| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
/Dad1 Gene Locus1
*
Department of Biological Sciences, City University of New York, Hunter College, New York, NY 10021; and
Department of Molecular and Cell Biology, Cancer Research Laboratory, and Division of Immunology, University of California, Berkeley, CA 94720.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
/TCR
gene segments and the ubiquitously
expressed Dad1 gene. Expression of a transgene
under the control of all the LCR elements is T cell specific. However,
a subfragment of this LCR is functional in a wide variety of tissues.
How a ubiquitously active element can participate in tissue-restricted
LCR activity is not clear. In this study, we localize the ubiquitously
active sequences of the TCR- LCR to an 800-bp region containing a
prominent DNase hypersensitive site. In isolation, the activity in this
region suppresses position effect transgene silencing in many tissues.
A combination of in vivo footprint examination of this element in
widely active transgene and EMSAs revealed tissue-unrestricted factor
occupancy patterns and binding of several ubiquitously expressed
transcription factors. In contrast, tissue-specific, differential
protein occupancies at this element were observed in the endogenous
locus or full-length LCR transgene. We identified tissue-restricted
AML-1 and Elf-1 as proteins that potentially act via this element.
These data demonstrate that a widely active LCR module can synergize
with other LCR components to produce tissue-specific LCR activity
through differential protein occupancy and function and provide
evidence to support a role for this LCR module in the regulation of
both TCR and Dad1 genes. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
gene exists in a complex genomic locus on mouse
chromosome 14 (see Fig. 1
A). This locus contains three
differentially regulated genes, TCR, TCR,
and Dad1 (1). The first two genes are
components of 
and 
T cell Ag receptors, respectively.
Early in development, TCRs are critical for proper development and
differentiation of T cells. In mature T cells, TCR signaling initiates
immune responses. Dad1 was initially described as an
antiapoptosis gene (2) and was subsequently found to be
homologous to a subunit of mammalian oligosaccharyl-transferase
(3). Germline deletion of Dad1 results in an
embryonic lethal phenotype (4, 5, 6). Its function in the
immune system is poorly understood. Dad1 is up-regulated
during T cell development and is highly expressed in mature T cells.
Overexpression of Dad1 in mature T cells results in
hyperproliferation in response to TCR stimulation (7),
suggesting that it might play a role in modulating T cell responses.
Transcriptional regulation of this locus is highly interesting.
TCR and TCR genes are expressed by the
and lineages of T cells, respectively. In contrast,
Dad1 is expressed ubiquitously (1). In addition
to the appropriate V-D-J rearrangements that generate functional TCR
and TCR chains, this locus is often involved in aberrant gene
rearrangements that can lead to T cell malignancies (8).
Therefore, the mechanisms controlling transcription and chromatin
structure/accessibility in this locus are of great interest.
|
and
Dad1 genes (11). In T cell lines, the TCR
LCR manifests itself as nine DNase I hypersensitive sites (HS) that
extend from the TCR-chain constant region exons to 3 kb upstream of
the 3' end of the Dad1 gene (which is in the opposite
transcriptional orientation from that of TCR) (11). HS1
maps to the heavily studied TCR enhancer region (12, 13). HS7 and HS8 are located 5' of the enhancer, whereas HS2–6
are 3' of the TCR enhancer. The nine HS of the LCR confer
high-level, position-independent, copy number-dependent, and
tissue-specific expression upon a linked transgene in chromatin of mice
(14). HS7 and HS8 appear to be dispensable for complete
LCR activity. HS1 contributes to high-level expression in the thymus
and, thus, is required for strict copy-number dependence of linked
transgene activity (15). Just 3' of the enhancer is HS1'.
This more recently characterized region plays a major role in the T
cell-specific functions of this LCR. It does not contain any classical
transcriptional enhancer activity but affects chromatin structure, DNA
methylation events, and the tissue distribution of LCR activity in
transgenic mice (15, 16). HS2 through HS6 are absolutely
required for complete, T cell-specific LCR activity (11).
In isolation, this TCR distal portion of the LCR also contains an
unusual activity that drives widespread transgene expression in mice
(14, 15). This expression is roughly, but not strictly,
related to copy number indicating the presence of elements in this
region that counteract integration-site "position-effects" in many
tissues. Furthermore, the ubiquitous activity of this region suggested
a role for it in the regulation of the nearby Dad1
gene. We aim to define the in vivo molecular requirements for the establishment of a T cell-specific gene expression program in chromatin. We also seek to understand how three separate gene expression programs are successfully coordinated within this immunologically important locus. Here we localize the sequences responsible for the activity in HS2–6 and characterize factors interacting with these sequences. Using transgenic analyses, we find that HS2–6 function resides in an 800-bp fragment containing HS6. Removal of this region from a reporter transgene makes transcription of that construct highly susceptible to position-effects in mice. Its presence stabilizes the expression level and ubiquitous tissue distribution of a transgene.
We additionally sought to explain how this widely active LCR module
synergizes with other LCR components to produce T cell-specific LCR
activity. To this end, we analyzed DNA sequences at HS6 by in vivo
footprinting (IVFP). In the endogenous locus, the pattern of factor
occupancy at HS6 sequences is different between lymphoid and
nonlymphoid organs. These tissue-differential patterns are reproduced
on full-length LCR-containing, T cell specifically active transgenes in
mice. Transgenes driven by HS2–6 alone are widely active. In this
context, the IVFP patterns at HS6 are similar in lymphoid and
nonlymphoid organs. Therefore, the synergy between the HS6 LCR module
and upstream LCR elements is accompanied by tissue-specific changes in
factor occupancy at HS6 sequences. In vitro assays of factor binding to
sites identified by IVFP revealed a tissue-restricted DNA binding
complex containing the Elf-1 protein and a second tissue-restricted
complex containing an AML-1-related protein. In addition, two
tissue-unrestricted, sequence-specific DNA binding complexes
interacting with these IVFP sequences were revealed. These data show
how a ubiquitous element can participate in tissue-specific gene
expression and implicate the Elf-1- and AML-1-related proteins in the T
lineage-specific activity of the TCR LCR.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
DNA fragments for microinjection were doubly purified by gel
electrophoresis on low-melting-point agarose (Sea Plaque GTG;
BioWhittaker, Rockland, ME) followed by digestion with -agarase (New
England Biolabs, Beverly, MA). DNA was microinjected into the
pronucleus of (C57BL/6 x CBA)F1 fertilized
mouse eggs, and transferred into pseudo-pregnant CD1 foster mothers.
Transgenic founders were identified by Southern blot analysis on tail
DNA. The founders were outcrossed to C57BL/6 mice, and heterozygous
transgene-positive offspring from these crosses were analyzed. Relative
copy number was determined for each line by analysis of at least two
Southern blots by PhosphorImager (Molecular Dynamics, Sunnyvale, CA).
All lines directly compared in this work were analyzed for relative
copy number on the same Southern blot using the same probe and enzyme
digestion. The signal from endogenous TCR locus was used as a
normalizing control.
DNA constructs
The construction of the 9-kb HS1–8 fragment, 5.9-kb HS2–6
fragment, and the :1–8 and :2–6 transgenes have all been
previously described (14). For the :6 construct, a
1.6-kb MfeI/SacI fragment of the HS1–8 region
was excised from the pSP72:HS1–8 vector. This fragment was cloned into
the previously described pSP72 vector (Promega, Madison, WI) containing
the 4.9-kb BglII human -globin fragment in a position 3'
of the transcription unit. :
6 was similarly constructed using a
0.8-kb BglII/SacI fragment of the LCR. Transgenic
inserts for microinjection were liberated from vector DNA using
XhoI and ClaI.
RNA analysis
RNA was prepared according to the "one-step" protocol
(17) from transgenic mouse tissues that were dissected of
fat, minced (except for thymus and spleen), and rinsed extensively with
PBS to minimize contaminating blood. Five micrograms of RNA samples was
used in RNase protection assays. RNA probes were labeled with
32P-GTP and SP6 RNA polymerase as follows: For
-globin, a 2.0-kb BamHI fragment spanning exons 1 and 2
was cloned into pSP72 in the opposite orientation with respect to the
SP6 promoter. The plasmid was linearized with AvaII to
generate an RNA probe to exon 2. For -actin (18),
plasmid was made and linearized with HinfI. The resulting
RNA probes were purified by acrylamide gel electrophoresis before
hybridization. Absolute numbers reported for mRNA expression levels are
normalized to internal loading controls and quantified within the
experiment presented. As differences in RNA probe preparation and RNase
digestion conditions are sometimes unavoidable between separate
experiments, comparison of absolute expression levels between
individual points in different experiments is not valid.
In vivo DNase I footprint analyses
Nuclei from liver (19) and thymus (20) were prepared and resuspended in DNase I digestion buffer (21) at 108 nuclei/ml. Nuclei were digested for 10 min on ice. The amount of DNase I (Worthington Biochemical, Lakewood, NJ) used in the digestion ranged from 0.0 to 4.0 mg/ml. For the plain genomic DNA samples, 0.25–1 µg/ml DNase was used. Digestion was stopped with a 1/10 volume 5% SDS/100 mM EDTA. The genomic DNA was purified by two rounds of phenol extraction, one round of chloroform/isoamyl alcohol extraction followed by ethanol precipitation. The DNA was then subjected to ligation-mediated (LM)-PCR following exactly the protocol described in Ref. 22 . Gene-specific oligonucleotides (Genset, San Diego, CA) were selected, synthesized, and acrylamide gel purified. The products of these reactions were separated on 5% denaturing acrylamide-urea sequencing-type gels. Dimethyl sulfate (DMS) control ladders were prepared exactly as described in the LM-PCR support protocol in Ref. (22). LM-PCR was performed simultaneously on the DMS-treated genomic DNA and the experimental samples.
Nuclear extracts and EMSA
Nuclei were prepared as they were for IVFP experiments described above. Nuclear pellets were resuspended in extraction buffer C (23) and rocked for 30 min followed by microcentrifugation at top speed. The supernatants were frozen in aliquots at -80°C. Nuclear extract (0.5–3 mg) was used in EMSA with 40,000 cpm (dry) of 32P-labeled oligonucleotide probe. The final binding reaction conditions were as follows: 15 mM HEPES pH 7.9, 80 mM NaCl, 15 mM KCl, 0.02 mM EDTA, 1 mM DTT, 1 mM PMSF, and 3% glycerol. Incubations were on ice for 45 min. For binding site competition assays, a 50- to 100-fold molar excess of unlabeled oligonucleotide was added to the binding reaction before the addition of labeled probe. The sequences of the oligonucleotide probes used are as follows: thymic footprint (TF)1-TGCCGTGGCGACAGGAAGTG, TF2-TGTACAGTAGTTGTGGTAAATG, TF3-AGCTTCCACAGATTGAACACAGGAAATA, core binding factor (CBF)-TCGACTCCCGCAGAAGCCACATG, ETS1-TCGACCTCTGGAAAGAGGAG, GATA3-TCGAGTAGAGATAAGATC, E-BOX-TCGAGGGCCACGTGCCCAG, GT-BOX-GCAGAGGTGGGTGGAGTTTCG. For Ab supershift/blocking assays, 2–6 µg of Ab were added to the binding reaction, 15 min into the incubation. The following Abs were obtained from Santa Cruz Biotechnology (Santa Cruz, CA): anti-AML-1 N-20 (sc-8563), anti-AML-1 C-19 (sc-8564), anti-Elf-1 (sc-631), and anti-Ets-1/Ets-2 (sc-112).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
LCR
We used a 4.9-kb BglII fragment of the human
-globin locus (10, 24, 25) as a reporter
gene in our previous LCR analyses (14, 15, 16). The globin
transgene under the control of HS1–6 is expressed T cell specifically,
whereas HS2–6 linked transgene is expressed in a wide variety of
tissues. Continuing in this system, we generated 5' deletion mutants of
HS2–6 (Fig. 1
B). Our studies
focused on HS6 as it is the strongest HS in this region of the
endogenous LCR. We hypothesized that it would be the major contributor
to the function of this region. Two new transgenic reporter constructs
were created to test this hypothesis. In :6, HS2–5 have been
deleted from the previously described :2–6 transgene. This
transgene consists of the -globin transcription unit linked to a
1.6-kb piece of the LCR that contains HS6. The second new transgene
contains an 800-bp fragment at the very 3' end of the LCR linked to
-globin. This LCR fragment contains no HS and is called :6.
Transgenic mice were generated with these new reporter constructs. Four
independent lines of :6 and five lines of :6 transgenic mice
were generated and analyzed. These lines were compared with lines of
:2–6 mice that were generated and described previously
(14) to determine the sequences contributing to the
activity of the HS2–6 region.
The major activity of the HS2–6 region is contained in HS6
The transcriptional activity of the reporter transgenes was
analyzed using the RNase protection assay as described in
Materials and Methods. Fig. 2
shows the tissue distribution of transgene activity in three
independent lines of :6 mice. Transgene expression is similarly
widespread in all lines bearing this construct. The :6 expression
pattern is reminiscent of that of the :2–6 transgene we analyzed
previously (14). This indicates that the region containing
HS2–5 is dispensable for the ability of this region to drive
ubiquitous transgene expression. We also carefully examined the
expression levels of the four lines of :6 in comparison to that of
:2–6 lines in the various organs. Fig. 3 shows PhosphorImager analyses of these
experiments (plotted on a log scale). We were unable to discern any
significant differences in the expression level per copy of :6 vs
:2–6. :6 expression appeared to be moderately higher than
:2–6 activity in the lymphoid organs (thymus and spleen). However,
in general, the differences in expression levels between the two
different constructs were no larger than the differences in the levels
seen within the group of lines carrying the same construct. Therefore,
we must conclude that HS2–5 may not significantly contribute to
transgene expression in this context. This, of course, does not rule
out a role for these HS in the context of the full-length LCR.
|
|
A classical measure of position-effects in chromatin involves the
examination of reporter gene expression levels in a single organ of
multiple lines of transgenic mice. Such a position-effect would
silence, or severely repress transgene expression in a proportion of
independent lines bearing the same construct (26). We
examined the ability of HS6 to protect the transgene from these
expression-level position-effects. Fig. 4A shows RNase protection data
from the thymus and liver of multiple independent transgenic mouse
lines bearing either the :6 or the :6 construct. Expression is
evident in the thymus of all four lines shown. As with the :2–6
construct, the expression in thymus is not copy number-dependent but,
rather, plateaus with increasing copy number. A more copy-related
expression pattern is seen in the liver of :6 mice. This degree of
copy-related expression was also seen in the :2–6 construct
(14).
|
:6 mice, the lines of :6 mice did not
display consistent expression of the construct in these organs. In
fact, thymic transgene activity was only evident in two of the five
lines. These same lines were the only ones that displayed high-level
expression in the liver. The expression levels in these two lines are
even higher than those seen in the :6 mice. This indicates that the
transgene is being subjected to a positive position-effect in these
lines. It is interesting to note that these two lines are derived from
the same founder, in which two separate integration events were
apparent. An outcross of the founder segregated the two integration
sites, thus establishing lines 23A and 23B. The similar transgene
expression evident in only these two lines suggests that the two
integration sites may be related. This founder transmitted the
transgene to all of its offspring as if the integration sites were on
allelic chromosomes. Furthermore, no cosegregation of the integration
sites was evident in these litters. We speculate that perhaps the two
integration events took place in similar positions on opposite
chromosomes.
Further analysis of :6 and :6 mice in other organs showed
similar results as those in the thymus (Fig. 4B). The
activity of HS6 is widespread and not restricted to lymphoid organs. To
quantify the degree to which HS6 suppresses the variation in transgene
expression levels (per copy), PhosphorImager analysis of these RNase
protection experiments were conducted and plotted on a log scale (Fig. 4B). In the :6 mice, the values of expression level per
copy varied over a range of 11-fold in the thymus, 4-fold in spleen,
5-fold in heart, and 6-fold in liver. These values are well within the
range described for the activity of a "partial-LCR"
(27). In contrast, in the :6 mice, the range of
variation in expression level per copy was over 110-fold in thymus,
190-fold in spleen, 56-fold in heart, and 150-fold in liver. Therefore,
the activity contained in the 800-bp region of HS6 suppresses
expression-level position-effects by over 10- to 40-fold in the various
organs.
Sequences of the HS6 region of the full-length TCR LCR are
differentially occupied by NFs in vivo
The demonstration that the HS6 region, in isolation, has a
widespread activity raises the question of how it can participate in
the T cell-specific activity of the TCR LCR. We hypothesized that in
the context of the full-length LCR the activity of this region was
altered to promote a more tissue-restricted function of HS6. We thought
an examination of NF occupancy at HS6 in the complete LCR might provide
evidence to support this hypothesis. To accomplish this, in vivo DNase
I footprinting analyses of HS6 sequences in the endogenous gene and
full-length LCR transgenes were conducted. Mapping of HS6 with respect
to nearby restriction sites placed the 5' border of HS6 just downstream
of an SphI site (data not shown). Fig. 5A shows data from this region
of the endogenous TCR LCR. Three regions of clear differences in
DNase I digestion patterns are evident between nuclei of thymus
(TCR+, Dad1+) and liver
(TCR-, Dad1+). These
regions are labeled TF1, TF2, and TF3. TF1 appears as a strong HS at
one nucleotide in thymus, followed by a small stretch of missing
fragments (which are strongly present in plain genomic DNA). This HS is
clearly absent in the liver, although the following nucleotides are
still absent from the ladder. TF2 also manifests itself as a strong
thymic HS followed by a clearing in the ladder (in comparison to the
genomic DNA). TF3 is also a clear region of differential cleavage
between thymus nuclei, liver nuclei, and genomic DNA. In general, the
region between TF1 and TF3 appears to be less accessible to the
nuclease in liver than in thymus. This suggests an overall differential
chromatin structure in the region separate from more localized
differences in TF1, TF2, and TF3. The localized cleavage pattern
changes further suggest tissue-specific differences in the NF binding,
particularly at the TF2 and TF3 sequences.
|
B). The :1–8 transgene contains the -globin reporter
linked to the complete LCR (Fig. 1B). IVFP analysis of a
high-copy :1–8 transgenic mouse is shown. The TF1, TF2, and TF3
regions of HS6 detected in :1–8 transgenic thymus and liver appear
identical with those seen in the endogenous locus. Perhaps due to the
excess of transgene copies available for DNase digestion in these
organs, the generalized difference in nuclease accessibility between
thymus and liver seen in the endogenous locus appears less evident in
:1–8. However, this change now makes the "footprint" at TF2
even clearer. These data confirm the evidence for tissue-differential
NF binding at HS6 sequences of the full-length LCR in vivo. Differential IVFPs at HS6 are not evident in the context of the widely active, partial LCR
We previously reported that removal of the four 5'-most HS of the
LCR (HS7, HS8, HS1, and HS1') from a transgene liberated the widespread
activity in the distal LCR region containing HS2-HS6. As the above
experiments indicate that HS6 is, at least, a major contributor to this
activity, we wanted to see whether the change in the activity caused by
LCR truncation is reflected in the IVFP patterns at HS6 in this mutant
LCR. The :2–6 transgene contains the human -globin transcription
unit linked to the partial LCR, HS2–6 fragment (Fig. 1B).
Fig. 5B shows the results of IVFP on thymus and liver of a
high copy :2–6 transgenic mouse alongside the results from the
full-length LCR (:1–8) transgene. The stark differences in the
endogenous footprint in the TF1, TF2, and TF3 regions are not present
in the truncated LCR transgene. HS6 IVFP patterns in both organs now
look similar to that seen in the full-length LCR in liver. These data
suggest that the LCR truncation that results in a change from T
cell-specific to widespread function of this region is accompanied by
an alteration of the NFs bound to sequences in the HS6 region.
TF1, 2, and 3 interact with several tissue-restricted and unrestricted sequence-specific DNA binding complexes
To examine the protein factors interacting with TF1, TF2, and TF3
DNA, we designed oligonucleotide probes representing these sequences
for use in EMSA. We prepared nuclear extracts from mouse thymocytes and
fibroblasts and incubated them with the various probes (Fig. 6A). TF1 bound a single
complex that appears identical in both tissue types. TF2 bound two
complexes in thymus (1T and 2T) and two complexes in fibroblast
extracts (1F and 2F). The 2T and 2F complexes appear equivalent.
However, complexes 1T and 1F are not similar with 1T migrating faster
than 1F. TF3 binds a strong, distinct thymocyte-derived complex that is
not present in the fibroblasts. Several minor bands are present in both
tissues. As a control for nuclear extract integrity, a GT box (Sp1
family protein binding) probe was also used in these experiments
(28). This probe shows the equivalent quality of the
nuclear extracts prepared from both cell types. EMSA competition
experiments demonstrated that all the complexes bind in a
sequence-specific fashion (Figs. 7, 8, A and B, and
9A). In addition, TF1 complex
was not competed for by any of the non-TF1 consensus binding
sites investigated, including CBF, E box, Ets, GATA-3, or GT box
elements (Fig. 7). Interestingly, the upper complexes formed on TF2 DNA
from both thymocytes and fibroblasts were competed for by a CBF
consensus binding site (29) (Fig. 8A). The
lower band was less efficiently competed for by an oligonucleotide
containing a GATA-3 site from the TCR enhancer. However, the TF2
oligonucleotide does not contain a GATA-3 binding site (Fig. 6B). Rather, there is a fortuitous sequence homology between
the GATA-3 probe and TF2 at an AGTAGT sequence that is just 5' of the
consensus CBF binding site (TGTGGT). The TF3 binding complexes, like
those formed on TF1, were only competed for by homologous
oligonucleotide (Fig. 9A).
|
|
|
|
The EMSA competition experiments suggested a relationship between
CBF proteins and the TF2 binding complexes 1T and 1F. CBF is a
heterodimer of AML-1 (CBF2) and CBF. Using Abs to the AML-1
protein, we performed EMSA "supershift/blocking" experiments to
further identify members of TF2 binding complexes (Fig. 8C).
The mobility of complex 1T was indeed altered by an Ab to the N
terminus of AML-1 but not an Ab directed at the C terminus of AML-1.
The mobility of complex 1T was not affected by Ab to Ets family
proteins. Neither were the lower complexes (2T, 2F) affected by Ab
treatment. None of the fibroblast-derived complexes reacted with any of
the Abs tested. These data suggest that the tissue-restricted band 1T
contains a protein related to AML-1. Although the 1F complex appears to
interact with a CBF binding site in competition assays, this complex
does not contain -AML-1-reactive protein.
The thymocyte-derived TF3 binding complex contains the Elf-1 protein
Although TF3 binding proteins were not
competed with by any of the binding sites assayed, visual inspection of
the sequence revealed a purine-rich sequence resembling the binding
site for proteins of the Elf-1 branch of the Ets transcription factor
family (30). EMSA supershift experiments confirmed the
presence of a protein related to Elf-1 in the major TF3 binding complex
of thymocytes (Fig. 9B). Formation of this complex was
blocked by the presence of anti-Elf-1. In contrast, an Ab against
Ets-1 and Ets-2 proteins had no effect on the mobility of the TF3
binding complex.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
/TCR/Dad1 genomic locus
contains genes of critical importance to T cell differentiation and
function. The LCR in this locus has been hypothesized to play a major
role in the differential regulation of these genes (11).
The structure and function of the LCR has been the focus of much study.
A complete, bona fide LCR has the ability to transfer nearly all
aspects of the transcriptional regulation of its locus of origin to a
linked heterologous transgene at any site of integration. These aspects
include tissue specificity, developmental timing, and expression level
per transgene copy (31, 32). The powerful activity of the
LCR to completely protect transgenes from position-effects in
transgenic mice has been well demonstrated in many systems
(33). Although a growing handful of complete LCR
activities are being identified, recent gene-targeting data has raised
controversial questions regarding the nonredundant roles of the LCR in
the genome (34, 35, 36). Nevertheless, there is little cause
to doubt the continued importance of transgenic studies of LCRs to the
understanding of the establishment of physiological, tissue-specific
gene expression in vivo. Of equal importance is the potential
application of LCR-like elements to future gene therapy applications
that, at this stage, resemble ectopic transgenesis more than gene
targeting.
LCRs appear to be particularly important regulators of
immunologically relevant loci. Complete LCRs have also been described
in the loci for human CD2 (37), the T
cell-specific adenosine deaminase (38), the
VpreB/
5 (39), and the macrophage lysozyme
(40). Partial LCR-like activities that have been described
include the Ig heavy chain locus (41) and the
IL-2 gene (42). Much of what is known of the
structure and function of LCRs comes from the studies of those in the
human -globin (31, 43) and human CD2 gene
loci (32). Although very little sequence homology between
different LCRs has been described, a recurring theme in these studies
is the modular nature of many LCRs. For example, the -globin, CD2,
and some other LCRs contain both elements with classical
transcriptional enhancer activity and critical components that lack
such activity. These latter elements are absolutely required for
aspects of LCR function. They could only have been discovered through
transgenic analyses due to their lack of classical enhancer activity.
This group of elements is inert in the kinds of cell culture and in
vitro assays used to study cis-acting transcriptional
control elements. Nevertheless, their effect on transgene expression in
the chromatin of the whole animal is now well documented. Such elements
include HS3 of the -globin LCR (27), HSS3 of the CD2
LCR (44), and the facilitator elements of the adenosine
deaminase LCR (38). We have also recently described such
an element in the TCR LCR. The HS1' element of this LCR lacks
classical enhancer activity, yet has a profound effect on the tissue
distribution of LCR activity in vivo (15). The DNA
sequences at HS6 of this LCR also lack activity in the classical assays
for enhancer function (12). Therefore, the HS6 element,
whose activity we describe here, adds to this growing family of
nonclassical cis-acting control elements. Extensive computer
and manual inspection revealed no significant homology between this
region and the core sequences of the -globin HS3 element. Recently,
a sequence involved in the function of HSS3 of the CD2 LCR was
described (44). This sequence binds a ubiquitously
expressed factor called HMG-BOX containing protein-1
(HBP-1). This factor binds a novel TTCATTCATTCA
motif. This sequence is not present in HS6.
Here we identify protein complexes interacting with IVFP
sequences of HS6. The AML-1 and Elf-1 proteins have not yet been
implicated in the activity of any LCR. However, Elf-1 has been shown to
be important in the regulation of several T cell-expressed genes such
as TdT (45), IL-2 (30),
and CD4 (46). Elf-1 has been demonstrated to be
expressed at all stages of T cell development (47) and is
a member of the Ets family of winged helix proteins. Winged helix
proteins such as hepatocyte nuclear factor (HNF)-3 and histone
H5 have been implicated in the selective positioning of nucleosomes
(48, 49). The finding of an AML-1-related protein
interaction with HS6 is interesting for several reasons. AML-1 is
critical for hemopoiesis, as AML-1-deficient mice do not undergo
definitive hemopoiesis and die in utero (50). The AML-1
protein has already been demonstrated to be important for the activity
of the TCR enhancer (HS1) (51). AML-1 is an
"organizer" of enhancer proteins and often functions in concert
with nearby bound factors. AML-1 also interacts with several
coactivator and corepressor proteins (reviewed in Refs.
29 and 52). Examples of the former include
the p300/CBP histone acetyltransferase and ALY. Examples of the
latter include Groucho and mSin3A, which associate with histone
deacetylase activity. AML-1 is part of the "runt domain" family of
transcription factors, which include three mammalian proteins: AML-1
(Runx1), AML-2 (Runx3), and AML-3 (Runx2). All three are expressed in
the hemopoietic system. AML-3 has been demonstrated to be most
important in bone development (53). AML-2 has been
implicated in the activation of the Ig heavy chain gene in B
lymphocytes (54). Several alternatively spliced isoforms
of AML-1 have been identified (52). This may explain why
the N-terminal and C-terminal Abs to AML-1 were differentially reactive
in the supershift/blocking assay (Fig. 8C). Alternatively,
the reactive protein in complex 1T could contain another runt domain
protein related to AML-1.
We observed the loss of tissue-differential IVFP patterns in the widely
active, truncated LCR-driven transgene. We have proposed that this
reflects the alteration of factor binding at these sequences, resulting
in a change in transgene activity. As an alternative explanation, a
subset of the transgene copies in the tandem array could be shut down
by a nonexpressing chromatin configuration. This would prevent factor
binding and mask any differential IVFP pattern that may exist at HS6 of
the remaining, expressing copies. Our previous data has shown that
:2–6 transgene expression correlates with copy number in many
organs (14). Furthermore, the HS6 region is equivalently
and abundantly DNase I hypersensitive in both thymus and liver
(14). Although this data would argue against
the alternative explanation for the loss of tissue-specific IVFP
patterns, in our system it is not possible to assay individual
transgene copies in the tandem array. Therefore, we cannot formally
exclude the possibility of heterogeneity in transgene
configuration.
The data reported here suggest that the HS6 region can support both
ubiquitously active and T cell-specific functions. The location of
this element, between a T cell-specific and a ubiquitously expressed
gene, makes this finding of special interest. It is possible that this
region is responsible for coordinating the separation of Dad1 and TCR
gene regulation. Recently, several in vitro enhancer-blocking
activities were described in the HS2–6 region of the LCR
(55). Although this finding lends credence to this
hypothesis, it is still, at this point, difficult to determine the
relationship between these in vitro findings and our in vivo results.
The enhancer-blocking activity does not localize to the HS6 region.
Furthermore, the sequences at the position of HS6 (shown in Fig. 8) do
not contain a site homologous to the recognition sequence for the
CTCF transcription factor. CTCF binds a 42-bp sequence that has
been found to be important for the enhancer-blocking activity of
several vertebrate insulators (56). Further studies are
necessary to determine the significance of our data and those of Zhong
and Krangel (55) to the regulation of TCR
vis a vis Dad1 in its native locus. Given the prevalence of
LCR-like activities in loci specifically expressed in the various cells
of the blood, it is likely that continued study of the structure, and
function, of LCRs may yield important clues as to the molecular
effectors of hemopoietic differentiation.
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Benjamin D. Ortiz, Department of Biological Sciences, City University of New York, Hunter College, 695 Park Avenue, Room 927-N, New York, NY 10021. E-mail address: ortiz{at}genectr.hunter.cuny.edu
3 Abbreviations used in this paper: LCR, locus control region; HS, DNase I hypersensitive sites; IVFP, in vivo footprint(ing); LM, ligation-mediated; DMS, dimethyl sulfate; TF, thymic footprint; CBF, core binding factor.
Received for publication May 30, 2001. Accepted for publication July 30, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
/ locus impairs T-cell development and reveals the presence of the nearby antiapoptosis gene Dad1. Mol. Cell. Biol. 17:2151.[Abstract]
-globin gene in transgenic mice. Cell 51:975.[Medline]
/ locus. Immunity 1:207.[Medline]
locus. EMBO J. 8:729.[Medline]
locus required for tissue-specific locus control region activity. Mol. Cell. Biol. 19:1901. locus control region. J. Biol. Chem. 275:1952.-interferon gene requires an interferon-inducible factor. Mol. Cell. Biol. 6:801.-globin DNase I hypersensitive site. Nature 318:680.[Medline]
B and transcription of human immunodeficiency virus. Proc. Natl. Acad. Sci. USA 87:9943.-globin genes in transgenic mice. EMBO J. 4:1715.[Medline]
-globin gene in transgenic mice. Nature 315:338.[Medline]
-globin locus control region. EMBO J. 15:562.[Medline]
-globin gene switching and DNase I sensitivity of the endogenous -globin locus in mice do not require the locus control region. Mol. Cell 5:387.[Medline]
-globin LCR is not necessary for an open chromatin structure or developmentally regulated transcription of the native mouse -globin locus. Mol. Cell 2:447.[Medline]
-globin locus but not the maintenance of an open chromatin structure in erythroid cells. Mol. Cell. Biol. 18:5992.5-VpreB1 locus control region. Mol. Cell. Biol. 19:671.-globin gene expression: straightening out the locus. Curr. Opin. Genet. Dev. 6:488.[Medline]
enhancer function. Genes Dev. 11:640.3 (AML2) is induced by TGF-1 to bind and activate the mouse germline Ig promoter. J. Immunol. 161:6751. and Dad1 genes. J. Immunol. 163:295.This article has been cited by other articles:
|
|
 |
|
  J. Gomos-Klein, F. Harrow, J. Alarcon, and B. D. Ortiz CTCF-Independent, but Not CTCF-Dependent, Elements Significantly Contribute to TCR-{alpha} Locus Control Region Activity J. Immunol., July 15, 2007; 179(2): 1088 - 1095. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Harrow and B. D. Ortiz The TCR{alpha} Locus Control Region Specifies Thymic, But Not Peripheral, Patterns of TCR{alpha} Gene Expression J. Immunol., November 15, 2005; 175(10): 6659 - 6667. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Magdinier, T. M. Yusufzai, and G. Felsenfeld Both CTCF-dependent and -independent Insulators Are Found between the Mouse T Cell Receptor {alpha} and Dad1 Genes J. Biol. Chem., June 11, 2004; 279(24): 25381 - 25389. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Harrow, J. U. Amuta, S. R. Hutchinson, F. Akwaa, and B. D. Ortiz Factors Binding a Non-classical Cis-element Prevent Heterochromatin Effects on Locus Control Region Activity J. Biol. Chem., April 23, 2004; 279(17): 17842 - 17849. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. FELSENFELD, B. BURGESS-BEUSSE, C. FARRELL, M. GASZNER, R. GHIRLANDO, S. HUANG, C. JIN, M. LITT, F. MAGDINIER, V. MUTSKOV, et al. Chromatin Boundaries and Chromatin Domains Cold Spring Harb Symp Quant Biol, January 1, 2004; 69(0): 245 - 250. [Abstract] [PDF]
|
|
|
|
 |
|
 M. Bulger, D. Schubeler, M. A. Bender, J. Hamilton, C. M. Farrell, R. C. Hardison, and M. Groudine A Complex Chromatin Landscape Revealed by Patterns of Nuclease Sensitivity and Histone Modification within the Mouse {beta}-Globin Locus Mol. Cell. Biol., August 1, 2003; 23(15): 5234 - 5244. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Q. Li, K. R. Peterson, X. Fang, and G. Stamatoyannopoulos Locus control regions Blood, October 16, 2002; 100(9): 3077 - 3086. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and 
