 |
INTRODUCTION |
B-cell development represents a complex series of steps during
which a single, pluripotent stem cell gives rise to a mature B-lymphocyte with rearranged heavy and light chain loci. Transcription factors play a crucial role in this process by regulating gene expression which affects both lineage specification and function (1,
2). Ets DNA-binding proteins represent a key family of transcription
factors implicated in lymphoid development and activity. Ets proteins
consist of monomeric transcription factors which bind the DNA site
GGA(A/T) through their Ets domain, a unique winged helix-turn-helix
motif (3-6). PU.1, Spi-B, and the recently described Spi-C (7) are a
divergent group of Ets proteins due to unique amino acids in their Ets
domain relative to other family members. Because of these differences,
PU.1 and Spi-B proteins can also bind the non-canonical AGAA (8,
9).
PU.1/Spi-1 was cloned as a B-cell- and macrophage- restricted DNA
binding activity (10) and as an oncogene which causes erythroleukemias
(11). PU.1 was eventually shown to transactivate a large number of
B-cell genes such as those encoding CD72 (12), LSP1 (13), CD20 (14),
Btk (15), J-chain (8), mb-1 (16), µ heavy chain (17), 
(18, 19),
and 
(20) light chains. Spi-B was cloned based upon its homology to
PU.1 (90% similar in the Ets domain), and has been shown to bind many
of the same sites as PU.1 (although with different affinities) and
transactivate some of the same genes as PU.1 (9, 21, 22). While PU.1 is
expressed in B-cells, early T-cells, megakaryocytes, granulocytes, mast
cells, immature erythrocytes, and myeloid cells (10, 21, 23), Spi-B is
restricted to B-cells and immature T-cells (9, 24). Thus, B-lymphocytes
express at least two Ets proteins which can bind to similar DNA sites
and transactivate similar genes.
To better understand the roles of PU.1 and Spi-B in B-lymphocyte
development, function, and transcription, we created targeted disruptions of both genes in mice. PU.1
/
animals die at approximately day 16.5 of gestation and lack lymphoid and myeloid cells (25) due to a cell intrinsic defect (26) of a
recently described lymphoid-myeloid common progenitor (27). Mice with a
different targeted allele of PU.1 display a less severe phenotype, but also lack B-cells (28, 29). In contrast to the dramatic
phenotype presented by the PU.1/ animals,
Spi-B/ animals are viable and have no defect
in B- or T-cell numbers (30). However,
Spi-B/ B-cells exhibit an interesting defect
in BCR1 signaling which
appears to be due to misexpression of one or more novel signaling
molecule(s) (31).
PU.1/ and Spi-B/ mice
present two of the problems commonly observed when analyzing the
phenotype of knock-out animals. The PU.1/
mice completely lack B-cells, and thus it is impossible to use these
animals as a tool for exploring the role of PU.1 in a mature B-cell. In
contrast, Spi-B/ mice may exhibit a mild
phenotype because their B-cells express wild-type levels of PU.1
protein, which could compensate for the lack of Spi-B (31). To assess
functional redundancies of PU.1 and Spi-B, we generated
PU.1+/Spi-B/ mice
(31). These mice are viable, but unlike the
PU.1+/+Spi-B/
animals, they exhibit perturbed B-cell numbers as well as a more severe
decrease in BCR mediated signaling. Thus,
PU.1+/Spi-B/ mice
act as a genetic model which allows us to explore the roles of this
subgroup of Ets proteins in B-cell transcription and function.
Although a large number of B-cell PU.1/Spi-B target genes have been
identified, all major signaling molecules are expressed at normal
levels in
PU.1+/Spi-B/ mice
(31). We hoped to identify novel PU.1/Spi-B target genes important for
BCR signaling using subtractive hybridization on PU.1+/Spi-B/
B-cells. These mice provided an attractive population to perform the
subtraction because they are phenotypically identical to
PU.1+/+Spi-B+/+ B-cells
based on fluorescence-activated cell sorter analysis except for the
loss of Spi-B and PU.1. One cDNA corresponding to an mRNA
underexpressed in
PU.1+/Spi-B/
B-cells encodes a lymphoid-restricted, heptahelical receptor (P2Y10)
that couples through heterotrimeric G-proteins. The promoter of
P2Y10 contains a functionally important PU.1/Spi-B-binding site critical for its transcriptional activity in B-cells. We, therefore, identify an in vivo PU.1/Spi-B target gene
dependent on wild-type levels of PU.1 and Spi-B for its transcription.
P2Y10 is unique in the way it was identified and can act as a model for
transcriptional regulation by this important group of DNA-binding proteins in B-cells. In addition, P2Y10 provides an intriguing connection between heterotrimeric G-proteins and membrane proximal BCR
signaling events.
| |
EXPERIMENTAL PROCEDURES |
Cell Purification and RNA Extraction--
Splenic B-lymphocytes
were purified from mice of the appropriate genotypes by subjecting
ammonium chloride-lysed splenocytes to anti-Thy1.2 plus rabbit
complement mediated depletion. B-lymphocytes were at least 75% pure as
measured by B220 staining. Splenic CD3+ T cells were
purified via magnetic cell sorting (PerSeptive Biosystems). Total RNA
was isolated from all cells and tissues using the Trizol reagent (Life
Technologies, Inc.) according to manufacturer's protocols.
Subtractive Hybridization--
The subtractive hybridization
procedure was performed using the PCR-cDNA Subtraction and
PCR-Select Differential Screening kits (CLONTECH),
as per the manufacturer's protocols. Due to the limited amount of
total RNA isolated from purified B-lymphocytes, the cDNA used in
the subtraction was prepared by the SMART PCR cDNA Synthesis kit
(CLONTECH).
Northern Analysis--
All Northern blots utilized 10-20 µg
of total RNA and were prepared as described previously (9). For P2Y10,
the probe was an 800-base pair RsaI fragment contained in
the 3'-untranslated region of the cDNA which was isolated by the
subtractive hybridization. The PU.1 probe was a 350-base pair
ApaI digested fragment of the 3'-untranslated region, and
the Spi-B probe has been previously described (9). Quantification of
signal intensity was done on a Molecular Dynamics PhosphorImager using
Image Quant software.
cDNA Library Screening and Promoter Isolation--
To
isolate the full-length cDNA, the 800-base pair fragment generated
from the subtractive hybridization was used as a probe to screen a
C57/BL6 splenocyte library (Stratagene). Pure phage were isolated
according to the manufacturer's protocols and sequenced to confirm
their identity.
The promoter of P2Y10 was initially isolated using the Genome Walker
kit (CLONTECH) and yielded fragments from 0.6-2.2
kb in size. These clones were sequenced and used to generate PCR primers to directly amplify either a 0.612-kb fragment or a 0.100-kb of
the promoter with the following primers: 5' primer (0.6 kb) CTTGAGAATTCATTTATTCACCCACTTGC, 5' primer (0.1 kb)
GGGACTTACCTACGTTCTGTGCAAGCAAG, 3' primer (wild-type PU.1 site)
CTGCTGATAACTGAAAGAAAAATAATGTTG, 3' primer (mutant PU.1 site)
GAAAAATAATGTTGAAACACCGGGTAGAGCACATTAATGTC. Both 3' primers were
contained in the 5'-untranslated region of the P2Y10 cDNA. PCR was
performed using the high-fidelity polymerase Bio-X-Act (Bioline) with
C57/BL6 mouse genomic DNA as a template for a limited number of cycles.
These promoter fragments (referred to as 0.6 kb wt and 0.6 kb mut) were
subcloned into the SalI-BamHI site of the
promoter-less reporter construct p
GH (Nichols Institute) which
expresses the human growth hormone (human growth hormone) cDNA.
Promoter constructs were sequenced and revealed no nucleotide differences from a consensus sequence developed from 5 individual PCR isolates.
Transient Transfections--
A20 and EL-4 cells were cultured in
RPMI 1640 supplemented with 10% fetal bovine serum (Life Technologies,
Inc.), 100 units/ml penicillin, 100 µg/ml streptomycin, and 0.1 mM 
-mercaptoethanol. MEL and NIH 3T3 cells were cultured
in Dulbecco's modified Eagle's medium supplemented with 10% fetal
bovine serum, 100 units/ml penicillin, and 100 µg/ml streptomycin.
For MEL, EL-4, and A20 transfections, 20 µg of reporter DNA and 2 µg of the luciferase expressing plasmid pGL2 Control (Promega) were
added to 107 cells in 0.4 ml of media without additives in
a 0.4-cm gap electroporation cuvette (Bio-Rad) and electroporated at
250 V, 960 microfarads using a Bio-Rad Gene-Pulser. Cells were then
incubated for 10 min on ice and replated in normal growth media. Cell
supernants and cell extracts were prepared 24 h after transfection
and assayed for human growth hormone (Nichols Institute) and luciferase
activity (Promega) as a monitor of transfection efficiency following
the manufacturer's protocols.
For NIH 3T3 co-transfections, cells were transiently transfected using
the Superfect reagent (Qiagen) according to the manufacturer's protocols. For each co-transfection, 1 µg of reporter plasmid (0.6 kb
wt or 0.6 kb mut), 1 µg of a -galactosidase expressing plasmid
(pMSVgal), and 3 µg of an expression plasmid for PU.1, Spi-B, or
Spi-B 
TA, all with a hemagglutinin epitope tag which have been
described previously (22). 48 h after transfection, supernatants
were collected and assayed for human growth hormone production and
transfection efficiency was monitored by -galactosidase activity as
described previously (32).
Electrophoretic Mobility Shift Assays (EMSA)--
Nuclear
extracts (NE) and in vitro Transcription and Translation
(IVT) proteins were prepared as described previously (22).
Binding reactions were performed for 30 min at room temperature and
contained equimolar amounts of IVTs, 10 µg of A20 NE or 2.5 µg of
primary B-cell NE, 10 mM Tris, 1 mM EDTA, 1 mM dithiothreitol, 75 mM KCl, 4% Ficoll,
5 × 105 cpm/ml of 32P-labeled
double-stranded oligonucleotide probe, and either 12.5 µg/ml (IVTs)
or 60 µg/ml (NE) poly(dI·dC) (Amersham Pharmacia Biotech). At this
time, cold competitor oligonucleotide and preimmune or antisera were
added. The anti-Spi-B antisera has been previously described (9); the
anti-PU.1 antibody was purchased from Santa Cruz. Protein-DNA complexes
were resolved on a 6% (19:1) polyacrylamide:bisacrylamide (Bio-Rad),
0.5 × TBE gel at 200 V for 3.5 h, dried, and subjected to autoradiography.
The following double-stranded synthetic oligonucleotides were used (top
strand) B 5'-CTAGCGAGAAATAAAAGGAAGTGAAACCAAGT-3', GAL4
5'-GAGCGGAGTACTGTCCTCCGAG-3', P2Y10 5'-TGTGCTACTTCCTCTTTCAACATT-3', P2Y10 mut 5'-TGTGCTCTACCCGGTGTTTCAACATT-3'.
| |
RESULTS |
Subtractive Hybridization Identifies Genes Poorly Expressed in
PU.1+/Spi-B/ B-Lymphocytes--
To
identify B-cell target genes of PU.1 and Spi-B, we chose a PCR-based
subtractive hybridization approach (Fig.
1) based upon the fact that resting
B-cells from
PU.1+/Spi-B/ mice
are essentially phenotypically normal (31).
PU.1+/Spi-B/
B-cells, because of reduced levels of PU.1 and Spi-B, should express
lower levels of any target genes critically dependent on normal levels
of PU.1 and Spi-B for their transcription. Therefore, splenic B-cells
from a PU.1+/Spi-B/
mouse and a wild-type littermate were purified and cDNA prepared as
described under "Experimental Procedures." These cDNAs were then used in a subtractive hybridization procedure (33). After a
preliminary screen, 42 individual clones were isolated, sequenced, and
subjected to analysis by either Northern blot or reverse
transcriptase-PCR based approaches to identify which ones were
differentially expressed. Of these 42, only 3 cDNA fragments
appeared to be expressed at lower levels in
PU.1+/Spi-B/
B-cells (data not shown). One cDNA (3H6) was further characterized because it was consistently underexpressed in
PU.1+/Spi-B/
B-cells prepared from multiple mice (data not shown). As shown in Fig.
2, this mRNA species was reduced
approximately 2-fold in
PU.1+/+Spi-B/ B-cells
and 8-fold in
PU.1+/Spi-B/
B-cells as compared with
PU.1+/+Spi-B+/+ B-cells.
The isolated cDNA fragment was used to obtain cDNAs from a
mouse splenocyte library and yielded three overlapping phage clones.
One clone contained an open reading frame of 328 amino acids with
strong homology to the human purinergic receptor P2Y10 (GenBank
accession number AF000545), a heterotrimeric G-protein coupled,
heptahelical receptor (Fig.
3A). Because the coding
regions of the two genes are highly related (83% identical, 88%
similar), it appears that 3H6 is the mouse homologue of human P2Y10. An
alignment of mouse and human P2Y10 amino acids, the most closely
related family member P2Y5, and the founding family member P2Y1 are
shown in Fig. 3B. As indicated in Fig. 3, A and B, homology between these four receptors is greatest in the
seven transmembrane domains.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 1.
Diagram of the PCR based subtractive
hybridization procedure used to isolate mRNA species expressed at
lower levels in
PU.1+/Spi-B/
as compared with
PU.1+/+Spi-B+/+
B-cells. Briefly, total RNA was prepared from purified splenic
B-cells, amplified by reverse transcriptase-PCR using a limited number
of cycles to specifically yield cDNAs, and digested with
RsaI. Tester cDNAs (from
PU.1+/+Spi-B+/+ B-cells)
had 2 different adaptors ligated onto their ends and were separately
hybridized to an excess of driver cDNA (from
PU.1+/Spi-B/
B-cells), and then hybridized to each other with an additional excess
of driver cDNA. Differentially expressed genes were amplified by
using primers specific to the two adaptors and subcloned.
|
|
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 2.
The clone 3H6 is down-regulated in
PU.1+/Spi-B/
B-cells. Northern blot on purified splenic B-cells from
PU.1+/+Spi-B+/+,
PU.1+/+Spi-B/, and
PU.1+/Spi-B/ mice showing the
expression of a clone (3H6) isolated from the subtractive
hybridization. The same blot was then stripped and reprobed with an
actin probe to control for loading. Expression levels were normalized
for loading and shown as a percent of wild-type.
|
|
View larger version (54K):
[in this window]
[in a new window]
|
Fig. 3.
Clone 3H6 is the heptahelical purinergic
receptor, P2Y10. A, deduced amino acid sequence of
P2Y10. The putative transmembrane domains are underlined and
numbered. B, alignment of the murine P2Y10 protein with its
human homologue, its most closely related family member P2Y5, and the
founding member of the P2Y family P2Y1. Red indicates
identity, blue indicates similarity.
|
|
Therefore, by using the
PU.1+/Spi-B/
B-lymphocytes in a subtractive hybridization we identified the
heptahelical receptor P2Y10 as a possible novel target gene of PU.1 and
Spi-B in vivo. In addition, it appears that P2Y10
transcription is dependent on maximal levels of both PU.1 and Spi-B for
its expression in B-cells.
P2Y10 Expression Is Restricted to Lymphocytes--
To further
characterize the murine P2Y10 gene, we examined its tissue
and cell line distribution using Northern blot analysis. Using organs
isolated from adult mice, P2Y10 mRNA is expressed predominantly in
the spleen and thymus, with very low levels of expression detected in
bone marrow (Fig. 4A and data
not shown). When murine cell lines were tested, P2Y10 appeared to be
exclusively expressed in immature and mature T-cells, immature and
mature B-cells, and faint expression was detected in one of the two
pre-B cell lines tested (Fig. 4B). However, it was not
expressed in macrophage, erythroid, or fibroblast cell lines.
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 4.
P2Y10 is lymphoid restricted.
A, examination of the tissue expression pattern of
P2Y10 by Northern blot analysis. RNA was prepared from the
indicated tissues of adult mice and total RNA was extracted. 18 S
ribosomal RNA is shown to control for loading. B,
examination of P2Y10, PU.1, and Spi-B
expression in various murine cell lines by Northern blot analysis on
total RNA extracted from Pro-B (NFS 70 c/10), pre-B (70 z/3 and 38 B9),
immature B (WEHI 231), mature B (A20), mature T (RMA), immature T
(EL-4), macrophage (J774.1), erythroid (MEL), and fibroblasts (3T3). 18 S ribosomal RNA is shown to control for loading. 28 S ribosomal RNA was
equivalently loaded (data not shown). C, P2Y10
expression in
PU.1+/+Spi-B+/+ B cells,
PU.1+/+Spi-B+/+ T-cells,
and PU.1+/Spi-B/ T cells. 18 S
RNA and actin serve as loading controls.
|
|
When these same cell lines were tested for Spi-B and PU.1 expression,
PU.1 appears to be expressed at all stages of B-cell development,
whereas Spi-B is expressed exclusively in immature and mature B-cells
(Fig. 4B). Neither PU.1 nor Spi-B is expressed in the T-cell
lines tested, but PU.1 is expressed in the macrophage cell line (Fig.
4). The erythroid cell line used (MEL) has been shown by others to
express low levels of PU.1 (10, 34, 35). Thus, it appears that P2Y10
mRNA expression mirrors that of Spi-B in B-cells, but is also
expressed in T-cells where neither PU.1 nor Spi-B is expressed. This
data indicates that in B-cells P2Y10 may be a direct
in vivo transcriptional target of PU.1 and Spi-B. To
investigate if P2Y10 expression was altered in T-cells
harvested from PU.1+/+Spi-B/
(data not shown) or
PU.1+/Spi-B/ mice,
we probed the Northern blot represented in Fig. 4C. As shown
in Fig. 4, the expression of P2Y10 in T cells is not
reproducibly affected by the genotype at either the PU.1 and
Spi-B loci. These results indicate that P2Y10 is
regulated by other factors in T lymphocytes (see below).
The P2Y10 Promoter Is Lineage Restricted and Contains a
Functionally Important PU.1/Spi-B-binding Site--
To ensure that
P2Y10 is a direct transcriptional target of PU.1 and Spi-B,
the promoter of P2Y10 was isolated as described under
"Experimental Procedures," and a 0.6-kb fragment was sequenced to
identify possible transcription factor-binding sites (Fig. 5A). The transcriptional start
site was confirmed by performing 5'-rapid amplification of cDNA
ends, no additional sequences upstream of the cDNA isolated from
the splenocyte library were detected (data not shown). Sequence
analysis of the region immediately upstream of the transcriptional
start site revealed no consensus TATA box, but a nearly perfect
PU.1/Spi-B-binding site is located at 
11 base pairs (36).
View larger version (72K):
[in this window]
[in a new window]
|
Fig. 5.
The P2Y10 promoter contains
a consensus PU.1/Spi-B DNA site. A, sequence of both
strands of the P2Y10 promoter near the transcriptional start
site is shown with the GGAA core of the PU.1/Spi-B binding site
boxed. Also shown is one strand of the consensus sequence
defined in vitro for PU.1 and Spi-B (36) and the SV40 PU.1
box (65). Of note, the PU.1/Spi-B site is found where a TATA box is
classically located. B, EMSA using nuclear extracts prepared
from the mature B-cell line A20. The probes used are described under
"Experimental Procedures." PI indicates preimmune serum for the
anti-Spi-B antiserum, NS is a nonspecific antibody used as a negative
control. C, EMSA using primary B-cell nuclear
extracts.
|
|
To test whether the putative PU.1/Spi-B-binding site actually binds
these factors in vitro, EMSAs were performed using NE prepared from A20 cells, a mature (IgG+) B-cell line. As
demonstrated in Fig. 5B, a 32P-labeled probe
containing the PU.1/Spi-B site from the P2Y10 promoter produced a
single DNA-protein complex when incubated with A20 NE (lane
2). This complex was specific to the PU.1/Spi-B-binding site since
it was not affected by a nonspecific cold competitor to the GAL4
transcription factor-binding site (lane 3) or a cold competitor P2Y10 probe in which the PU.1/Spi-B-binding site was mutated
(GAGGAA 
CACCGG, lane 5). In contrast, the A20 complex was efficiently competed by either the unlabeled P2Y10 probe
(lane 4) or an unlabeled probe to the B site from the
2-4 enhancer which has been shown previously (22) to
efficiently bind both PU.1 and Spi-B (lane 6). To confirm
that PU.1 and/or Spi-B were present in the DNA-protein complex,
supershift experiments were performed using antisera to both PU.1 and
Spi-B. Antiserum to Spi-B caused a small increase in the mobility of
the DNA-protein complex (lanes 8 and 9) that was
not observed in the preimmune sera, and an antibody to PU.1 reduced the
mobility and intensity of the complex (lane 10).
Importantly, addition of antisera to both PU.1 and Spi-B completely
abolished the shift (lane 11), implying that the binding
activity in A20 nuclear extracts to the P2Y10 promoter
contains almost exclusively PU.1 and Spi-B. The most likely explanation
for the increase in mobility seen with the anti-Spi-B antiserum
(lane 9) or the decrease in mobility seen with the anti-PU.1
antiserum (lane 10) is that blocking the binding of either
PU.1 or Spi-B allows for better binding of the other transcription
factor to the site. Since Spi-B·DNA complexes migrate more slowly
than PU.1·DNA complexes (Fig.
6B), the antisera are
affecting mobility simply by allowing only one of the two transcription
factors to form a protein-DNA complex.
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 6.
IVT PU.1 and Spi-B both bind to the P2Y10
site specifically and with comparable affinity. A,
equimolar amounts of IVT PU.1 and Spi-B were used in EMSA. The probes
and competitors are identical to those used in Fig. 5. B,
equimolar amounts of IVT PU.1 and Spi-B were used in EMSA where
radiolabeled P2Y10 probe was present in excess and then increasing
amounts of unlabeled probe were added to compete the radiolabeled
probe:DNA complex. C, the percent of maximal
(i.e. no competitor) radiolabeled DNA-protein complex is
graphed as a function of cold competitor concentration
(nM).
|
|
To explore the binding activity of PU.1 and Spi-B from primary cells,
NE were prepared from purified splenic B-cells and used in an EMSA. As
demonstrated in Fig. 5C, a probe to the P2Y10 site generates
a specific protein-DNA complex from these NE (lanes 2-6).
This complex contains exclusively PU.1, as evidenced by the complete
ablation of shift using an anti-PU.1 antibody (lane 10) and
is essentially unaffected by the anti-Spi-B antiserum (lane
9). Thus, in both A20 and primary B-cell NE, PU.1 is the predominant component of the DNA binding activity for the P2Y10 site.
To confirm that PU.1 and Spi-B could bind to the Ets site in the
P2Y10 promoter, equimolar amounts of IVT PU.1 and Spi-B were used in EMSAs. As shown in Fig. 6A, both PU.1 and Spi-B
efficiently bound the P2Y10 site (lanes 2 and 3).
Binding to the P2Y10 site was not competed by either a nonspecific cold
competitor probe (GAL4, lanes 4 and 5) or an
unlabeled P2Y10 probe containing a mutated PU.1/Spi-B site (lanes
8 and 9) but was efficiently competed by unlabeled
P2Y10 probe (lanes 6 and 7) or the B site
(lanes 10 and 11).
Our EMSA results with B-cell NE (Fig. 5, B and C)
imply that Spi-B did not bind to the P2Y10 site at significant levels
either because it has weak affinity for the site or because it is
expressed at much lower levels than PU.1. To directly assess the DNA
binding affinities of PU.1 and Spi-B for the P2Y10 site, we performed EMSA under conditions where the radiolabeled P2Y10 probe would saturate
the binding of either PU.1 or Spi-B and added increasing amounts of
unlabeled probe (Fig. 6B). The reduction in radiolabeled DNA-protein complex was then quantified to generate a binding affinity
curve for the two proteins (Fig. 6C). Based upon these results, PU.1 and Spi-B have a comparable binding affinity for the
P2Y10 site (Kd 
3 nM).
Therefore, the lack of Spi-B binding activity seen in B-cell extracts
is most likely because it is expressed at much lower levels than PU.1.
Unfortunately, no antibody for Western blotting of murine Spi-B is
currently available, making it impossible to directly assess the
protein levels of PU.1 and Spi-B in B-cells.
To directly address whether this promoter fragment depends on an intact
PU.1/Spi-B site for proper transcription, the 0.6-kb fragment was
subcloned into a reporter construct expressing the human growth hormone
reporter gene with either a wild-type or a mutant PU.1/Spi-B-binding
site (GAGGAA CACCGG). Transient transfections with the wild-type
(0.6 kb wt) and mutant promoter (0.6 kb mut) constructs were performed
using three hematopoietic cell lines: EL-4 (immature T) and A20 (mature
B) which both normally express P2Y10, and MEL (erythroid) which does
not express P2Y10. As indicated in Fig.
7A, the 0.6-kb promoter
fragment was not active in MEL cells, and only weakly active in EL-4
cells, although this activity was not affected by a mutation in the
PU.1/Spi-B site. Because both PU.1 and Spi-B are not expressed in
T-cells it is likely that other transcription factors are responsible for P2Y10 transcription in this lineage (see Fig.
4C). The Ets factor, Elf-1, is present in T-cells and binds
to similar DNA elements (2), making Elf-1 a candidate for regulating
T-cell-specific P2Y10 expression. However, this seems less likely given
that the Pu.1Spi-B-binding site mutation failed to decrease promoter
activity in EL-4 cells. A more plausible explanation is the existence
of a second promoter that is active in T-cells and independent of PU.1
and Spi-B. In direct contrast, the 0.6-kb promoter was extremely active
in A20 B-cells and mutation of the PU.1/Spi-B-binding site decreased
its activity by 50-75%. Thus, the 0.6-kb P2Y10 promoter is lineage
restricted to B-cells and the activity of this promoter depends on a
functional PU.1/Spi-B-binding site.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 7.
The PU.1/Spi-B binding site is functionally
important for P2Y10 promoter activity.
A, transient transfections were performed using a 0.6-kb
fragment of the P2Y10 promoter with either a wild-type (0.6 kb wt) or mutant (0.6 kb mut) PU.1/Spi-B site to examine its activity
in three hematopoietic cell lines. All data represent mean fold
induction over the promoterless reporter p GH for four independent
experiments (±S.E.). B, PU.1 and Spi-B can transactivate
the P2Y10 promoter in an NIH 3T3 cell. Epitope tagged
versions of PU.1, Spi-B, or a deletion mutant of Spi-B containing no
Ets domain (Spi-B Ets) or no transactivation domain (Spi-B TA)
were transiently transfected into 3T3 cells and assayed for reporter
activity of a 0.1-kb promoter fragment with a wild-type
PU.1/Spi-B-binding site. All data represent mean fold induction over
the promoterless reporter pGH for three independent experiments
(±S.E.).
|
|
To further examine the role of PU.1 and Spi-B in transcription of the
P2Y10 promoter, co-transfection experiments were performed by transiently transfecting NIH 3T3 cells with a smaller reporter construct (0.1 kb) containing the wild-type PU.1/Spi-B-binding site and
expression plasmids containing epitope tagged versions of PU.1, Spi-B,
and two transcriptionally inactive deletions of Spi-B which either
remove the DNA-binding domain (Spi-B 
Ets) or the transactivation
domain (Spi-B TA). Similar results were obtained with the 0.6 kb wt
reporter construct (data not shown). As shown in Fig. 7B,
either PU.1 or Spi-B caused strong transactivation of the reporter in
this heterologous cell type. This effect is specific since Spi-B Ets
and Spi-B TA caused no transactivation of the reporter.
Interestingly, the same promoter with a mutant PU.1/Spi-B was weakly
transactivated by PU.1 or Spi-B (data not shown), implying that there
may be other low affinity binding sites for these transcription factors
in this DNA fragment. Transactivation of the P2Y10 0.6-kb
promoter was also examined by reconstituting NIH 3T3 cells with both
PU.1 and Spi-B. We found that transactivation by PU.1, Spi-B, or a
combination of the two factors resulted in statistically similar levels
of P2Y10 promoter activity (data not shown). These results
are consistent with a model where P2Y10 is regulated by both
PU.1 and Spi-B.
We conclude that the putative PU.1/Spi-B site in the P2Y10
promoter efficiently binds both proteins, and this binding site is
critical for P2Y10 transcription in B-cells. In addition,
the P2Y10 promoter can be transactivated by PU.1 or Spi-B in
heterologous cells. Given that
PU.1+/+Spi-B/ B-cells
express 50% as much P2Y10 as wild type cells and that PU.1+/Spi-B/
B-cells express 10-13% as much, we suggest that PU.1 is the dominant factor. Importantly,
PU.1+/Spi-B+/+ B-cells
express wild type levels of P2Y10 (data not shown). Therefore, Spi-B is
necessary for 100% levels of P2Y10 transcription. Taken together, PU.1 and Spi-B regulate P2Y10 and it is likely to
be a direct transcriptional target for these factors.
| |
DISCUSSION |
PU.1/ mice fail to develop any mature
B-cells (25, 28). Therefore, there is no way to assess the role of PU.1
in B-cell function without a conditional mutation in B-lymphocytes (37, 38). To explore the role of PU.1 in mature B-cells, we created a
semi-epistatic condition in which B-cells expressed 50% of wild type
PU.1 protein levels and lacked Spi-B
(PU.1+/Spi-B/).
B-cells from
PU.1+/Spi-B/ mice
contained sufficient PU.1 protein to support B-lymphocyte development
but insufficient PU.1/Spi-B binding activity for efficient transcription of all target genes.
PU.1+/Spi-B/ B-cells represent
a tractable model system for analysis of these factors in
vivo because they are phenotypically normal when unstimulated (31)
but display signaling defects upon IgM cross-linking. As such,
PU.1+/Spi-B/ and
PU.1+/+Spi-B+/+ purified
B-cells provide attractive substrates for subtractive hybridization,
with the only difference in the two populations being disruption of the
PU.1 and Spi-B loci. Our data indicate that PU.1
and Spi-B are necessary for efficient transcription of membrane
proximal component(s) of the BCR signaling cascade. Importantly, the
expression of several presumptive target genes of PU.1 and Spi-B such
as mb-1, Btk, Blk, , and light
chains and all known components of BCR signaling appear unaffected by reduced levels of PU.1 and Spi-B (31). Therefore,
PU.1+/Spi-B/
B-cells could be used to clone novel PU.1/Spi-B targets important for
BCR-mediated responses.
We show in this report that PU.1 and Spi-B are both necessary for
efficient transcription of P2Y10 in B-cells. It is possible that PU.1 and Spi-B also regulate another transcription factor required
for the proper expression of P2Y10. However, we have cloned the
P2Y10 promoter and demonstrated that it contains a consensus
PU.1/Spi-B-binding site important for promoter activity in B-cells.
While these data cannot exclude the possibility that PU.1 and Spi-B
also regulate additional gene(s) required for P2Y10 expression, our
data illustrate that PU.1 and Spi-B directly control P2Y10
promoter activity. Other possible low affinity PU.1/Spi-B sites in the
proximal P2Y10 promoter (Fig. 7B) may explain why the single mutation created in our reporter constructs (Fig. 7, A and B) does not completely abolish
transcription in transient assays.
The importance of PU.1 and Spi-B for P2Y10 transcription is
supported by their convergent expression patterns. In B-cells, P2Y10
expression essentially parallels that of Spi-B in that both are
preferentially transcribed in immature and mature B-cells, whereas PU.1
is expressed throughout B-cell development. Importantly, one of the two
pre-B cell lines producing P2Y10 (38 B9) also expresses low levels of
Spi-B (9), and the other pre-B cell line lacks both P2Y10 (70 Z/3) and
Spi-B. These data imply that Spi-B is important for P2Y10 expression
(Fig. 4B). However, EMSA shows that PU.1 is the predominant
DNA binding activity contained in A20 or primary B-cell nuclear
extracts (Fig. 5, B and C). Based on these
results and mRNA levels of P2Y10 from
PU.1+/+Spi-B/ and
PU.1+/Spi-B/
purified B-cells (Fig. 2), P2Y10 transcription in B-cells is dependent on the combined dosage of PU.1 and Spi-B, although PU.1 is
more critical because it is expressed at higher levels than Spi-B.
Transient transfection experiments in NIH 3T3 fibroblasts also support
a combined role for PU.1 and Spi-B (Fig. 7). Our data are similar to
results reported for bHLH proteins of the E2A family where
B-cell development is affected (39). In this situation, combinations of
mutations in E2A, E2-2, and HEB revealed that a
critical dosage of all three genes is required for efficient B-cell
formation and development. However, in our situation the dosage of PU.1
and Spi-B seems to effect the transcription of a specific target gene.
One possible reason that P2Y10 transcription depends on
adequate levels of PU.1 and Spi-B is that its proximal promoter
contains a paucity of binding sites for other known transcription factors (data not shown). PU.1 and Spi-B are likely to be scaffolding proteins which bind DNA elements and create activation complexes that
recruit the basal transcriptional machinery to promoters through
multiple protein-protein interactions (40-44). When other factors are
present, such protein-protein interactions stabilize PU.1 and/or Spi-B
binding and low levels of PU.1/Spi-B are sufficient to create effective
transcriptional activation complexes. The lack of binding sites for
other transcription factors in the P2Y10 promoter could
reduce the stability of activation complexes, requiring maximal levels
of PU.1 and Spi-B for proper activation.
Another explanation for the observed gene dosage effects could be the
presence of additional Ets sites outside of the 0.6-kb P2Y10
promoter that preferentially bind either PU.1 or Spi-B which are
critical for P2Y10 transcription. Functional Ets sites in B-cell promoters or enhancers which selectively bind PU.1 or Spi-B have
not been identified thus far. However, an Ets site in the myeloid
cell-specific c-fes promoter appears to bind PU.1 at a significantly higher affinity than Spi-B (22). The loss of PU.1 or
Spi-B would selectively reduce P2Y10 transcription through the activity of such DNA elements.
The ultimate goal of our studies was to understand how PU.1/Spi-B
target genes contribute to the phenotype of mutant mice. The
PU.1+/Spi-B/ B-cell
defect is striking: they exhibit poor proliferation in response to
antigenic stimulation both in vivo and in vitro,
decreased levels of substrate tyrosine phosphorylation, and a blunted
calcium response upon IgM cross-linking. The most likely cause of
reduced BCR signaling is decreased Syk phosphorylation of its
downstream targets such as phospholipase C-
2 and BLNK (31). Could
P2Y10 be the in vivo target gene responsible for the
PU.1+/Spi-B/ B-cell
signaling defect? Classically, membrane proximal BCR signaling events
are thought to involve tyrosine phosphorylation, while activation of
heterotrimeric G-proteins and calcium fluxes act downstream. Like other
heptahelical receptors, P2Y family members couple to heterotrimeric
G-proteins (usually G
q) and eventually lead to PLC
activation (45). However, recent evidence indicates that Btk, a
critical tyrosine kinase in BCR signaling and B-cell development (46,
47), is activated by both Gq and G subunits (48-54). Therefore P2Y10 could activate Btk via G-proteins, thereby influencing the tyrosine phosphorylation of membrane proximal components of BCR signaling.
If P2Y10 does influence BCR signaling, it must become activated by
binding its cognate ligand. Unfortunately, the P2Y10 ligand remains
unknown, and it may not be the classical ligand for most purinergic
receptors: ATP/ADP or UTP/UDP (55-60). The identity of the ligand for
P2Y5, the most closely related member of the P2Y family (32%
identity), is still unclear (61, 62), and another family member, P2Y7,
has been shown to bind leukotriene B4 (63, 64). Once the
natural ligand or a chemical agonist of P2Y10 is discovered the role of
this receptor in BCR signaling can be directly assessed.
This study represents a unique example of gene targeted mice as a model
to identify in vivo target genes of two transcription factors which contribute to an observed lymphoid phenotype. P2Y10 appears to be a direct transcriptional target of both PU.1 and Spi-B
in vivo and provides an interesting model of transcriptional activation by this subgroup of Ets proteins. In addition, P2Y10 could
have important implications in coupling heterotrimeric G-proteins directly with membrane proximal BCR signaling events. It will be
important to explore the transcriptional regulation of P2Y10 as well as its effects on antigen receptor signaling in B-cells.
§,
**
TOP
ABSTRACT
INTRODUCTION
