To characterize the transcriptional mechanisms controlling NRAMP1 expression, we previously showed that a 263 bp region upstream of the ATG drives basal promoter activity, and that a 325
Trang 1gene expression
Etienne Richer1,2, Carole G Campion1, Basel Dabbas3, John H White3 and Mathieu F M Cellier1
1 Institut national de la recherche scientifique, INRS-Institut Armand-Frappier, Laval, Canada
2 The Centre for the Study of Host Resistance, McGill University, Montreal, Canada
3 Department of Physiology, McGill University, Montreal, Canada
The natural resistance-associated macrophage protein 1
(Nramp1, also known as solute carrier family 11,
member 1, Slc11A1) confers innate resistance to
intra-cellular parasites in mice [1] The activity of the Nramp1
transporter in the membrane of phagosomes prevents
growth of ingested microbes and limits their capacity to
produce a lethal infection [2] Genetic polymorphisms in
the NRAMP1 and vitamin D receptor (VDR) genes
were linked to innate susceptibility to mycobacterial
infections or increased risk of immune diseases [3,4]
Functional analyses of NRAMP1 promoter alleles
suggested a possible impact of polymorphisms on gene
expression and function [5–7]
NRAMP1expression is restricted to mature myeloid cells: primary monocytes, macrophages and neutroph-ils, ranked by increasing mRNA abundance Transient transfection studies showed that a DNA fragment extending 647 bp upstream of the NRAMP1 ATG enables transcriptional activation in response to VDR ligands in HL-60 cells, but not in nonmyeloid cells HL-60 clones stably transfected with this promoter fragment showed dose- and time-dependent transcrip-tional responses to VDR ligands consistent with the accumulation of endogenous NRAMP1 mRNA induced during monocytic differentiation [8] Identifi-cation of the specific determinants controlling
Keywords
1,25D; innate immunity; myeloid
differentiation; phagocytes; transcriptional
regulation
Correspondence
M Cellier, INRS-Institut Armand-Frappier,
531, Bd des prairies, Laval, QC H7V 1B7,
Canada
Fax: +1 450 686 5301
Tel: +1 450 687 5010 ext 4681
E-mail: mathieu.cellier@iaf.inrs.ca
(Received 13 June 2008, revised 24 July
2008, accepted 12 August 2008)
doi:10.1111/j.1742-4658.2008.06640.x
The natural resistance-associated macrophage protein 1 (Nramp1), which belongs to a conserved family of membrane metal transporters, contributes
to phagocyte-autonomous antimicrobial defense mechanisms Genetic poly-morphisms in the human NRAMP1 gene predispose to susceptibility to infectious or inflammatory diseases To characterize the transcriptional mechanisms controlling NRAMP1 expression, we previously showed that a
263 bp region upstream of the ATG drives basal promoter activity, and that a 325 bp region further upstream confers myeloid specificity and acti-vation during differentiation of HL-60 cells induced by vitamin D Herein, the major transcription start site was mapped in the basal region by S1 protection assay, and two cis-acting elements essential for myeloid transac-tivation were characterized by in vitro DNase footprinting, electrophoretic mobility shift experiments, in vivo transfection assays using linker-mutated constructs, and chromatin immunoprecipitation assays in differentiated monocytic cells One distal cis element binds Sp1 and is required for NRAMP1 myeloid regulation Another site in the proximal region binds CCAAT enhancer binding proteins a or b and is crucial for transcription This study implicates Sp1 and C⁄ EBP factors in regulating the expression
of the NRAMP1 gene in myeloid cells
Abbreviations
CDP, CCAAT displacement protein; C ⁄ EBP, CCAAT enhancer-binding protein; ChIP, chromatin immunoprecipitation; 1,25D,
1a,25-dihydroxyvitamin D3;dsODN, double-stranded oligonucleotide; EB1089, 1(S),3(R)-dihydroxy-20(R)-[5¢-ethyl-5¢-hydroxy-hepta-1¢(E),3¢(E)-dien-1¢-yl]-9,10-secopregna-5(Z),7(E),10(19)-triene; EMSA, electrophoretic mobility shift assay; IFN-c, interferon-c; IRF, interferon-c response factor; KH1060, 20-epi-22-oxa-24a,26a,27a-tri-homo-1,25-dihydroxyvitamin D 3 ; MEF, myeloid Elf-1-like factor; NRAMP1, natural
resistance-associated macrophage protein 1; TSS, transcriptional start site; VDR, vitamin D receptor.
Trang 2NRAMP1 expression will shed light on the regulatory
cis-acting elements and trans factors involved during
myelopoiesis and immune responses, and further
increase our understanding of the possible influence of
NRAMP1 promoter genetic polymorphisms in human
susceptibility to diseases, including infections by
intra-cellular parasites
Vitamin D agonists have profound effects on the
immune system, specifically stimulating innate
antimi-crobial defenses and macrophage maturation [9,10]
Several major transcription factors known to regulate
myelopoiesis [11–13] also control genes expressed
dur-ing differentiation induced by the hormonal form of
vitamin D, 1,25-dihydroxyvitamin D3 (1,25D),
includ-ing the VDR [14], Sp1 [15] and CCAAT enhancer
binding proteins (C⁄ EBPs) [16] Sp1 regulates genes
associated with innate immunity in cooperation with
other tissue-specific or ‘terminal differentiation’-specific
nuclear factors [17] Sp1 is thus often associated with
Ets-related transcription factors, e.g myeloid Elf-1-like
factor (MEF) [18,19] Members of the C⁄ EBP family
are important nuclear factors that cooperate with
others, including Sp1, to regulate myeloid genes
[11,12] C⁄ EBP factors have prominent roles in
myelo-poiesis [20], and several isoforms, e.g C⁄ EBPa,
C⁄ EBPb and C ⁄ EBPe, are differentially regulated
dur-ing myelo-monocytic differentiation [16,21–23]
An NRAMP1 promoter-proximal region starting
263 bp upstream of the ATG is sufficient for maximal
transcription reporter activity in nonmyeloid cell lines,
whereas the more distal region (264–588 bp upstream
of the NRAMP1 ATG) is required for maximal
pro-moter activity and for responsiveness to 1,25D in
HL-60 cells but not in Jurkat T-cells [8] The data suggest
that the NRAMP1 promoter comprises a proximal
region binding a basal transcription complex (core
promoter) and a more distal, myeloid-specific region
(upstream promoter) To test this hypothesis, we
mapped the NRAMP1 transcriptional start site in
dif-ferentiated HL-60 cells, located basal and
myeloid-spe-cific cis-acting sites, and identified important
transcription factors controlling NRAMP1 expression
in myeloid cells
Results
Delineating NRAMP1 cis-acting elements in HL-60
cells undergoing differentiation
Primer extension mapping previously revealed several
NRAMP1transcriptional start sites (TSSs) in different
cell types [24,25] NRAMP1 TSSs in HL-60 cells were
thus located by an S1 nuclease protection assay
(Fig 1A) Two major protected fragments were obtained, 10 or 38 bp shorter than the specific probe used, the latter being more abundant (hereafter denominated )28 and +1, respectively) The absence
of a larger fragment lacking only the control 5¢ synthetic 7-mer (Supporting information Table S1) excluded other start sites upstream of the probe The results indicated heterogeneity of NRAMP1 TSSs in a single cell type, suggesting that the NRAMP1 pro-moter fragment NR1S may bind a basal transcription complex close to the +1 or)28 site (Fig 1D)
Myeloid-specific cis elements were mapped by assay-ing the transcriptional activity of nested deletions of a promoter fragment extending 647 bp upstream of the NRAMP1 ATG in transfected HL-60 cells differenti-ated with 1,25D or dimethylsulfoxide We used stable transfections based on previous data obtained with HL-60 transfected clones, which showed luciferase reporter activity that paralleled the accumulation of endogenous NRAMP1 mRNA in response to differen-tiation [8] The clone HSRL5 was used as a positive control to characterize the activity of clones represent-ing deletion constructs (Fig 1B; 5E3, 5E4, M-1, and data not shown) Some activity persisted with con-struct 5E3, and little remained with further deletions Similar results were obtained with 1,25D and dimethyl-sulfoxide (Fig 1B), implying that the NRAMP1 region upstream of)365 contributes to myeloid regulation
To locate NRAMP1 promoter cis-acting sites, in vitro DNase 1 footprint experiments were conducted on both strands, revealing 14 protected areas (e.g E2, E6, and E10; Fig 1C) Few clear differences were observed in the patterns obtained with different cell types Double-stranded oligonucleotides (dsODNs, 30 bp) overlapping the protected sites were used in electrophoretic mobility shift assays (EMSAs) to assess transcription factor-binding activities in vitro Three DNase-protected sites confirmed by specific EMSA are indicated relative to NRAMP1 deletion construct ends (Fig 1D) To characterize candidate cis-acting elements, consensus binding sites for known transcription factors were tested as decoys for inhibition of nuclear extract binding to NRAMP1 dsODNs, and linker mutations were designed to delineate NRAMP1-specific cis-acting sites using in vitro EMSAs and in vivo stable trans-fection assays
Sp1 binds NRAMP1 promoter site E10 and influences protein binding at site E6 The cytosine-rich site E10 located between the bound-aries of fragments M-2 and S (Fig 1D) was detected
as a faint footprint that was similar in intensity and
Trang 370 b
2 8
+1 luc
M-1 -296 luc
53 b
7 1 3
2
luc
luc
5E4
5E3 -365
-398
luc
luc
1
(-28)
25 b
0 5 2 4
luc
SRL -498
luc
Fold induction (RLU)
(+1)
C
149bp
117bp
427bp
Myeloid specific region
+1
D
249bp
ATG
S1 probe
Luc
CA
ACGT 1 2 3 4 5
E6
E10
E2 ACGT 1 2 3 4 5 ACGT 1 2 3 4 5
E10
KH DMSO
Fig 1 Organization and activity of the NRAMP1 promoter (A) Identification of the major TSS by S1 protection assay (+1, located 147 bp upstream of the ATG), and several minor TSSs either adjacent or 28 bp upstream, using HL-60 cells differentiated for 3 days with the 1,25D genomic analog KH1060 (KH) (B) NRAMP1 transcriptional activity in HL-60 clones, identified by numbers, which were stably transfected with promoter constructs of the indicated length relative to the +1 site Transcriptional activation was measured by luminometry and expressed as relative luciferase (luc) units (RLU) fold induction between cells untreated and cells treated for either 3 days with 1,25D or
5 days with dimethylsulfoxide (DMSO); the mean ± standard error (SE) of at least three independent experiments is presented (C) DNase digestion footprints of three putative cis elements, E10, E6 and E2, that bind nuclear extracts from various cell types: lane 2, HL-60 cells; lane 3, HL-60 cells treated with 10)8M KH for 4 days and activated with IFN-c; lane 4, HL-60 cells treated for 6 days with 1.25% dimethyl-sulfoxide; lane 5, Jurkat lymphoid T-cells; lane 1, control with no extracts A sequencing ladder run was used to locate the protected sites, indicated by a vertical bar (D) Schematic representation of the luciferase reporter constructs showing deletions in the upstream region of the NRAMP1 promoter, which is required for myeloid regulation, and locations of the three protected sites E2, E6 and E10, as well as the polymorphic CA repeat adjacent to the downstream E6 site.
Trang 4quality between different extracts (Fig 1C, left)
Unla-beled probe E10 in excess competed with this
band-shift, but not the mutant probe E10M0 (Fig 2A, left)
Mutants E10M1–3 competed slightly better, but less
than wild-type E10 (Table 1; data not shown),
indicat-ing specificity in nuclear factor bindindicat-ing to the E10
motif Moreover, a wild-type Sp1 dsODN decoy
strongly competed with E10 probe binding, and an
antibody against Sp1 induced a specific band supershift
(Fig 2A, right) Finally, two stably transfected HL-60
clones revealed that NR1L mutant E10M0 abolished
differentiation-induced transcription (Fig 2C),
imply-ing that Sp1 bindimply-ing to NRAMP1 site E10 is
impor-tant for gene transcription in vivo during myeloid
differentiation
Another cytosine-rich site downstream of the
poly-morphic CA dinucleotide repeat (E6; Fig 1D) was
protected with all nuclear extracts tested (Fig 1C,
center) E6 binding specificity was demonstrated by
competitive EMSA and abrogated by the mutation
E6M2 (Fig 2B, Table 1) Although the E6 motif was
predicted to bind Sp1, inclusion in EMSAs of an antibody against Sp1 reduced band-shift intensity only slightly Little binding competition occurred with excess unlabeled wild-type Sp1 dsODN decoy, albeit reproducibly (Fig 2B) Similar weak competi-tion was associated with a wild-type dsODN decoy for the Ets family member MEF (Fig 2B), but not for PU.1, AP-1, Stat and PU.1-IRF factors (data not shown), contrasting with the strong competition by wild-type E10 dsODN (Fig 2B) These data indicate that Sp1 contributes to interactions with NRAMP1 promoter site E6, and suggest MEF as a potential binding partner The role of site E6 in vivo was studied using clones stably transfected with the NR1L mutant construct E6M2, which lost transcriptional response to 1,25D (Fig 2C), but conserved some activity induced with dimethylsulfoxide Similar responses were observed with PCR mutant clones having a 2 bp CA deletion (promoter allele 9 [6], data not shown), indicating that the E6M2 mutation limits NRAMP1 transcriptional activation in vivo in response to 1,25D (Fig 2C)
C
Fig 2 The distal cis elements E2, E6 and E10 are required for NRAMP1 promoter activation during myeloid differentiation induced with 1,25D (A, B) EMSA using dsODNs covering sites E10 (A) or E6 (B) and nuclear extracts from HL-60 cells treated with KH1060 (KH) and IFN-c To characterize the specific band-shifts cated by an open arrowhead, a 50-fold excess of either cold dsODN or, as indi-cated, a corresponding inactive linker mutated dsODN (E10M0 and E6M2, Table 1), or specific dsODN decoys and their mutated inactive counterparts [Sp1 and mutated (mt) Sp1, MEF and MEF mt] were used in competitive EMSA NS, non-specific Band supershifts (SS) were obtained using an antibody against Sp1 Com-plexes were resolved by 6% PAGE (C) HL-60 clones stably transfected with NR1L promoter constructs, identified by numbers and containing sites inactivated
by linker mutagenesis (E2M2, E6M2 and E10M0), were used to measure RLU fold induction between untreated cells and cells treated for 3 days with KH1060, producing monocyte-like cells, or with dimethylsulfoxide (DMSO) to generate neu-trophils The mean ± SE of at least three independent experiments is presented.
Trang 5The 5¢-region upstream of)365 is important for
NRAMP1 myeloid expression
The region spanning )498 to )365 (constructs SRL
and 5E4) is required to regulate the promoter (Fig 1B),
and encompasses site E2 (Fig 1D), detected as a weak
DNase footprint that appeared to be better protected
with extracts from differentiated HL-60 cells (Fig 1C
right, lanes 3 and 4) An excess of cold mutant dsODN
competed for protection of site E2 specifically (E2M1
but not E2M2) (Table 1, and data not shown) Another
dsODN overlapping this site, E2.2, competed with E2
and E2M1 but not E2M2 (data not shown) Sequence
analyses suggested that E2 might be a composite site
for an interferon-c (IFN-c) response factor (IRF),
which would mediate NRAMP1 upregulation in mature
phagocytes exposed to IFN-c However, neither
dsODN decoys for PU-IRF, GAS (IFN-c-activating
sequence) and Stats (signal transducers and activators
of transcription, data not shown) nor any other factor
tested competed with E2 binding (data not shown) To
examine the impact of the E2M2 mutation in vivo,
three independent stably transfected HL-60 clones were
obtained, and all showed abrogation of NRAMP1
tran-scriptional activity in response to 1,25D (Fig 2C) The
E2M2 and E6M2 mutations seemed to preserve the
response to dimethylsulfoxide in vivo In comparison,
the deletion 5E3 downstream of the E2M2 mutation
had a less severe, more variable impact on NRAMP1
gene activation (Fig 1B,D) These data establish a
regulatory role in myeloid cells for the NRAMP1
pro-moter region upstream of)365, including site E2
Sp1 recognizes myeloid-specific sites and transactivates the NRAMP1 promoter in vivo
A role for the transcription factor Sp1 in the regula-tion of NRAMP1 activaregula-tion in mature myeloid cells was tested by using an antisense phosphorothioate ODN that inhibits Sp1 expression and observing the effect on pSRL-driven luciferase activity in HSRL5 cells differentiated with 1,25D (Fig 3A) A modest reduction in reporter activity was noted with the anti-Sp1 ODN that was statistically significant as compared to scrambled anti-Spl control ODN (AS), suggesting that in monocytic HL-60 cells Sp1 could contribute to the upregulation of NRAMP1 trans-cription
To show trans-activation of the NRAMP1 promoter
by Sp1, 293T epithelial cells were transiently cotrans-fected using NRAMP1 promoter luciferase constructs and pCMV vectors expressing similar levels of the Sp1 family members Sp1 and Sp3 [26–28] As previous studies in 293T cells revealed similar activities for the constructs NR1L and NR1S in the absence of cotrans-fected transcription factor [8], and because Sp1 is known to interact with distal or proximal parts of the promoters that it regulates [15–19], we compared lucif-erase activity levels obtained in the presence of Sp factors of NR1S and NR1L as well as of other con-structs of intermediate length Sp1 increased the tran-scriptional activity of constructs NR1L, 5E4 and M-1 about two-fold to three-fold as compared to the NR1S construct (Fig 3B), and experiments using NR1L and Sp3+ plasmids resulted in a lower level of activation
Table 1 Specific electromobility shifts.
#
Upper case letters indicate linker-mutations; nucleotides fitting the indicated transcription factor predicted sites are underlined.
Trang 6(data not shown) These data indicated that Sp1 can
trans-activate the NRAMP1 upstream promoter
Differences in the activity of NRAMP1 promoter
upstream elements that depended on the myeloid
versus nonmyeloid cellular background were further
noted Cotransfection of Sp1+plasmid and NRAMP1 promoter fragments of different lengths showed that construct 5E4 was as active as the full-length promoter NR1L in 293T cells (Fig 3B), although it was inactive during myeloid differentiation in stably transfected HL-60 clones (Fig 1B) Also, the construct M-2 failed
to mediate trans-activation by Sp1 in 293T cells (Fig 3B), despite containing site E10 (Fig 1D) How-ever, the E10M0 mutation in the NR1L construct sig-nificantly reduced Sp1-dependent luciferase activity in 293T cells (Fig 3B), consistent with the prominent role
of this site in myeloid cells and derived nuclear extracts (Fig 2) Overall, a lower level of Sp1-driven expression
in 293T cells was detected using the shortest construct NR1S, supporting our proposition that Sp1 interacts with the NRAMP1 promoter upstream of myeloid-specific elements, including site E10
CCAAT-binding factors activate NRAMP1 basal transcription in vivo
Cotransfection studies using the Sp1+plasmid revealed different transcriptional activities of NR1L and NR1S constructs in 293T cells (Fig 3B), indicating that, despite endogenous factors that function on NRAMP1 proximal elements [8], this system may detect other can-didate transcriptional regulators of NRAMP1 Neither the VDR, PU.1, IRF-4 nor IRF-8 demonstrated activity when cotransfected separately or in various combina-tions with the reporter construct NR1L (data not shown) However, cotransfection of NR1L with a plas-mid expressing the CCAAT displacement protein⁄ cut homeobox (CDP⁄ Cut, p200, typical of immature mye-loid cells) inhibited reporter activity (Fig 3C), consistent with the known repressor activity of full-length CDP In contrast, a proteolytically processed isoform lacking the inhibitory domain [29,30], CDP⁄ Cut p110, upregulated NRAMP1 transcription about six-fold as compared to the full-length CDP Unlike those of Sp1, the effects of CDP isoforms were mediated by the NRAMP1 pro-moter proximal region (Fig 3C) Also, NRAMP1 trans-cription was stimulated upon cotransfection of NR1S constructs and plasmids expressing C⁄ EBPs+ The
C⁄ EBP+plasmids upregulated transcription similarly to CDP⁄ Cut p110, indicating that CCAAT-binding factors also bind to the NRAMP1 basal promoter
NRAMP1 transcription is regulated by cooperation between distal and proximal elements
Previous transient transfections in immature HL-60 cells showed about three-fold higher NRAMP1
trans-A
RLU 4000
6000
0
2000
4000
ND
Ctrl
ND AS
KH Ctrl
KH Sp1
KH Sp3
KH AS
B
15
20
0
5
10
2
3
4
5
6
7
NR1L NR1S
C
0
1
2
Fig 3 In vivo transactivation of the NRAMP1 promoter by Sp1
requires the myeloid-specific upstream promoter region (A)
Anti-sense ODNs directed against Sp1, Sp3 or a scrambled anti-Sp1
control ODN were used to treat, for 24 h, HL-60 cells from clone
HSRL5, which were previously differentiated for 3 days with
KH1060 (KH), and RLU were determined by luminometry ND,
con-trol cells without ODN (B) NRAMP1 transcriptional activation was
measured by luminometry (RLU) in 293T cells transiently
cotrans-fected with NR1L promoter constructs either deleted or
linker-mutated, together with Sp1 + or control plasmid; results are
expressed in Sp1+fold induction (C) NRAMP1 transcriptional
activa-tion in the presence of trans-acting factors was measured in RLU
fold modulation in 293T cells transiently cotransfected with pGL3
constructs carrying long or short NRAMP1 promoter (NR1L or NR1S)
and control or expression plasmids encoding various transcription
factors, assayed alone or in combination The mean ± SE of at least
three independent experiments is presented in each panel.
Trang 7criptional activity using the NR1L versus the NR1S
construct, and upregulation by 1,25D also required
the full-length promoter [8] This suggested that
NRAMP1 transcription in myeloid cells involves
cooperation between factors bound to distal sites and
others acting at proximal sites of the promoter To
address this possibility in 293T cells, the trans-acting
activity of combinations of Sp1 and CDP⁄ Cuts or
C⁄ EBPs was studied by cotransfections with
NRAMP1 promoter constructs containing or not
containing the upstream myeloid-specific region
(Fig 3C) Significant increases in trans-activation
resulted from combining the NR1L construct with
Sp1+ and CDP p110+ plasmids, whereas the CDP
p200+ construct inhibited luciferase activity
(Fig 3C), demonstrating cross-talk between factors
bound to distal or proximal promoter sites A similar
but less pronounced trend was observed when Sp1
and C⁄ EBP factors were cotranfected in presence of
the NR1L construct, some combinations being
stimulatory (Sp1+ and C⁄ EBPa+ or b+ plasmids)
and some inhibitory (Sp1 and C⁄ EBPe) These effects
were specific because they required the NR1L
con-struct, and additional cotransfections using C⁄ EBPa
and Sp3 were less efficient than those using Sp1
(data not shown) The results suggest that the
func-tion of Sp1 bound to distal sites is subservient to
the occupancy of promoter proximal elements by CCAAT-binding or other factors, which may activate
or repress transcription
C⁄ EBPs recognize NRAMP1 proximal promoter site E14 in vitro
Assuming a model for NRAMP1 expression that includes a key CCAAT-binding (or other) factor occu-pying a proximal element, which may control tran-scription and integrate the activities of more distal factors (e.g Sp1), due to intrinsic properties (e.g CDP) or protein–interactions (e.g C⁄ EBPe and Sp1),
we sought candidate CCAAT factor-binding sites in the NRAMP1 proximal promoter The footprint E14 adjacent to the major TSS within the basal transcrip-tion region contains a motif fitting the consensus for the C⁄ EBP family [12] Given the prominent role of
C⁄ EBP factors in myeloid development and their activ-ity on the NRAMP1 promoter in 293T cells, we tested
C⁄ EBP binding at site E14 EMSA using site E14 showed a strong and diffuse band-shift of high mole-cular weight (Fig 4A) Nuclear factor binding was diminished by the E14M1 and E14M2 mutations, but not by the E14M3 mutation, outside the predicted site (data not shown and Table 1) Competitive EMSA using mutant E14 sites confirmed the results of binding
Fig 4 The NRAMP1 proximal site E14 is bound in vitro by C ⁄ EBPb and C ⁄ EBPa (A) Nuclear extracts of HL-60 cells differentiated with KH1060 (KH) and activated with IFN-c were incubated with labeled E14 dsODN for EMSA A specific band-shift (indicated by an open arrow-head) is shown by the absence of competitive EMSA using mutated dsODN (E14M1, C ⁄ EBP mut) as compared to wild-type dsODN (B, C) EMSA, competitive EMSA and band supershifts (SS) were obtained using nuclear extracts from HL-60 cells differentiated with KH1060 for
5 days (B) or dimethylsulfoxide (DMSO) for 6 days (C), which were incubated with labeled E14 dsODN and additional cold dsODN or in the presence of antibody against C ⁄ EBPa, antibody against C ⁄ EBPb [2 and 6 lg (B) or 3 lg (C)] or control rabbit IgG antibodies, as indicated, and the complexes were resolved by 4% PAGE.
Trang 8experiments, with little competition due to the E14M1
mutation (Fig 4A) The wild-type C⁄ EBP dsODN
decoy strongly competed with binding, unlike the CDP
(Fig 4A), PU.1 or Sp1 decoys (data not shown)
Supershift experiments demonstrated strong binding to
site E14 of C⁄ EBPb and, to a lesser extent, C ⁄ EBPa
(Fig 4B,C)
C⁄ EBPs trans-activate the NRAMP1 proximal
region during myeloid differentiation
The impact of site E14 on NRAMP1 transcriptional
activity was deduced from luciferase activity levels after
transient cotransfection of 293T cells using the NR1L
mutant constructs E14M1 or E10M0, and the
expres-sion plasmids Sp1+and C⁄ EBPa+, singly or combined
(Fig 5A) The E14M1 mutation virtually abolished
NRAMP1 transcription; minimal effects of added
nuclear factors persisted, but were drastically reduced
In comparison, the Sp1-binding site E10M0 mutation
limited only the expression levels obtained with Sp1+
plasmid (Figs 3C and 5A) The role of NRAMP1 site
E14 was confirmed in myeloid HL-60 cells stably
trans-fected with the mutated pSRL construct E14M1 No
luciferase expression was detected either prior to or
after differentiation (Fig 5B,C), and genomic DNA
PCR fragments indicated promoter construct integrity
for all clones tested (data not shown) Such a dramatic
effect proved that CCAAT-binding factors are
crucial to control of NRAMP1 transcriptional activity
Endogenous Sp1 and C⁄ EBP factors bind the
NRAMP1 promoter in maturing monocytic cells
Although myeloid-specific expression is a conserved
property among human and mouse NRAMP1
ortho-logs, the factors found to be involved so far differ
between species (Fig 6) The murine promoter
con-tains two Inr sequences preceded by a proximal Sp1
site and up to six E-boxes (‘myc-max’) scattered in the
upstream region [31,32], and its expression is
con-trolled by the macrophage-specific transcription factor
IRF-8 [33,34] To determine whether current species
differences reflect divergence [35] or limited knowledge
of the mechanisms involved and extent of their
conser-vation requires studies detailing gene regulation in vivo
We used chromatin immunoprecipitation (ChIP) assays
to test whether C⁄ EBPs and Sp1 factors were recruited
to the endogenous NRAMP1 promoter during
mono-cytic differentiation We studied both C⁄ EBPa and
C⁄ EBPb, and compared HL-60 cells untreated or
dif-ferentiated for 24–48 h using the 1,25D genomic
agon-ists KH1060 (KH; Fig 7A) or EB1089 (EB; Fig 7B)
C⁄ EBPa appeared to be specifically bound to the NRAMP1 promoter and to be more abundant in growing than in differentiating cells In contrast, asso-ciation of C⁄ EBPb with the NRAMP1 promoter was detected in differentiating cells (Fig 7A,B) Assays
A
6
1 2 3 4 5
B
0 1
C
104
105
No treatment KH
100
101
102
103
104
100
101
102
103
DMSO
HSRL -5 E14M1 -2 E14M1 -4
Fig 5 The proximal cis element E14 is essential for NRAMP1 basal transactivation (A) Basal transcription levels in 293T cells transiently cotransfected with linker mutations of the NRAMP1 pro-moter construct NR1L (E10M0 or E14M1) and with carrier DNA or expression plasmids encoding Sp1, C ⁄ EBPa or both RLU fold induction was determined by comparison with the wild-type NR1L construct alone The data presented are the mean ± SE of at least three independent experiments (B, C) Luciferase activity (RLU) of HL-60 clones stably transfected with NRAMP1 long promoter con-structs containing the mutation E14M1 and compared to HSRL5, either untreated (B) or differentiated with KH1060 (KH) or dimethyl-sulfoxide (DMSO) for 4 days (C).
Trang 9targeting Sp1 and using HL-60 cells differentiated with
1,25D showed a weaker specific signal (Fig 7C), which
was not seen when cells were treated for 24 h or less
(data not shown); other assays using different condi-tions (specific antibodies against Sp1, oligonucleotide primer pairs) confirmed Sp1 binding to the NRAMP1
A
B
C
Fig 6 NRAMP1 distal and proximal promoter features (A) The four cis elements identified in this study within the 647 bp NRAMP1 promoter are indicated (E2, E6, E10 and E14); the dsODNs that were used in EMSA, and the 6 bp linker mutations abrogating trans-acting factor binding in vitro, are indicated by overlining and bold letters, respectively E2, E6, E10 and a fifth element, E3, located between the deletion boundaries 5E3 and 5E4, are required for NRAMP1 promoter trans-activation in vivo during HL-60 differentiation, whereas E14 is essential for NRAMP1 promoter activity The major TSS (+1) and another TSS upstream ( )28) are boxed, and the 63-mer probe used for S1 nuclease mapping is indicated The polymorphic CA repeat is shaded, and CpG dinucleotides, which are mainly clustered in the basal proxi-mal promoter region, are in bold italics The initiation codon is indicated by ATG (Met) (B) Alignment of the promoter sequences of NRAMP1 (top) and the cattle and mouse counterparts (respectively, middle and bottom) shows the absence of conservation of the five sites important for NRAMP1 regulation that were identified in this study (C) Schematic representation of the cis-acting elements that were identified in the promoter regions of the human NRAMP1 and mouse orthologs.
Trang 10promoter after monocytic differentiation (Fig 7D).
These in vivo data established that C⁄ EBP and Sp1
factors regulate NRAMP1 expression during myeloid
differentiation
Discussion
This study identified cis-acting sequences in the
NRAMP1 promoter and transcription factors binding
to them, providing a mechanistic basis for the restric-tion of NRAMP1 gene expression to mature myeloid cells The results also provide novel, valuable knowl-edge with which to interpret human genetic polymor-phisms, including allelic variation at the NRAMP1 promoter and in genes encoding the DNA-binding factors that regulate its transcription, in relation to disease resistance
The NRAMP1 promoter exhibits several characteris-tic myeloid properties: it is compact, lacks canonical TATA, initiator sequences or CCAAT boxes, and shows heterogeneous TSSs [36] The NRAMP1 pro-moter is divided into two regions: (a) the proximal core region, which spans the TSS and is presumed to bind the basic RNA polymerase II-dependent mac-hinery; and (b) the 5¢-upstream region required for developmental regulation of expression
Three upstream cis-acting elements (E2, E6 and E10) mediate differentiation-dependent NRAMP1 tran-scription induced by 1,25D, as well as by dimethylsulf-oxide (E10 only) Binding of Sp1 to site E10 was demonstrated using in vitro and in vivo assays, includ-ing in the context of chromatinized DNA in HL-60 monocytic cells The key role of Sp1 in the control of myeloid gene expression has been recently reviewed [17] Although Sp1 is a ubiquitous factor, it contrib-utes to myeloid-restricted expression by rendering promoters more accessible by various mechanisms, including demethylation of Sp1-binding sites (‘GC box’) [37], local chromatin structure rearrangement [38,39], interactions with other nuclear factors, and⁄ or post-translational modifications [17]
C⁄ EBP factors also have prominent roles in myeloid cell development [40] These factors trans-activated NRAMP1 through the proximal site E14;
C⁄ EBPa and C ⁄ EBPb bound this site in vitro and the NRAMP1 proximal promoter in vivo, as they do with other myeloid genes [41,42] Higher levels of C⁄ EBPb binding in HL-60 monocytic cells are consistent with upregulation of this factor in mature cells [16] The combined action of Sp1 and C⁄ EBPa is essential for maximal myeloid expression of LF and CD11c [43,44], and some C⁄ EBP sites were found to be crucial for gene activation by Sp1 Synergistic trans-activation by Sp1 and C⁄ EBPb was also reported in hepatoma cells [45,46] Interestingly, both C⁄ EBPb and Sp1 can recruit Mediator, a multiprotein complex acting as molecular bridge between enhancer-bound activators and the core transcriptional machinery [40] Thus, it appears possible that NRAMP1 expression results from direct cooperation between C⁄ EBPb located on
a proximal element and Sp1 bound to more distal sites
KH (h)
HL 60 0 24 0 24 0 24 0 24
A
200 bp
300 bp
HL-60
B
200 bp
300 bp
200 bp
300 bp
(-111/+163) (+1924/+2139) Untreated
HL-60
24 h EB
HL
200 bp
300 bp
48 h EB
HL-60
HL-60
C
(-204/+73)
D
48 h VitD
HL-60 300 bp
200 bp
300 bp
72 h EB
HL-60
(-392/-93) (+1924/+2139)
Fig 7 In vivo recruitment of C⁄ EBPb and Sp1 on the proximal and
distal parts of the NRAMP1 promoter during HL-60 differentiation
induced by 1,25D (VitD) analogs ChIP assays were performed using
antibodies against C ⁄ EBP (A, B) or Sp1 (C, D) as indicated in
Experi-mental procedures C ⁄ EBPa and C ⁄ EBPb binding were assayed in
resting HL-60 cells and in cells preincubated for 24 h with the 1,25D
agonist KH1060 (KH) (A) Cells untreated or differentiated for 24–
72 h using the 1,25D agonist EB1089 (EB) were assayed for
C ⁄ EBPs (B) and Sp1 (D); NRAMP1 exon 3 PCR amplification was
used to control DNA fragment size Control normal rabbit IgG was
used as specificity control (A, C) Numbers in parentheses indicate
NRAMP1 gene base pair coordinates relative to the major TSS.