1 results in
recruitment of full length Pip to these DNA-binding sites. Pip
C-terminal sequences appear to mask the Pip DNA-binding domain, thus
inhibiting its ability to bind to and enhancer sequences (30).
Pip interaction with PU.1 may therefore result in a conformational
change in Pip such that its DNA-binding domain becomes unmasked (30).
However, the precise mechanism by which Pip is recruited to DNA by PU.1
remains unknown.
In this report, we have used mutagenesis to better define the region of
PU.1 required for recruitment of Pip to its 3' enhancer-binding site
by EMSA. We identified several regions within the PEST domain that
either reduce or abolish Pip recruitment when deleted or mutated to
alanine residues. Surprisingly, the PU.1 ETS domain, but not the PEST
domain, is required for intermolecular interaction between these two
proteins in solution. Taken together, our results suggest a two-step
mechanism for the recruitment of Pip to DNA by PU.1. This model
proposes that physical interaction occurs between PU.1 and Pip via the
PU.1 ETS domain. Subsequently, a conformational change within the PU.1
PEST domain, apparently mediated by serine phosphorylation, induces a
conformational change in Pip, thereby exposing the Pip DNA-binding
domain.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Polyclonal rabbit-anti-Pip antisera (Cocalico Biochemicals, Reamstown, PA) was prepared against polyhistidine-tagged Pip protein purified from Erichia coli BL21 cells using Ni2+-chelate chromatography (Qiagen, Santa Clarita, CA).
Construction of plasmids
The Pip cDNA-containing plasmid pPip/ATG IVT (a gift of Dr.
Harinder Singh, University of Chicago, Chicago, IL (26)) was digested
with HinDIII-EcoRI. The purified
HinDIII-EcoRI fragment was blunted with Klenow
polymerase and ligated into XmaI-digested and blunted
pGEX-2TK (Pharmacia, Piscataway, NJ) to produce plasmid pGEX-Pip.
Plasmid pCMV+-PU.1
PEST was created by ligation of the
EcoRI insert from plasmid pKS+-PU.1PEST into
EcoRI-digested pCMV+.
The PU.1 mutants 7–30, 33–100, 1–117, 1–160,
119–160, and 245–272 were previously described (6). The PU.1
mutant 201–272 (KpnI) was constructed by excising
a 0.6 kbp KpnI fragment from plasmid 25.1, containing
sequences from the pKS+ polylinker and PU.1 cDNA sequences
encoding amino acids 1 to 200, and ligation into
KpnI-digested pKS+. The mutant 256–272 was
constructed by ligation of a 
0.9 kb EcoRI-BglI
fragment from the plasmid 25.1 into SmaI-digested
pKS+. The Ets-1 cDNA was provided by Dr. Barbara Graves
(University of Utah, Salt Lake City, UT), the Fli-1 cDNA was provided
by Dr. Rich Maki (La Jolla Cancer Research Foundation, La Jolla, CA),
and the Ets-2 cDNA was provided by Dr. Narayan Avadhani (University of
Pennsylvania, Philadelphia, PA).
PU.1 deletion mutations were generated by PCR amplification of regions of the PU.1 cDNA clone 25.1 (9) that flank the desired deletion, followed by digestion with appropriate enzymes and ligation of the two pieces into pKS+ Bluescript (Stratagene, La Jolla, CA). The PCR reactions were conducted using 30 ng of template (25.1), 300 ng of each primer, 50 mM each of dATP, dCTP, dGTP, and dTTP, 1x PCR buffer containing 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 1.5 mM MgCl2, 0.001% (w/v) gelatin (Perkin-Elmer/Cetus, Norwalk, CT), 2.5 U of AmpliTaq polymerase (Perkin-Elmer/Cetus), and 2.5 U of Taq Extender (Stratagene). The amplification cycle consisted of 35 cycles of denaturation for 1 min (95°C), annealing for 30 s (50°C), and polymerization for 2 min (72°C), followed by 1 cycle of polymerization for 4 min (72°C).
The T3 primer (AATTAACCCTCACTAAAGGG, Stratagene) was used in tandem
with the reverse primers in Table I
to
amplify the 5' portion of PU.1, while the primer PU.1 3'UT Rev
(GCGTCTAGACGGTCTCTGCGGGCGATCAGTGGGG) was used in tandem with the
forward primers in Table I to amplify the 3' portion of PU.1. The PU.1
mutants, 154–156, 161–164, 165–168, and the alanine
substitution mutants were made using the "overlap extension PCR"
method (31). All PU.1 mutants generated by PCR were transcribed in
vitro using T3 RNA polymerase. The oligonucleotides used to generate
PU.1 mutants are shown in Table I.
|
Nuclear extracts were prepared from S194 plasmacytoma cells
essentially as described (32), and binding reactions were performed as
previously described (6). The sequence of the PU.1:Pip-binding site
from the 3' enhancer used as a probe in these assays is
CTTTGAGGAACTGAAAACAGAACCT (Oligo 5). Quantification of EMSA data
was conducted by PhosphorImager analysis (Molecular Dynamics,
Sunnyvale, CA). For determination of off-rates, binding reactions were
conducted until PU.1:Pip:DNA complex formation reached equilibrium (20
min) followed by addition of a large excess of cold competitor
oligonucleotide. Aliquots were removed at various time points and
loaded onto a nondenaturing gel. After electrophoresis, the fraction of
probe retained in the free DNA, PU.1:DNA, and PU.1:Pip:DNA complexes
were quantified using PhosphorImager analysis. The percentage of probe
in the PU.1:Pip:DNA complex at a given time point (i.e., the
percent of counts in the complex based upon the total number of counts
in all three complexes) was normalized based upon a value of 100% for
no competitor added.
Protease sensitivity studies
Conformational differences between selected PU.1 mutants were studied by partial proteolytic digestion of PU.1:DNA complexes that had formed in standard EMSA reactions, according to the method of Lodie et al. (33). After the reactions had reached equilibrium, 1 to 2 µl of solution containing either trypsin or proteinase K was added, and the reactions were allowed to proceed for 5 min at room temperature. The reactions were then stopped on ice and subjected to electrophoresis under standard EMSA conditions.
Glutathione S-transferase (GST) chromatography experiments
GST-fusion proteins were prepared essentially as described (34). Equivalent TCA-precipitable counts of [35S]-labeled rabbit reticulocyte lysate translated proteins (Promega, Madison, WI) or metabolically labeled proteins (see below) were incubated with approximately equivalent amounts (as judged by Coomasie blue staining) of GST or GST-fusion proteins bound to glutathione-agarose beads for 1 h at 4°C in NETN (100 mM NaCl, 1 mM EDTA, 20 mM Tris, pH 8.0, 0.5% Nonidet P-40). The beads were washed four to five times in NETN, and bound proteins were eluted in 1x SDS loading dye and resolved on 10% SDS polyacrylamide gels. When oligonucleotides were included in the assay, 500 ng of oligonucleotide was added. The following oligonucleotides were used: AGCAACTGTCATAGCTACCGTCACA (nonspecific (NS)), CTTTGAGGAACTGAAAACAGAACCT (PU.1:Pip).
Metabolic labeling and immunoprecipitation experiments
For metabolic labeling, a variation of the method of Luscher and
Eisenman was used (35). Briefly, 3T3 fibroblasts were transfected via
the CaPO4 precipitation method (36) with 10 µg of
either plasmid CMV-PU.1 or CMV-PU.1PEST, or both. Twenty-four hours
later, cells were pulsed with 0.2 mCi/ml [35S]
protein-labeling mix (Expre[35S][35S];
Dupont-NEN, Boston, MA) for 2 h at 37°C followed by a chase with
cold methionine (0.5 mM). At the specified time points, the cells were
harvested in Ab buffer (20 mM Tris, pH 7.5, 50 mM NaCl, 0.5% Nonidet
P-40, 0.5% SDS, 0.5% deoxycholate, 1 mM DTT, 10 µg/ml leupeptin, 10
µg/ml N--p-tosyl-L-arginine methyl ester, 1
µg/ml pepstatin, 1 mM PMSF), sonicated, and clarified by
centrifugation. Approximately 1.5 x 106
TCA-precipitable counts of each lysate was used in precipitations with
excess anti-PU.1 Ab (Santa Cruz Biotechnology, Santa Cruz, CA), and the
immune complexes were recovered using goat anti-rabbit IgG-coupled
agarose beads. The immune complexes were eluted by addition of SDS
loading dye, resolved on SDS polyacrylamide gels, and visualized by
autoradiography.
For preparation of metabolically labeled lysates used in GST-pull down
experiments, 3T3 fibroblasts were transfected with 20 µg of the PU.1
or PU.1PEST-expressing plasmids and labeled essentially as described
above. Cells were harvested in hypotonic buffer (25 mM Tris-HCl, pH
7.4, 1 mM MgCl2, 5 mM KCl) and incubated on ice for 5 min.
An equal volume of hypotonic buffer containing 1% Nonidet P-40 was
added, and incubation was continued for an additional 5 min. The nuclei
were pelleted and resuspended in RIPA buffer (150 mM NaCl, 1% Nonidet
P-40, 0.5% DOC, 0.1% SDS, 50 mM Tris, pH 8.0). The cell suspension
was sonicated and clarified by centrifugation.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Brass et al. (30) recently showed that Pip is responsible for
NF-EM5 activity in J558L cells. To determine whether the NF-EM5
activity present in S194 plasmacytoma cell nuclear extracts is also due
to Pip, we performed EMSA studies with anti-Pip antisera. The
PU.1:Pip DNA-binding sequence from the Ig 3' enhancer was incubated
with S194 nuclear extract and either preimmune sera or anti-Pip
sera. The PU.1:Pip:DNA complex was completely abolished by inclusion of
a polyclonal Pip antiserum, while preimmune serum had no effect (Fig. 1). Therefore, the Pip protein present in
S194 cells is identical, or highly homologous, to recombinant Pip.
Because we obtained more efficient DNA binding with proteins isolated
from nuclear extracts, the remainder of our studies were performed with
Pip derived from S194 cells.
|
We previously showed that PU.1 sequences 118–160 (the PEST
domain) and phosphorylation of serine 148 are necessary for Pip
recruitment to DNA by EMSA (6, 16). In addition, mutation of PU.1
serines 132 and 133 to alanine residues reduced, but did not abolish,
Pip recruitment (16). To better characterize the PU.1 PEST domain
sequences important for Pip recruitment, we prepared a number of PU.1
PEST domain mutants using a PCR-based technique. Fifteen internal
deletions were generated that had deletions anchored on the N-terminal
side at arginine residues 87, 117, or 140 and on the C-terminal side at
serine residues 107, 113, 126, 133, 142, 148, or 168 (Fig. 2A). Each mutant was
transcribed and translated in vitro using [35S]methionine
and assayed by EMSA for the ability to recruit Pip from S194 nuclear
extracts to the PU.1-Pip DNA-binding site of the Ig 3' enhancer
(sequences 445–469; numbering according to 3 . The results of
these assays are shown in Figure 2B.
|
B, lanes
2–3). Additional deletion to amino acid 125 or 132
resulted in a decrease in Pip recruitment ability (lanes
4–5), and progressive deletion to amino acid 141, 147, or
167 resulted in a total loss of Pip recruitment (lanes
6–8). Similarly, deletion of amino acids 118–125 or
118–132 (lanes 9–10) resulted in a decrease
in Pip binding, and progressive deletion to amino acid 141, 147, or 167
completely eliminated Pip recruitment (lanes
11–13). Therefore, PU.1 sequences between 118 and 141 are
necessary for Pip recruitment. Deletion of amino acids 141–147 or
141–167 also resulted in complete loss of Pip recruitment
(lanes 15–16). Interestingly, deletion of a
single amino acid, glutamine 141, produced a PU.1 mutant that appears
to recruit Pip with higher efficiency than wild-type PU.1
(lane 14). Similar results with each mutant
were seen using recombinant Pip protein (data not shown). Definition of the N-terminal and C-terminal limits of the recruitment domain
The above results suggested that amino acids 113–125 contain the
N-terminal boundary of the Pip recruitment domain (compare 88–112
to 88–125; Fig. 2B, lanes 3–4). We made two
additional deletions to more precisely determine the N-terminal
recruitment boundary (111–114 and 111–117; Fig. 2A) and assayed them for Pip recruitment via EMSA.
Neither deletion resulted in any loss of Pip recruitment relative to
wild-type PU.1, indicating that sequences amino terminal to residue 118
are not required for Pip recruitment to DNA (Fig. 3A, lanes 1–3).
Therefore, the N-terminal boundary of the Pip recruitment domain lies
between PU.1 amino acids 118 and 125.
|
154–156, 161–164, and 165–168; Fig. 2A).
None of these mutations caused any significant loss of Pip recruitment
in EMSA (Fig. 3A, lanes 10–12), suggesting that PU.1
sequences between residues 154 and 168 are not necessary for Pip
recruitment to DNA. Since mutation of serine 148 abolishes Pip
recruitment (16), the C-terminal boundary of the recruitment domain
lies between the PU.1 residues 148 and 154. Subregions within the PEST domain
Because deletion of amino acids 118–132 reduced Pip recruitment,
whereas deletion of sequences 118–141 abolished recruitment, important
sequences must also reside between amino acids 132 and 141. To better
characterize these sequences, we made three additional deletions
between amino acids 129–141: 129–132, 133–139, and 129–141
(Fig. 2A). Deletion of amino acids 129–132 had no
effect on Pip recruitment compared with wild-type PU.1 (Fig. 3A, compare lanes 6 and 7).
However, deletion of amino acids 133–139 resulted in a decrease in Pip
recruitment, as did deletion of amino acids 129–141
(lanes 8 and 9).
Our results indicate that three PU.1 regions (sequences 118–125,
133–139, and 141–147) can influence Pip recruitment when deleted.
However, deletion can potentially affect function by altering protein
structure nonspecifically. Therefore, to better characterize these
regions, we also prepared alanine point mutations between residues
118–125, 126–132, 133–139, and 141–147 (Fig. 2A).
Consistent with our deletion data, alanine mutants 118–125A and
133–139A reduced Pip recruitment to DNA, whereas mutant 126–132A had
no effect (Fig. 3B, lanes 1–4). Alanine mutant
141–147A reduced Pip recruitment (Fig. 3B, lane 5),
whereas deletion of the same sequences abolished recruitment (Fig. 2B, lane 15). This difference may be the result of
lost phosphorylation of serine 148 in the deletion mutant (discussed
below). Quantification of the recruitment efficiency of each alanine
mutant as compared with wild-type PU.1 is shown in Figure 3C. Mutation of residues 118–125A was the most severe,
reducing Pip recruitment >15-fold, while 133–139A and 141–147A
reduced recruitment 2.5-fold and 9-fold, respectively.
Figure 2A contains a summary of each PU.1 mutant and Figure 4 shows the PU.1 PEST sequence and the
specific sequences that influence Pip recruitment. Deletion or alanine
point mutation of residues 118–125 (subregion A) or 133–139
(subregion B) reduced Pip recruitment, indicating that these sequences
can modulate recruitment efficiency. Deletion of both regions
(118–141) completely eliminated Pip recruitment to the DNA, whereas
mutation of the sequences between subregions A and B had no effect on
Pip recruitment. Finally, alanine point mutation of residues 141–147
(subregion C) reduced Pip recruitment, whereas deletion of these
sequences abolished recruitment.
|
Since deletion of sequences between PU.1 residues 118–125 and
133–139 modulated the efficiency of Pip recruitment, we sought to
determine whether these mutants affected protein-DNA binding kinetics.
To test this hypothesis, 118–125, 133–139, and wild-type PU.1
proteins were assayed by off-rate EMSA experiments. These studies
showed differences in the off-rates of the PU.1:Pip:DNA complex when
comparing wild-type PU.1 with the mutant proteins (Fig. 5A). Interestingly,
these studies showed a slower off-rate by the 133–139 mutant as
compared with wild-type PU.1. The off-rate for the 118–125 mutant
was intermediate between wild-type PU.1 and the 133–139 mutant. A
similar trend was observed with the PU.1:DNA complex alone (data not
shown). The different off-rates of the 118–125 and 133–139
mutants compared with wild-type PU.1 suggested that these proteins
might have altered conformations. To determine whether this was the
case, we performed protease sensitivity studies. Wild-type PU.1,
118–125, and 133–139 proteins were bound to DNA, then subjected
to increasing doses of either trypsin (Fig. 5B) or
proteinase K (Fig. 5C). Interestingly, both mutant
proteins were slightly more resistant to protease digestion than was
wild-type PU.1; this can best be seen when comparing the lowest doses
of trypsin (Fig. 5B, compare lanes 2, 6, and
10) or proteinase K (Fig. 5C, compare
lanes 2, 6, and 10).
|
118–125 and 133–139 can influence the recruitment of
Pip to DNA. Mutation of these sequences can influence the kinetics of
DNA-protein complexes and can alter the conformation of these proteins.
It is interesting that the 133–139 mutant recruits Pip less
efficiently than the wild-type protein, yet has a slower off-rate with
Pip compared with wild-type PU.1. This issue will be addressed in the
Discussion below. Interaction of GST-Pip and PU.1 in solution requires the PU.1 ETS domain
The studies described above identified PU.1 sequences within the
PEST domain that are required for efficient recruitment of Pip to its
DNA-binding site as measured by EMSA. These studies require intact
DNA-binding domains because mutation of the PU.1 ETS domain abolishes
Pip recruitment (6). To study the interaction between PU.1 and Pip by a
second method, we prepared a glutathione S-transferase-Pip
fusion protein (GST-Pip) which enabled us to assay the PU.1-Pip
interaction in solution. Bacterially produced GST-Pip protein was
incubated with [35S]-labeled PU.1 in a GST chromatography
experiment in the absence or presence of various oligonucleotide
sequences (Fig. 6A).
Binding of wild-type PU.1 protein was easily detected in the absence of
DNA, indicating that PU.1 and Pip can physically interact in solution
(Fig. 6A, lanes 3–4). Addition of either PU.1-Pip
DNA-binding site specific (lanes 5–6) or
nonspecific (lanes 7–8) oligonucleotide
sequences had no effect on GST-Pip:PU.1 interaction.
|
141–147) did not abolish the interaction of
PU.1 with GST-Pip (Fig. 6A, lanes 9–14). This
unexpected result led us to test the specificity of this physical
interaction under more stringent conditions. First, we washed the
chromatography assays with NETN containing high salt and blocking
protein (0.5 M NaCl, 0.5% milk). This treatment did not effect the
relative binding of wild-type and mutant PU.1 proteins (data not
shown). Second, we tested the relative specificity of the PU.1:Pip
interaction in the context of total nuclear proteins. NIH3T3 cells were
transfected with either wild-type PU.1 or PU.1PEST. Transfected
cells were metabolically labeled with [35S]methionine,
and cysteine and cell extracts were prepared. Total labeled nuclear
proteins were then assayed for their ability to physically interact
with GST-Pip. Remarkably, of all the labeled cellular proteins, only
PU.1 and PU.1PEST interacted with high efficiency with GST-Pip (Fig. 6B, lanes 4–9). No specific protein interactions
were observed with cell lysates generated from untransfected cells
(lanes 2–3). Therefore, the physical interaction we observed with
PU.1PEST and GST-Pip is highly specific.
To identify the region of PU.1 that is required for interaction with
GST-Pip, we tested a panel of PU.1 deletion mutants in GST
chromatography assays. Proteins lacking the activation domain (7–30
and 33–100; Fig. 7A,
lanes 3–6), the PEST domain (119–160; Fig. 7A, lanes 7–8), or both (1–160; Fig. 7A,
lanes 11–12; Fig. 7B, lanes 8–9) strongly
interacted with GST-Pip. In contrast PU.1 mutants 245–272 (Fig. 7A, lanes 9–10) and 201–272 (Fig. 7B, lanes 6–7), which lack portions of the ETS
domain, were unable to bind to GST-Pip in these assays. These results
are summarized in Figure 7C. Our results show that the PU.1
ETS domain is necessary and sufficient for physical interaction with
GST-Pip. This domain, however, is not sufficient to recruit Pip to bind
to DNA (6). Recruitment of Pip to DNA requires a second step involving
the PU.1 PEST domain.
|
The observation that Pip can physically interact with the PU.1 ETS
domain prompted us to determine whether other ETS domain proteins can
interact with Pip. The ETS proteins Ets-1, Ets-2, and Fli-1 were
tested. Results from these studies (Fig. 8) showed that Ets-1 interaction with
GST-Pip was detectable but very weak (lanes 7 and
8). Fli-1 interacted with GST-Pip (lanes
11 and 12) although not as efficiently as PU.1
(lanes 5 and 6). Ets-2 was
incapable of interacting with Pip (lanes 9 and
10). Therefore, some, but not all, ETS domain
proteins can physically interact with GST-Pip.
|
Our results indicate a role of the PU.1 PEST domain in Pip
recruitment to DNA. PEST domains have also been hypothesized to confer
protein instability (37). We sought, therefore, to determine whether
the PEST domain sequences play a role in PU.1 stability. NIH3T3 cells
were transfected with plasmids expressing PU.1 or PU.1PEST, or both,
and 24 h post-transfection, cellular proteins were metabolically
labeled with [35S]-labeled amino acids for 2 h,
then chased with cold methionine for various time periods and
harvested. The PU.1 proteins were immunoprecipitated from cell lysates
with anti-PU.1 Ab and resolved by SDS-PAGE. As can be seen in
Figure 9, the PU.1PEST deletion is
actually moderately less stable than the wild-type protein, indicating
that the PEST domain does not confer an innate instability on
PU.1.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To better characterize the PU.1 sequences important for Pip
recruitment, we prepared 27 mutations that flank or traverse the PU.1
PEST domain. Consistent with our previous observations (6), we
demonstrated that deletion or alanine point mutation of sequences
within the PU.1 PEST domain significantly reduced or eliminated
recruitment of Pip to DNA. Based upon the data summarized in Figures 2A and 4, multiple regions of the PU.1 PEST domain influence
Pip recruitment. Mutation of residues 118–125 (subregion A), 133–139
(subregion B), or 141–147 (subregion C) reduced Pip recruitment.
The mechanism by which PU.1 sequences 118–125 (Met-Cys-Phe-Pro-Tyr-Gln-Thr-Leu) influence Pip recruitment is unknown. However, deletion of these sequences alters the conformation of PU.1, and this altered conformation may result in less efficient recruitment of Pip. This mutant protein might also serve as a less efficient substrate for phosphorylation of serine 148 by casein kinase (CK) II. However, addition of excess CKII did not increase Pip recruitment relative to wild-type PU.1 (data not shown). Therefore, it is more likely that mutation of sequences 118–125 influences the PU.1 conformation necessary for optimal recruitment of Pip.
Mutation of sequences 133–139 also reduced Pip recruitment. These sequences (Ser-Asp-Glu-Glu-Glu-Gly-Glu-Val) constitute a consensus sequence for CKII. Therefore, it is very likely that deletion of this sequence abolishes phosphorylation of serine 132, which then leads to reduced Pip recruitment. In support of this conclusion, we previously showed that mutation of serines 132 and 133 to alanine residues resulted in reduced Pip recruitment to DNA (16).
The only difference between our deletion and alanine point
mutation data occurred with sequences 141–147
(Gln-Ser-Pro-Pro-Leu-Glu-Val). Mutant 141–147 abolished Pip
recruitment, whereas mutant 141–147A only reduced recruitment. Both
mutants retain serine 148, of which phosphorylation is required for
PU.1 to recruit Pip to DNA (16). However, deletion of the sequences
immediately adjacent to serine 148 in mutant 141–147 may alter the
ability of this residue to be phosphorylated. This hypothesis is
supported by the less severe phenotype of the 141–147A mutant, which
replaces these sequences with alanine residues. Therefore, the effects
on Pip recruitment by mutation of residues 141–147 may be a
consequence of reduced phosphorylation of serine 148. The multiple
phosphorylated serines in PU.1 make it very difficult to determine the
phosphorylation status of serine 148. This would require
two-dimensional peptide mapping of metabolically labeled proteins
harvested from transfected cells.
Somewhat surprisingly, mutants 118–125 and 133–139 showed
slower off-rates from DNA as compared with wild-type complexes.
Therefore, these mutants have a reduced ability to recruit Pip to DNA,
but once assembled, the PU.1:Pip complex comes off the DNA more slowly.
Possibly the mutant PU.1 proteins take on a conformation that
stabilizes protein-DNA complexes, but they do not as readily undergo
conformational changes that facilitate Pip recruitment (see below).
Whether deletion of these sequences directly alters a sequence
necessary for Pip recruitment or alters the structure of other PU.1
sequences necessary for this function is unclear.
We also studied the interaction of PU.1 and Pip in solution using GST
chromatography. Surprisingly, in contrast to our EMSA data, deletion of
the PU.1 PEST domain had no effect on PU.1:GST-Pip physical
interaction. In fact, deletion of all PU.1 sequences except the ETS
domain had no effect on PU.1:Pip interaction. On the other hand,
deletions within the PU.1 ETS domain abolished the physical interaction
between PU.1 and Pip. Therefore, the PU.1 ETS domain is necessary and
sufficient for physical interaction with Pip. Since PU.1 belongs to a
family of related ETS domain proteins, we tested the ability of several
other ETS domain proteins to physically interact with GST-Pip.
Interestingly, Fli-1 was also able to physically interact with GST-Pip,
although the interaction between PU.1 and Pip appears to be more
efficient (Fig. 8). Fli-1 and Pip are both found in B and T lymphocytes
(26, 38, 39). It is therefore possible that Pip interacts with Fli-1 in
either lymphocyte lineage. The ability of Pip to select from multiple
potential dimerization partners may serve to modulate its activity by
controlling DNA-binding site specificity, DNA-binding kinetics, or
transactivation potential. It will be interesting to determine whether
Pip normally interacts with other ETS domain proteins in addition
to PU.1.
A two-step model for PU.1 recruitment of Pip to DNA
Two distinct PU.1 domains (the PEST and ETS domains) are required
for the recruitment of Pip to its 3' enhancer DNA site. Whereas both
the PU.1 PEST and ETS domains are required for Pip recruitment to DNA,
only the ETS domain is required for a solution interaction. This is
supported by several observations. First, Pip recruitment via EMSA is
eliminated by various deletions of the PEST domain, even though these
clones contain the PU.1 ETS domain (Fig. 2A; see also 6 . Second, the PU.1 ETS domain alone is incapable of recruiting Pip to
its Ig 3' enhancer DNA-binding site as assayed via EMSA (6). Third,
mutations within the ETS domain abolish PU.1 DNA binding, and as a
consequence, Pip recruitment (6). Finally, GST-Pip is able to interact
with a variety of PU.1 mutants, including PU.1PEST, but not with
certain ETS domain mutants (Fig. 7).
A model consistent with the above observations is a modification of the
one put forward by Brass et al. (30), who suggested that PU.1
interaction with Pip results in a Pip conformational change that
exposes the Pip DNA-binding domain. Here, we propose a two-step
mechanism that results in conformational changes in both PU.1 and Pip
(Fig. 10). Initially, PU.1 can interact
with Pip in solution via the PU.1 ETS domain. This interaction does not
require the PU.1 PEST domain or PU.1 phosphorylation and does not
result in recruitment of Pip to DNA. Second, phosphorylation of PU.1 at
serine 148 causes a covalent structural change in the PU.1 protein.
This structural change may directly induce changes in Pip (see below),
or may induce a PU.1 conformational change perhaps involving PEST
subregions A, B, and C (sequences 118–125, 133–139, and 141–147,
respectively). Phosphorylation of PU.1 may occur either before or after
interaction with Pip, but phosphorylation at serine 148 is critical
because bacterial PU.1 is not capable of Pip recruitment unless first
treated with CKII (16). The phosphorylation-induced change in PU.1
would then induce a subsequent conformational change in Pip that either
exposes or alters the Pip DNA-binding domain, enabling recruitment of
Pip to the 3' enhancer.
|
While the physical interaction between PU.1 and Pip is highly specific,
other proteins are known to physically interact with PU.1. These
include TATA-binding protein (TBP), the retinoblastoma protein (Rb),
the glucocorticoid receptor (GR), NF-IL6ß, HMG I/Y, and
structure-specific recognition protein (40, 41, 42). TBP, Rb, and GR appear
to interact with the amino-terminal region of PU.1, whereas we show
that Pip physically interacts with the carboxyl-terminal PU.1 ETS
domain (Fig. 7). Interestingly, NF-IL6ß also physically interacts
with the PU.1 ETS domain (42). However, the interaction between PU.1
and NF-IL6ß does not result in cooperative DNA binding as is observed
with PU.1 and Pip. The ability of the PU.1 ETS domain to physically
interact with multiple proteins suggests that these interactions could
be used as a regulatory mechanism. It will be interesting to determine
whether NF-IL6ß can disrupt the interaction between PU.1 and
Pip.
PU.1 is a pivotal protein involved in hemopoetic development and in the genesis of erythroleukemia (10, 18, 19, 20, 21, 22, 23, 24). PU.1 can regulate expression of multiple genes and can physically interact with multiple proteins (11, 12, 13, 14, 15, 16, 17, 26, 40, 41, 42). It will be very interesting to determine whether the PU.1 sequences that are necessary for recruitment of Pip to DNA are also necessary for other PU.1 functions such as hemopoetic development, oncogenesis, gene regulation, and interaction with other proteins.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Michael Atchison, 3800 Spruce Street, University of Pennsylvania, School of Veterinary Medicine, Philadelphia, PA 19104. E-mail address:
3 Abbreviations used in this paper: PEST, domain rich in the amino acids proline, glutamate, serine, and threonine; NETN, 100 mM NaCl/1 mM EDTA/20 mM Tris, pH 8.0/0.5% Nonidet P-40; EMSA, electrophoretic mobility shift assay; GST, glutathione S-transferase.
Received for publication August 4, 1997. Accepted for publication September 19, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
enhancer and its binding factor NF-B in the developmental regulation of gene transcription. Cell 48:121.[Medline]
locus contains a second, stronger B-cell-specific enhancer which is located downstream of the constant region. EMBO J. 8:1959.[Medline]
3'-enhancer necessary for maximal enhancer activity. Biochemistry 35:8662.[Medline]
), that binds to the immunoglobulin 3' enhancer and the immunoglobulin heavy chain µ E1 site. Proc. Natl. Acad. Sci. USA 88:9804. 3' enhancer activity. Mol. Cell. Biol. 12:368. 3' enhancer. J. Biol. Chem. 270:10304. regulate expression of the granulocyte-macrophage colony-stimulating factor receptor gene. Mol. Cell. Biol. 15:5830.[Abstract]
2–4 enhancer. Mol. Cell. Biol. 13:6452., CRP3). J. Immunol. 155:4330.[Abstract]
This article has been cited by other articles:
|
|
 |
|
  F. Wei, K. Zaprazna, J. Wang, and M. L. Atchison PU.1 Can Recruit BCL6 to DNA To Repress Gene Expression in Germinal Center B Cells Mol. Cell. Biol., September 1, 2009; 29(17): 4612 - 4622. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. C. Azim, X. Wang, G. Y. Park, R. T. Sadikot, H. Cao, B. Mathew, M. Atchison, R. B. van Breemen, M. Joo, and J. W. Christman NF-{kappa}B-Inducing Kinase Regulates Cyclooxygenase 2 Gene Expression in Macrophages by Phosphorylation of PU.1 J. Immunol., December 1, 2007; 179(11): 7868 - 7875. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Dahl, S. R. Iyer, K. S. Owens, D. D. Cuylear, and M. C. Simon The Transcriptional Repressor GFI-1 Antagonizes PU.1 Activity through Protein-Protein Interaction J. Biol. Chem., March 2, 2007; 282(9): 6473 - 6483. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Hodawadekar, F. Wei, D. Yu, A. Thomas-Tikhonenko, and M. L. Atchison Epigenetic Histone Modifications Do Not Control Ig{kappa} Locus Contraction and Intranuclear Localization in Cells with Dual B Cell-Macrophage Potential J. Immunol., November 1, 2006; 177(9): 6165 - 6171. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Guillouf, I. Gallais, and F. Moreau-Gachelin Spi-1/PU.1 Oncoprotein Affects Splicing Decisions in a Promoter Binding-dependent Manner J. Biol. Chem., July 14, 2006; 281(28): 19145 - 19155. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Wang, R. M. Baron, G. Zhu, M. Joo, J. W. Christman, E. S. Silverman, M. A. Perrella, R. J. Riese, and M. Cernadas PU.1 Regulates Cathepsin S Expression in Professional APCs J. Immunol., January 1, 2006; 176(1): 275 - 283. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Lehtonen, V. Veckman, T. Nikula, R. Lahesmaa, L. Kinnunen, S. Matikainen, and I. Julkunen Differential Expression of IFN Regulatory Factor 4 Gene in Human Monocyte-Derived Dendritic Cells and Macrophages J. Immunol., November 15, 2005; 175(10): 6570 - 6579. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Bai, L. Srinivasan, L. Perkins, and M. L. Atchison Protein Acetylation Regulates Both PU.1 Transactivation and Ig{kappa} 3' Enhancer Activity J. Immunol., October 15, 2005; 175(8): 5160 - 5169. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Laricchia-Robbio, T. Tamura, T. Karpova, B. L. Sprague, J. G. McNally, and K. Ozato Partner-regulated interaction of IFN regulatory factor 8 with chromatin visualized in live macrophages PNAS, October 4, 2005; 102(40): 14368 - 14373. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. V. Pedchenko, G. Y. Park, M. Joo, T. S. Blackwell, and J. W. Christman Inducible binding of PU.1 and interacting proteins to the Toll-like receptor 4 promoter during endotoxemia Am J Physiol Lung Cell Mol Physiol, September 1, 2005; 289(3): L429 - L437. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. C. McDevit, L. Perkins, M. L. Atchison, and B. S. Nikolajczyk The Ig{kappa}3' Enhancer Is Activated by Gradients of Chromatin Accessibility and Protein Association J. Immunol., March 1, 2005; 174(5): 2834 - 2842. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. van der Stoep, E. Quinten, M. M. Rezende, and P. J. van den Elsen E47, IRF-4, and PU.1 synergize to induce B-cell-specific activation of the class II transactivator promoter III (CIITA-PIII) Blood, November 1, 2004; 104(9): 2849 - 2857. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Suzuki, K. Honma, T. Matsuyama, K. Suzuki, K. Toriyama, I. Akitoyo, K. Yamamoto, T. Suematsu, M. Nakamura, K. Yui, et al. From the Cover: Critical roles of interferon regulatory factor 4 in CD11bhighCD8{alpha}- dendritic cell development PNAS, June 15, 2004; 101(24): 8981 - 8986. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Hines, B. R. Sorensen, M. A. Shea, and W. Maury PU.1 Binding to ets Motifs within the Equine Infectious Anemia Virus Long Terminal Repeat (LTR) Enhancer: Regulation of LTR Activity and Virus Replication in Macrophages J. Virol., April 1, 2004; 78(7): 3407 - 3418. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Alter-Koltunoff, S. Ehrlich, N. Dror, A. Azriel, M. Eilers, H. Hauser, H. Bowen, C. H. Barton, T. Tamura, K. Ozato, et al. Nramp1-mediated Innate Resistance to Intraphagosomal Pathogens Is Regulated by IRF-8, PU.1, and Miz-1 J. Biol. Chem., November 7, 2003; 278(45): 44025 - 44032. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Hashmueli, M. Gleit-Kielmanowicz, D. Meraro, A. Azriel, D. Melamed, and B.-Z. Levi A truncated IFN-regulatory factor-8\IFN consensus sequence-binding protein acts as dominant-negative, interferes with endogenous protein-protein interactions and leads to apoptosis of immune cells Int. Immunol., July 1, 2003; 15(7): 807 - 815. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. O'Reilly, C. M. Quinn, T. El-Shanawany, S. Gordon, and D. R. Greaves Multiple Ets Factors and Interferon Regulatory Factor-4 Modulate CD68 Expression in a Cell Type-specific Manner J. Biol. Chem., June 6, 2003; 278(24): 21909 - 21919. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. R. McKercher, C. R. Lombardo, A. Bobkov, X. Jia, and N. Assa-Munt Identification of a PU.1-IRF4 protein interaction surface predicted by chemical exchange line broadening PNAS, January 21, 2003; 100(2): 511 - 516. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Kuwata, C. Gongora, Y. Kanno, K. Sakaguchi, T. Tamura, T. Kanno, V. Basrur, R. Martinez, E. Appella, T. Golub, et al. Gamma Interferon Triggers Interaction between ICSBP (IRF-8) and TEL, Recruiting the Histone Deacetylase HDAC3 to the Interferon-Responsive Element Mol. Cell. Biol., November 1, 2002; 22(21): 7439 - 7448. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Meraro, M. Gleit-Kielmanowicz, H. Hauser, and B.-Z. Levi IFN-Stimulated Gene 15 Is Synergistically Activated Through Interactions Between the Myelocyte/Lymphocyte-Specific Transcription Factors, PU.1, IFN Regulatory Factor-8/IFN Consensus Sequence Binding Protein, and IFN Regulatory Factor-4: Characterization of a New Subtype of IFN-Stimulated Response Element J. Immunol., June 15, 2002; 168(12): 6224 - 6231. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Hong, A. Y. Kim, S. Ky, C. Rakowski, S.-B. Seo, D. Chakravarti, M. Atchison, and G. A. Blobel Inhibition of CBP-Mediated Protein Acetylation by the Ets Family Oncoprotein PU.1 Mol. Cell. Biol., June 1, 2002; 22(11): 3729 - 3743. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Kim and R. A. Feldman Activated Fes Protein Tyrosine Kinase Induces Terminal Macrophage Differentiation of Myeloid Progenitors (U937 Cells) and Activation of the Transcription Factor PU.1 Mol. Cell. Biol., March 15, 2002; 22(6): 1903 - 1918. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Marecki, C. J. Riendeau, M. D. Liang, and M. J. Fenton PU.1 and Multiple IFN Regulatory Factor Proteins Synergize to Mediate Transcriptional Activation of the Human IL-1{{beta}} Gene J. Immunol., June 1, 2001; 166(11): 6829 - 6838. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Gupta, A. Anthony, and A. B. Pernis Stage-Specific Modulation of IFN-Regulatory Factor 4 Function by Kruppel-Type Zinc Finger Proteins J. Immunol., May 15, 2001; 166(10): 6104 - 6111. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Lubyova and P. M. Pitha Characterization of a Novel Human Herpesvirus 8-Encoded Protein, vIRF-3, That Shows Homology to Viral and Cellular Interferon Regulatory Factors J. Virol., September 1, 2000; 74(17): 8194 - 8201. [Abstract] [Full Text]
|
|
|
|
|
|
S. Maitra and M. Atchison BSAP Can Repress Enhancer Activity by Targeting PU.1 Function Mol. Cell. Biol., March 15, 2000; 20(6): 1911 - 1922. [Abstract] [Full Text]
|
|
|
|
|
|
B. Falini, M. Fizzotti, A. Pucciarini, B. Bigerna, T. Marafioti, M. Gambacorta, R. Pacini, C. Alunni, L. Natali-Tanci, B. Ugolini, et al. A monoclonal antibody (MUM1p) detects expression of the MUM1/IRF4 protein in a subset of germinal center B cells, plasma cells, and activated T cells Blood, March 15, 2000; 95(6): 2084 - 2092. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Meraro, S. Hashmueli, B. Koren, A. Azriel, A. Oumard, S. Kirchhoff, H. Hauser, S. Nagulapalli, M. L. Atchison, and B.-Z. Levi Protein-Protein and DNA-Protein Interactions Affect the Activity of Lymphoid-Specific IFN Regulatory Factors J. Immunol., December 15, 1999; 163(12): 6468 - 6478. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Marecki, M. L. Atchison, and M. J. Fenton Differential Expression and Distinct Functions of IFN Regulatory Factor 4 and IFN Consensus Sequence Binding Protein in Macrophages J. Immunol., September 1, 1999; 163(5): 2713 - 2722. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Cohen, A. Azriel, T. Cohen, D. Meraro, S. Hashmueli, D. Bech-Otschir, R. Kraft, W. Dubiel, and B.-Z. Levi Interaction between Interferon Consensus Sequence-binding Protein and COP9/Signalosome Subunit CSN2 (Trip15). A POSSIBLE LINK BETWEEN INTERFERON REGULATORY FACTOR SIGNALING AND THE COP9/SIGNALOSOME J. Biol. Chem., December 8, 2000; 275(50): 39081 - 39089. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Rieske and J. M. R. Pongubala AKT Induces Transcriptional Activity of PU.1 through Phosphorylation-mediated Modifications within Its Transactivation Domain J. Biol. Chem., March 9, 2001; 276(11): 8460 - 8468. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Luchin, S. Suchting, T. Merson, T. J. Rosol, D. A. Hume, A. I. Cassady, and M. C. Ostrowski Genetic and Physical Interactions between Microphthalmia Transcription Factor and PU.1 Are Necessary for Osteoclast Gene Expression and Differentiation J. Biol. Chem., September 21, 2001; 276(39): 36703 - 36710. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
