Search tips
Search criteria 


Logo of molcellbPermissionsJournals.ASM.orgJournalMCB ArticleJournal InfoAuthorsReviewers
Mol Cell Biol. Feb 2011; 31(4): 861–875.
Published online Dec 6, 2010. doi:  10.1128/MCB.00836-10
PMCID: PMC3028656
Serum Response Factor Utilizes Distinct Promoter- and Enhancer-Based Mechanisms To Regulate Cytoskeletal Gene Expression in Macrophages[down-pointing small open triangle]
Amy L. Sullivan,1 Christopher Benner,1 Sven Heinz,1 Wendy Huang,1 Lan Xie,2,3 Joseph M. Miano,4 and Christopher K. Glass1*
University of California, San Diego, Department of Cellular and Molecular Medicine, 9500 Gilman Drive, La Jolla, California 92093,1 Medical Systems Biology Research Center, School of Medicine, Tsinghua University, Beijing 100084, China,2 National Engineering Research Center for Beijing Biochip Technology, Beijing 102206, China,3 Aab Cardiovascular Research Institute, University of Rochester School of Medicine and Dentistry, Rochester, New York 146424
*Corresponding author. Mailing address: University of California, San Diego, Department of Cellular and Molecular Medicine, 9500 Gilman Drive, MC0651, La Jolla, CA 92093-0651. Phone: (858) 534-6011. Fax: (858) 822-2127. E-mail: ckg/at/
Received July 20, 2010; Revised September 7, 2010; Accepted November 26, 2010.
Cells of the monocyte/macrophage lineage play essential roles in tissue homeostasis and immune responses, but mechanisms underlying the coordinated expression of cytoskeletal genes required for specialized functions of these cells, such as directed migration and phagocytosis, remain unknown. Here, using genetic and genomic approaches, we provide evidence that serum response factor (SRF) regulates both general and cell type-restricted components of the cytoskeletal gene expression program in macrophages. Genome-wide location analysis of SRF in macrophages demonstrates enrichment of SRF binding at ubiquitously expressed target gene promoters, as expected, but also reveals that the majority of SRF binding sites associated with cell type-restricted target genes are at distal inter- and intragenic locations. Most of these distal SRF binding sites are established by the prior binding of the macrophage- and the B cell-specific transcription factor PU.1 and exhibit histone modifications characteristic of enhancers. Consistent with this, representative cytoskeletal target genes associated with these elements require both SRF and PU.1 for full expression. These findings suggest that SRF uses two distinct molecular strategies to regulate programs of cytoskeletal gene expression: a promoter-based strategy for ubiquitously expressed target genes and an enhancer-based strategy at target genes that exhibit cell type-restricted patterns of expression.
Cells of the monocyte/macrophage lineage are key effectors and regulators of innate and acquired immune responses and participate in diverse aspects of tissue homeostasis (35, 38). These roles require the acquisition of both general and specialized functions of the actin cytoskeleton that are necessary for adhesion, directed migration, phagocytosis, and antigen presentation. In addition to using broadly expressed components of the cytoskeleton, such as actin itself, macrophages utilize a number of cell-restricted factors to enable or regulate specialized aspects of cytoskeleton-dependent processes. For example, Coro1a is preferentially expressed in cells of the hematopoietic lineage and functions to regulate phagosome-lysosome fusion in macrophages (10). The mechanisms that enable the coordinated expression of genes required for both the general and specialized functions of the macrophage cytoskeleton remain poorly understood.
Based on its known functions in other cell types, the ubiquitously expressed transcription factor serum response factor (SRF) is likely to play important roles in regulating the expression of cytoskeletal genes in macrophages. Although deletion of the Srf gene in mice results in embryonic lethality at the time of gastrulation, Srf−/− embryonic stem cells are able to grow and differentiate in vitro but show defects in cell spreading, adhesion, and migration (3, 43-44, 60). Conditional knockout (KO) of Srf in all three muscle types and in the brains of mice have confirmed roles for SRF in regulating cytoskeletal and contractile genes that are essential for the proper development and function of these tissues (2, 26, 33, 36-37, 61). The roles of SRF in the regulation of cytoskeletal factors are also highly conserved because defects in migration and proper cell targeting and morphology have been seen in inactivation and knockdown models of SRF in Dictyostelium discoideum, Caenorhabditis elegans, and Drosophila melanogaster (32). These results are consistent with a recent computational analysis that estimates that nearly half of the SRF target genes are related to the cytoskeleton and contractile apparatus (50).
SRF is a member of the MADS (MCM1, agamous, deficiens, SRF) box family of transcription factors that contain a highly conserved, N-terminal MADS box domain that mediates DNA binding, dimerization, and protein-protein interactions (47). SRF regulates gene expression by binding as a homodimer to a sequence motif termed the CArG box [CC(A/T)6GG], which has been associated primarily with growth (e.g., c-Fos and JunB) and muscle-specific differentiation and cytoskeletal (e.g., Tagln and Acta2) genes (17, 31).
The ability of SRF to regulate such diverse programs of gene expression is currently attributed to the ability of SRF to interact with different cofactors. The best studied of these cofactors are the ternary complex factors (TCFs) (ELK1, ELK3 [also known as NET], and ELK4 [also known as SAP1], which are members of the Ets family of transcription factors). The TCFs bind to an Ets motif [GGA(A/T)] that is found adjacent to the CArG motif (collectively called the serum response element [SRE]) in the promoters of immediate early growth genes (31, 51-52). More recently, SRF has also been shown to interact with members of the myocardin family of transcription factors (MYOCD, MKL1 [also known as MRTF-A, MAL, and BSAC], MKL2 [also known as MRTF-B]) in the regulation of cytoskeletal and contractile genes (6, 12, 26, 34, 42, 45, 55-56). Interaction studies have shown that MKL does not bind to DNA but does bind directly to SRF through its DNA binding domain in a manner that is mutually exclusive with TCF binding, thus providing an alternative mechanism for regulation of SRF-dependent programs of gene expression (34, 57).
The well-established notion that proper cytoskeletal protein expression and regulation are required for appropriate macrophage functional responses, in conjunction with the observation that SRF is an important regulator of the actin cytoskeleton in other cell systems, led us to explore the roles of SRF in the regulation of macrophage gene expression and function. Here, we demonstrate that SRF controls the expression of both general and hematopoietic-cell- restricted cytoskeletal genes. Using genome-wide location analysis, we find that SRF localization to hematopoietic-cell-restricted genes occurs primarily at distal sites and is dependent on the macrophage- and B cell-specific transcription factor PU.1. Functional studies further demonstrate that PU.1 is required for activation of these genes, providing insight into how cell-specific programs of SRF-dependent gene expression are achieved.
Conditional KO of Srf in macrophages.
Srfflox/flox (Srffl/fl) mice were generated as previously described (33, 40). Srffl/fl mice were crossed with mice harboring the inducible Mx-Cre transgene to obtain mice that were homozygous for the floxed Srf allele and either negative (wild type [WT]) or positive (KO) for the Mx-Cre transgene (21). For Srf WT/KO experiments, mice were injected intraperitoneally (i.p.) with 200 μl of 2-mg/ml poly(I)·poly(C) (pIpC) (Amersham Biosciences) 3 days prior to thioglycolate injection. Genomic deletion was confirmed by PCR using the following primers: F1 (5′-TGCTTACTGGAAAGCTCATGG-3′), R1 (5′-TGCTGGTTTGGCATCAACT-3′), and R2 (5′-CAAGACGACTCCCATCCTTG-3′).
Cell culture.
Macrophages were isolated from 8- to 14-week-old male C57BL/6 mice (Harlan). Elicited macrophages were isolated by peritoneal lavage 3 days after i.p. injection of thioglycolate broth and grown in Dulbecco's modified Eagle's medium (DMEM; 4.5 g/liter glucose) (Cellgro) plus 10% heat-inactivated fetal bovine serum (FBS) (HyClone) and 100 U penicillin-streptomycin (Invitrogen). PUER cells were cultured as described previously and differentiated with 100 nM 4-hydroxy-tamoxifen (Sigma) for the times indicated in the figures (22, 54). Equal numbers of bone marrow cells (following red blood cell lysis) from WT and KO mice were plated on 10-cm petri dishes and differentiated to macrophages as previously described (53).
siRNA transfection.
Primary macrophages were plated at 7.5 × 105 cells per well of a 24-well plate in growth medium without antibiotics overnight. Cells were transfected with nonspecific (NS) control or SMARTpool small interfering RNAs (siRNAs) (Dharmacon) using Deliver X transfection reagent (Panomics) according to the manufacturer's instructions. Cell samples transfected with the PU.1 siRNA showed visible cytotoxicity at longer incubations (~48 h), so cells treated in parallel with the PU.1 siRNA were harvested at 30 h when no cytotoxicity was evident.
Expression array profiling.
Total RNA from primary, elicited macrophages was purified using the RNeasy kit (Qiagen) (with DNase digestion) according to the manufacturer's instructions. Biological duplicates of control and knockout samples were labeled and hybridized on the 44K whole-mouse-genome oligonucleotide microarray (Agilent) according to the manufacturer's instructions. Slides were scanned according to the manufacturer's specifications and quantified using Feature Extraction software (Agilent). Genes were considered present if every replicate for at least one condition (i.e., WT and KO) were above the threshold value cutoff of 100 (based on the histograms of the expression values) for the array. To calculate the false-discovery rates (FDR), genes from replicate experiments were ranked by the log ratio of levels of expression between knockout and control cells. For each gene, a rank sum statistic, “R,” was defined as the sum of its ranks in all replicate experiments. The expected number of false positives was calculated exactly as the mean number of genes whose rank sum statistic was more extremal than “R” if the individual ranks were independent, uniform, random variables. The FDR was calculated as the ratio of this number and the actual, experimentally determined number of genes whose rank sum statistic was more extremal than “R.” Target gene expression was considered to be significantly changed if the difference between the expression levels in the control and knockout samples was below a 0.1 FDR threshold.
GO analysis of gene expression data.
Gene ontology (GO) analysis of microarray results was performed using the Web-based DAVID functional-annotation tool ( (8). Significantly changed genes were compared to a background set of the total complement of genes that were represented on the microarray. Gene ontology terms were considered significant if they had a Benjamini-corrected P value of less than 0.05.
cDNA was prepared from total RNA templates using Superscript III (Invitrogen) reverse transcriptase according to the manufacturer's instructions. Quantitative real-time PCR (qPCR) analysis was performed using a DNA template, 50 ng of each primer, and SYBR greenER master mix (Invitrogen) in 10-μl reaction mixtures on a Step One Plus or 7300 real-time PCR system (both from Applied Biosystems). cDNA primer sequences are available upon request.
FACS analysis.
Primary macrophages and total bone marrow cells were incubated with red blood cell lysis buffer, washed, and resuspended in phosphate-buffered saline (PBS) containing 0.1% bovine serum albumin (BSA) and anti-CD16/CD32 antibody for 10 min, on ice, to block Fc receptors. Bone marrow macrophages were harvested from 10-cm petri dishes, with a 15-min incubation in PBS plus 5 mM EDTA at 37°C. All cells were washed in PBS-BSA and stained for 10 min, on ice, with antibodies to the cell surface proteins indicated in the figures. If required, the staining step was repeated with streptavidin-conjugated fluorophores, followed by washes with PBS-BSA. Just prior to acquisition, DAPI (Roche) was added to 50 μg/ml. Data were acquired using an LSR II cytometer (BD) and analyzed using FlowJo software. All antibodies/conjugates used were from eBioscience.
Phagocytosis assay.
Primary, thioglycolate-elicited macrophages were plated at 1 × 105 cells/well of a 96-well plate in growth medium. The phagocytosis assay was carried out using the Vybrant phagocytosis assay kit (Invitrogen) according to the manufacturer's instructions, except that the cells were allowed to adhere for 1 h, followed by a 30-min treatment with dimethyl sulfoxide (DMSO) or 2.5 μM latrunculin B (Biomol) before the fluorescent bacteria were added.
In vitro chemotaxis and invasion assays.
Bone marrow macrophages were serum starved with RPMI 1640 on day 8 of differentiation for 3 h. Cells were recovered from petri dishes by being scraped, and cells were counted with trypan blue staining. Cells were resuspended to 5 × 105 viable cells/ml in RPMI 1640. For chemotaxis assays, 100 μl of the cell suspension was added to the top of the insert (Corning catalog number 3422; 8-μm pore size), and RPMI 1640 with or without 1 μM N-formyl-methionyl-leucyl-phenylalanine (fMLP) was added to the bottom wells. After 1 h, cells were fixed with 10% formaldehyde in PBS for 30 min and stained with 10 μM 4′,6-diamidino-2-phenylindole (DAPI) in PBS for 30 min. Cells were removed from the tops of the inserts with cotton swabs and rinsed with PBS. Pictures of the inserts were taken on an Olympus MVX10 microscope, and images were quantified using ImageJ by calculating the integrated density of the fluorescence signal over the insert. For invasion assays, transwell inserts that were coated with basement membrane extract (Corning catalog number 3458) were used. FBS and mouse recombinant macrophage colony-stimulating factor (M-CSF) were added to the cell suspension to final concentrations of 0.5% and 20 ng/ml, respectively. Inserts were hydrated with RPMI 1640 according to manufacturer's instructions. Following rehydration, 100 μl of the cell suspension was added to the top of the insert (without removing rehydration medium) and either RPMI 1640 containing 0.5% FBS and 10 ng/ml M-CSF (control) or RPMI 1640 containing 20% FBS, 10 ng/ml M-CSF, and 1 μM fMLP was added to the bottom wells. Cells were allowed to migrate for 48 h before being fixed, stained, imaged, and quantified as described for the chemotaxis assay.
Spreading assay.
Bone marrow macrophages from 2 mice per genotype were harvested on day 7 of differentiation by scraping and replating them on glass coverslips coated with a 1:10 dilution of Matrigel (BD Biosciences) according to the manufacturer's instructions. One hour after being plated, cells were rinsed with PBS and fixed in 4% paraformaldehyde-PBS for 10 min before being stained with rhodamine-conjugated phalloidin according to the manufacturer's instructions. Five fields of cells per mouse were visualized using the 63× oil immersion objective on a Zeiss AxioImager. Cell spreading was quantified using ImageJ.
ChIP assay.
Chromatin immunoprecipitation (ChIP) assays were performed as previously described (15). Briefly, 20 million cells were fixed with 1% formaldehyde, quenched with glycine (125 mM, final concentration), and washed with cold PBS. Nuclei were prepared and sonicated on wet ice 6 times for 10 s each time using 13 W of output power on a Misonix 3000 sonicator. Cleared supernatant was recovered, diluted, and precleared as previously described. Immunoprecipitations were performed using 2.5 μg of antibody at 4°C and incubated overnight with rotation. Antibody complexes were recovered using 50 μl of blocked ImmunoPure protein A-agarose (Pierce) and then washed, eluted, digested, and purified as previously described (15).
SRF (sc-335), PU.1 (sc-352), C/EBPβ (sc-150), and control (sc-2027) antibodies were purchased from Santa Cruz Biotechnology. Antibodies against H3K4me1 (ab8895) and H3K4me3 (ab8580) were from Abcam.
High-throughput ChIP sequencing (ChIP-seq).
Purified ChIP DNA (10 to 50 ng) was adapter ligated and PCR amplified according to the manufacturer's instructions (Illumina). Amplified fragments were sequenced for 36 cycles on an Illumina genome analyzer according to the manufacturer's instructions. Sequence tags returned by the Illumina Pipeline (first 23 to 25 bp) were mapped to the mouse genome using the mm8 assembly (NCBI build 36). Only those tags that mapped uniquely to the genome were considered for further analysis. Peaks were visualized by preparing custom tracks for the UCSC Genome Brower ( in a manner similar to that previously described (15, 41).
Identification and annotation of ChIP-seq peaks.
ChIP-seq peaks were identified using the HOMER software suite (, which was developed by our lab to facilitate ChIP-seq analysis by the method that has been described previously (15). To determine SRF-specific peaks, the results of the SRF ChIP-seq experiment performed with KO cells were used as the background. Only those peaks in WT cells with tag counts at least 4-fold more than the tag counts in KO cells were considered specific. For comparison, all experiments analyzed were normalized to represent 10 million total mapped tags. Peaks from multiple experiments were considered cobound if their peaks were found within 100 bp of one another. Expected overlaps of peaks from different experiments were based on a genome size of 2 × 109 to account for duplicated/repetitive regions where peaks cannot be identified, and significance calculations were made using the Fisher exact test. Peaks were associated with genes by identifying the nearest NCBI Reference Sequence (RefSeq) transcription start site (TSS), and exon and intron, etc., annotations were based on RefSeq transcripts. For the annotation analysis, the promoters were defined as −1 kb to +1 kb relative to the position of the TSS. Scatter plots and histograms of ChIP-seq data were generated as previously described (15).
HOMER de novo motif analysis.
HOMER was used to identify enriched sequence motifs in ChIP-seq peaks as previously described (15). Briefly, sequences spanning from −100 bp to +100 bp relative to the ChIP-Seq peak centers were extracted and their CpG contents calculated. Fifty thousand random genomic sequences of the same size were then randomly selected from the genome such that their profiles of CpG content were identical to those of the target sequences to avoid identification of nonspecific CpG island-associated motifs. The de novo HOMER algorithm was then used to identify motifs with lengths of 8, 10, and 12 bp that are overrepresented in the target sequences relative to motifs in the set of random background sequences. Sequence logos were generated using WebLOGO (
Immunoprecipitation and Western blot analysis.
WT and KO whole-cell lysates were prepared with 1 ml cold radioimmunoprecipitation (RIPA) buffer (40 mM Tris, pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.2% SDS, 0.5% Na-deoxycholate, 1× complete protease inhibitors, 1 mM phenylmethylsulfonyl fluoride [PMSF], 5 μM E64, 5 μM MG132). Samples were cleared at 18,000 × g for 10 min at 4°C. Protein concentrations were measured using protein assay reagent (Bio-Rad). Equal amounts of protein were incubated with SRF antibody (1 μg) overnight at 4°C, followed by 1 h with 20 μl of ImmunoPure protein A-agarose before being washed 3 times with RIPA buffer. After the last wash, the beads were resuspended in 3× SDS sample loading buffer, run on a 4 to 12% Bis-Tris SDS-PAGE gel (Invitrogen), and transferred to a polyvinylidene difluoride (PVDF) membrane (Immobilon-P; Millipore). Primary antibodies against SRF (sc-335) were from Santa Cruz Biotechnology. Alkaline phosphatase-conjugated secondary antibodies were from Jackson ImmunoResearch. Blots were developed using CDP-Star substrate (Applied Biosystems) and exposed to film according to the manufacturer's instructions. Input samples were visualized using Ponceau S stain (Sigma).
SRF primarily regulates cytoskeleton-dependent functions in macrophages.
To investigate the potential roles of SRF in the regulation of macrophage development and function, we crossed Srffl/fl mice with transgenic mice expressing the Cre recombinase gene under the control of the alpha/beta interferon-inducible (Mx) promoter (21, 33, 40). Following poly(I)·poly(C) (pIpC) injection to induce Cre recombinase expression, nearly complete deletion of the promoter region and exon 1 of Srf was observed in total bone marrow cells and primary Srffl/fl Mx-Cre+ (KO) macrophages, compared to the sequence of Srffl/fl Mx-CRE (WT) macrophages (data not shown and Fig. Fig.1A).1A). Genomic deletion of the Srf gene in macrophages also resulted in virtually undetectable levels of SRF mRNA and protein (Fig. 1B and C).
FIG. 1.
FIG. 1.
Srf KO macrophages differentiate normally. (A) Confirmation of genomic deletion of Srf in primary, thioglycolate-elicited macrophages by PCR. The WT band is ~450 bp, and the KO band is ~800 bp. (B) Representative qPCR analysis of Srf mRNA (more ...)
FACS analysis of Srf KO bone marrow cells indicated a small, but not statistically significant, reduction in the number of CD11b+ GR1+ macrophage precursor cells (Fig. 1D and E). In addition, there were also small, but not statistically significant, decreases in the numbers of mature macrophages (CD11b+ F4/80+) derived from total bone marrow cells following 7 days of in vitro differentiation with M-CSF (Fig. 1F and G). Despite similar levels of macrophage progenitor cells in bone marrow (Fig. 1D and E), the number of CD11b F4/80 double-positive Srf KO macrophages recovered from the peritoneal cavity in a model of sterile peritonitis induced by thioglycolate injection was significantly reduced compared to the number of WT macrophages recovered (Fig. (Fig.2A).2A). No differences were observed in the cell viabilities or the relative percentages of CD11b+ F4/80+ cells (data not shown and Fig. Fig.2B2B).
FIG. 2.
FIG. 2.
Cytoskeleton-dependent functions are attenuated in Srf KO macrophages. (A) Quantification of total live CD11b+ F4/80+ WT and KO macrophages that were recovered from peritoneal exudates 3 days following thioglycolate injection (5 WT and (more ...)
In order to investigate whether reduced infiltration of macrophages into the peritoneal cavity in response to thioglycolate injection might relate to a cell-autonomous defect, we performed in vitro chemotaxis and invasion assays using WT and Srf-deficient macrophages. Srf KO macrophages exhibited marginal (not significant) decreases in chemotactic abilities, as well as significantly (P < 0.02) decreased capacities for invasion through a basement membrane protein matrix (Fig. 2C and D). In addition, we used a rhodamine-conjugated version of the actin filament binding protein phalloidin to directly visualize the actin cytoskeleton in macrophages that were allowed to adhere to Matrigel-coated coverslips for 1 h. We observed that the overall intensity of the rhodamine signal was diminished in the Srf KO macrophages and that their ability to spread was significantly impaired (P < 0.0001) (Fig. 2E and F).
We further assessed the ability of the loss of SRF to compromise normal macrophage function by comparing the phagocytic capacities of wild-type and Srf-deficient macrophages. Srf KO macrophages were significantly less efficient at phagocytosing bacteria than WT macrophages (Fig. (Fig.2G),2G), with the level of phagocytosis being comparable to that in WT cells treated with the actin polymerization inhibitor latrunculin B. Latrunculin B was also able to further inhibit the phagocytic capacity of Srf KO macrophages, suggesting that actin protein is still present in these cells but at decreased levels that lead to attenuated function.
To more comprehensively assess the effect of the loss of Srf on macrophage gene expression, transcriptome analysis was performed using total RNA from resting WT and KO peritoneal macrophages obtained following thioglycolate injection. This analysis led to the identification of 166 genes that were significantly downregulated in the absence of Srf, while 99 genes were found to be upregulated compared to their expression in WT controls, as determined using a false-discovery rate (FDR) threshold of 0.1 (see Table S1 in the supplemental material).
To place these results in the context of macrophage-specific gene expression, we used corresponding microarray data from B cells and neutrophils to generate a list of 305 genes that were highly expressed in macrophages compared to levels of expression in these other cell types (expression in macrophages was >16-fold more than that in B cells or neutrophils [15]). Expression of these genes was largely unaltered in Srf KO macrophages (Fig. (Fig.3A).3A). Taken in concert with the normal numbers of monocyte precursor cells in Srf KO bone marrow, the normal expansion of these cells in response to M-CSF, and the normal expression of CD11b and F480 on bone marrow-derived Srf KO macrophages, we conclude that SRF is not required for the general program of macrophage differentiation following short-term deletion of SRF in the bone marrow compartment. However, a recent study evaluating consequences of the loss of SRF in the hematopoietic system at longer time points did identify differences in the numbers of myeloid progenitors (39). This difference could reflect roles of SRF in the hematopoietic stem cell or early myeloid progenitor cells that would not have been apparent in our short-term deletion protocol.
FIG. 3.
FIG. 3.
Srf KO macrophages exhibit defects in cytoskeletal gene expression. (A) Expression of most macrophage-specific genes is not affected in Srf KO macrophages. Scatter plot analysis of the microarray expression values for Srf WT and KO macrophages of genes (more ...)
Gene ontology (GO) analysis (8) of functional annotations associated with genes that were downregulated in Srf KO macrophages identified significant enrichment for terms, including actin cytoskeletal organization, immune system process, chemotaxis, and leukocyte differentiation. In contrast, corresponding analysis of the upregulated genes resulted in the identification of generic developmental terms (e.g., developmental process) (Fig. (Fig.3B;3B; for full results, see Tables S1 and S2 in the supplemental material). We confirmed that cytoskeletal genes are targets of SRF in macrophages by verifying the SRF dependence of a representative subset of genes with functional annotations linked to the cytoskeleton by using additional Srf WT and KO cDNA samples for qPCR analysis. As shown in Fig. Fig.3C,3C, all 12 of the cytoskeletal genes chosen were confirmed to be dependent on SRF for their full expression, while the expression of a control transcript, Jmjd5, was unaffected. We also confirmed the expression of several of the upregulated target genes (fos, egr1, egr2) (data not shown), which is consistent with previous observations that SRF can function as a transcriptional repressor (9).
With the exception of β-actin, all of the SRF target genes illustrated in Fig. Fig.3C3C exhibit marked differences in expression across tissues, and many are preferentially expressed in cells of myeloid or hematopoietic origin (BioGPS) (23, 62). As noted previously, Coro1a is preferentially expressed in cells of the hematopoietic lineage and functions to regulate phagosome-lysosome fusion in macrophages (10). Both Lcp1 and Lsp1 are actin-bundling proteins that have been shown to regulate adhesion, migration, and activation in both myeloid and lymphoid cells, while conditional knockout of the actin-binding protein gene Cnn2 in macrophages results in alterations in proper migration, spreading, and phagocytosis (7, 16, 18). These findings suggest that SRF must work in a cooperative manner with other transcription factors to direct the preferential expression of these genes in the hematopoietic system.
The SRF cistrome in macrophages is enriched for vicinal cytoskeletal genes.
To gain insight into mechanisms underlying SRF-dependent regulation of cytoskeletal gene expression, we performed chromatin immunoprecipitation (ChIP) assays for SRF in primary macrophages and then high-throughput sequencing of the enriched DNA fragments (ChIP-seq). To control for antibody specificity, parallel ChIP sequencing was performed using chromatin obtained from Srf KO macrophages. An example of SRF ChIP-seq data from both WT and KO cells for a portion of chromosome 10 is shown as a genome browser image (19) in Fig. Fig.4A.4A. Those peaks that were present only in the Srf WT and not in Srf KO cells were considered to be specific to SRF (1,262 peaks), while peaks with similar tag counts in both WT and KO cells were considered to be nonspecific background of the SRF antibody and were excluded from the study (see Table S3 in the supplemental material). Each of these peaks was then annotated using the closest RefSeq gene observed. Gene ontology analysis of this set of SRF-bound genes was performed using the genomic region enrichment of annotation tool (GREAT) and recovered actin cytoskeleton organization as the most enriched term by 3 orders of magnitude (Fig. (Fig.4B),4B), consistent with the previously identified transcriptional and functional consequences of SRF deletion (29).
FIG. 4.
FIG. 4.
Genomic location annotation and motif analysis of SRF ChIP-seq peaks in primary macrophages. (A) UCSC genome browser image of ChIP-seq peak data for part of chromosome 10 (chr10) in primary macrophages. All experiments were normalized to 10 million tags. (more ...)
Classification of SRF peaks with respect to the locations of genes identified by RefSeq indicated that just over 30% were located in the vicinity of promoters (within ±1 kb of a transcriptional start site), representing a significant enrichment over what was predicted based on a random binding profile of the total genome. Approximately two-thirds of SRF peaks were found to be located in intergenic regions and in introns, however, raising the possibility that SRF plays a significant role in the regulation of macrophage gene expression through distal elements (Fig. (Fig.4C).4C). Of the total set of 1,262 SRF binding sites, 1,055 were associated with genes that were expressed in macrophages (based on the microarray data from Fig. Fig.33 and Table S1 in the supplemental material). Out of these 1,055 binding sites, 64 (6%) were associated with significant changes in expression of the nearest gene in SRF KO macrophages (P < 10−6 compared to the number expected by chance, as determined by the Fisher exact test) (see Table S3 in the supplemental material).
To determine the factors that may functionally cooperate with SRF in macrophages, we performed de novo motif analysis of the significant SRF peak regions (within ±100 bp of the peak position) using the HOMER software suite and MEME (Fig. (Fig.3D3D and data not shown) (4, 15). As expected, the most enriched sequence identified by each program was the known SRF consensus-binding motif (CArG box), which was found in 42% of the SRF peaks (Fig. 4D and E). More than 90% of the SRF motifs identified were within 50 bp of the center of the peak (see Fig. 6A), strongly suggesting that SRF binding to these sites is through direct sequence-specific interactions with the CArG box. Significant enrichment was also observed for motifs for the macrophage- and B cell-specific Ets factor PU.1 and the ubiquitous transcription factors SP1 (GC box) and AP-1, found to be present in 39%, 20%, and 15% of the SRF peaks, respectively (Fig. 4D and E). It is important to note that although both PU.1 and TCFs are Ets factors, recent work has shown that the subtle, yet distinct, sequence specificities within the Ets family allow the motifs for these factors to be discriminated (59).
Distinct sequence characteristics of proximal and distal SRF binding sites.
To further investigate the mechanisms responsible for the genomic pattern of SRF binding, de novo motif analysis was selectively performed on SRF peaks associated with promoters (−1 kb to +1 kb, 363 peaks) and those peaks associated with distal sites (all areas, excluding promoters, 899 peaks). SRF peak sequences (±100 bp from the center of the peak) in proximal promoter regions in primary macrophages were enriched for CRE, SP1, Ets, and nuclear factor Y (NFY) motifs, in addition to SRF-specific CArG boxes (Fig. (Fig.5,5, top). In contrast, SRF peaks that were distal from target genes were selectively enriched for SRF, PU.1, and AP-1 motifs (Fig. (Fig.5,5, bottom). Surprisingly, the CArG box was identified in significantly more distal SRF ChIP-seq peaks than in proximal peaks (P < 10−15, Fisher exact test), while CRE, SP1, and NFY motifs were observed in significantly more proximal peaks. These results are consistent with the high enrichment of the latter motifs in promoter-proximal regions in general (63). In contrast to the proximal promoter motifs, PU.1 and AP-1 motifs were found in significantly more distal peaks than proximal peaks.
FIG. 5.
FIG. 5.
SRF associates with different transcription factor (TF) motifs depending on genomic location. HOMER de novo motif analysis of SRF-specific ChIP-seq peak sequences located in the promoter-proximal (−1 kb to +1 kb) (top) and distal (bottom) (more ...)
SRF associates with PU.1 at distal sites.
To further investigate the distal binding program of SRF, we plotted the aggregate spatial relationships of PU.1 and CArG motifs to the centers of distal SRF peaks. With this analysis, we were looking for any distinct spatial relationships between SRF binding (presumably the center of the peak) and the location of the associated PU.1 motif that might suggest a novel SRF/PU.1-dependent regulatory element. While the CArG motif itself is sharply focused at the center of the SRF peak, PU.1 motifs exhibit a relatively broad distribution over approximately ±100 bp (Fig. (Fig.6A).6A). To confirm that the SRF association with PU.1 motifs at distal sites correlates with actual PU.1 binding, we compared levels of SRF and PU.1 binding in primary macrophages (PU.1 ChIP-seq data from reference 15). Figure Figure6B6B shows a scatter plot representation of the combined SRF and PU.1 ChIP-seq peaks that were determined to be SRF positive (blue triangles), PU.1 positive (gray diamonds), and SRF/PU.1 positive (red squares). Remarkably, over 50% of all SRF sites are in close proximity to PU.1 binding sites (P [double less-than sign] 10−100, Fisher exact test). Coimmunoprecipitation experiments failed to show direct binding of these two factors, consistent with the broad distribution of PU.1 motifs in the SRF peak sequences and the results of previous studies that failed to observe ternary-complex formation between SRF and PU.1 (48, 58) (Fig. (Fig.6A6A and data not shown).
FIG. 6.
FIG. 6.
Distal SRF peaks are associated with PU.1 and the H3K4me1 enhancer mark. (A) Histogram analysis of the locations of the indicated DNA sequence motifs relative to the center of distal SRF ChIP-seq peaks (excluding promoters) in primary macrophages. Motif (more ...)
Recent studies have shown that PU.1 located at distal regions of the genome is highly associated with chromatin modifications ascribed to transcriptional enhancers (11, 15). In order to determine the SRF associations with chromatin marks with and without PU.1, we analyzed the associations of SRF and PU.1 with mono- and trimethylated lysine 4 of histone 3 (H3K4me1 and H3K4me3, respectively) as marks of enhancers and promoters, respectively, using hierarchical clustering of the ChIP-seq data for each factor/mark in primary macrophages (13-15). The overall association between SRF peaks and peaks from each of the other experiments was highly nonrandom, with observations of cobinding within 100 bp being several orders of magnitude higher than expected (P [double less-than sign] 10−100, Fisher exact test). As expected, a large subset of SRF peaks were closely associated with H3K4me3 (Fig. (Fig.6C),6C), which has been shown to be highly associated with poised or active promoters (30). This association also occurs with relatively low levels of PU.1 and the highly PU.1-associated factor C/EBPβ, which have been described to be localized primarily to distal genomic regions (15). In the case where SRF is bound together with PU.1, there is enrichment of the H3K4me1 enhancer mark directly adjacent to the location of factor binding, while the H3K4me3 promoter mark is relatively absent, suggesting SRF and PU.1 association at putative enhancers (Fig. (Fig.6C6C).
In order to further assess the requirement for PU.1 and SRF binding at distal enhancer sites, we used PU.1-deficient myeloid precursor cells (PU.1−/−) and the tamoxifen-inducible PU.1 cell line PUER for ChIP-seq analysis. PUER cells are PU.1−/− myeloid precursor cells that have been transduced with PU.1 that has been fused to the ligand-binding domain of the estrogen receptor (54). Treatment with tamoxifen activates PU.1 and initiates a gene expression program similar to that observed during monocyte-to-macrophage differentiation (22). To characterize the requirement of PU.1 for SRF binding, ChIP-seq for SRF was performed in PU.1−/− and differentiated PUER cells (at 24 h). To ensure SRF specificity, analysis was focused on those SRF peak locations identified in primary macrophages that were also SRF targets in differentiated PUER cells (831 out of 1,262 peaks). As shown in the Venn diagram in Fig. Fig.7A,7A, 252 of these peaks were observed in PU.1−/− cells, constituting a PU.1-independent subset of SRF binding sites. Remarkably, the remaining two-thirds (579) of the SRF peaks were bound only by SRF in differentiated PUER cells, establishing the majority of SRF sites as being PU.1 dependent in this cell type. Motif analysis of the PU.1-independent and PU.1-dependent subsets of SRF peak sequences indicated that CArG boxes and GC boxes were enriched preferentially in the PU.1-independent peaks (P < 0.001, Fisher exact test). As expected, PU.1 motifs were preferentially enriched in PU.1-dependent peaks (P < 10−8). Conversely, SP1 sites were preferentially enriched at PU.1-independent sites (P < 0.0001) (Fig. (Fig.7B7B).
FIG. 7.
FIG. 7.
The genomic localization of SRF during macrophage differentiation is PU.1 dependent. (A) Venn diagram of the SRF ChIP-seq peaks identified in PU.1−/− cells and differentiated PUER cells (PUER-24h) that were found to be in common with the (more ...)
To specifically look at SRF association with PU.1 at enhancers, we first confirmed that SRF binding at distal enhancer targets identified by the ChIP-seq experiment are indeed PU.1 dependent but that binding of SRF to proximal promoter targets occurs in a PU.1-independent manner (Fig. (Fig.7C).7C). To further characterize SRF and PU.1 binding to enhancers, we analyzed SRF, PU.1, and H3K4me1 ChIP-seq data from PU.1−/− and differentiated PUER cells, using PU.1 and H3K4me1 ChIP-seq data from the work of Heinz et al. (15). For this analysis, we distilled our data set to the 428 SRF peaks in differentiated PUER cells that were located more than 3 kb from the TSS to avoid interference with H3K4me1 signal trailing from strong areas of H3K4me3-marked promoters (15). Out of these 428 peaks, 405 are acquired during PU.1-induced macrophage differentiation (PUER cell SRF tag count [24 h]/PU.1−/− SRF tag count > 2), 227 of which are within 100 bp of a PU.1 binding site (bound in differentiated PUER cells). Using histogram analysis, we determined that acquisition of SRF to distal enhancer sites during PU.1-induced differentiation is dependent on PU.1 binding by comparing the levels of the H3K4me1 signal observed at SRF-PU.1-cobound sites to the signal observed at sites bound only by SRF. SRF acquisition at distal sites that are in close proximity to a PU.1 peak (<100 bp, bound) is highly correlated with acquisition of the H3K4me1 enhancer mark (Fig. (Fig.7D,7D, red versus green lines). In contrast, when SRF is acquired during differentiation at sites that are distant from a PU.1 peak (>100 bp away or absent), the overall H3K4me1 signal is less and there is very little differentiation-dependent change (Fig. (Fig.7D,7D, blue versus gray lines).
Recent studies have suggested that PU.1 and collaborative lineage-determining factors are required for binding of the liver X receptor (LXR) transcription factors in macrophages but that LXRs are not required for the binding of PU.1 (15). To determine whether a similar relationship exists between PU.1 and SRF, we performed ChIP sequencing analysis of PU.1 in primary macrophages from Srf KO mice. Using dot plot analysis of tag counts at PU.1 peak locations in Srf WT and KO primary macrophages, we observed that PU.1 binding has little to no dependence on the presence of SRF (Fig. (Fig.7E,7E, gray diamonds). We also failed to identify any subsets of SRF peaks that were preferentially bound by PU.1 based on the Srf genotype (i.e., all of the SRF-bound PU.1 peaks are also located along the diagonal), indicating that PU.1 binding is not dependent on the presence of SRF during macrophage differentiation (Fig. (Fig.7E,7E, red squares).
SRF binding and factor association is cell type specific.
Based on our data that shows that SRF is highly associated with the macrophage- and B cell-restricted factor PU.1 in differentiated macrophages and that the acquisition of SRF at genomic sites during differentiation is PU.1 dependent (Fig. (Fig.44 and and7),7), we sought to determine whether the SRF genomic binding profile is cell type specific. Using mouse SRF ChIP-seq data in neurons that were recently published (20), we compiled the percentages of macrophage-specific, neuron-specific, and commonly bound genomic sites that were located in proximal promoter regions (±1 kb of the TSS). Approximately 60% of the genomic sites that were bound by SRF in both cell types were found within proximal promoters (Fig. (Fig.8A).8A). Conversely, only about 25% of genomic sites occupied by SRF in a macrophage- or neuron-specific manner were located within 1 kb of the TSS. Motif analyses of the macrophage-specific, neuron-specific, and common SRF peak sequences returned the CArG box as the most enriched motif in each class of peaks, as expected. Analysis of the peaks common to macrophages and neurons resulted in recovery of motifs for the ubiquitously expressed SP1 and GA-binding protein alpha (GABPa)/Ets factors. In contrast, motif analysis of macrophage-specific peak sequences yielded enrichment for PU.1, Krüppel-like factor (KLF)/SP1, and AP-1 recognition sequences, while neuron-specific SRF peaks contained recognition motifs for CTCF and regulatory factor X (RFX) (Fig. 8B and C).
FIG. 8.
FIG. 8.
SRF binding to the genome is cell type specific. (A) Percentages of macrophage-specific, neuron-specific, or common SRF ChIP-Seq peaks that were located within 1 kb of a TSS. (B) De novo motif analysis of cell type-dependent and -independent SRF peak (more ...)
To further investigate the relationship between SRF and PU.1 in establishing cell-specific binding sites for PU.1, we plotted the normalized SRF ChIP-seq tag counts in primary macrophages versus neurons and color coded those peaks corresponding to genomic locations also occupied by PU.1 (within ±100 bp) in macrophages (Fig. (Fig.8D).8D). This analysis indicated that the sites cobound by PU.1 in macrophages are mostly (~80%) restricted to the macrophage-specific subset. Furthermore, using the SRF ChIP-seq data in differentiated PUER cells, we also determined that those SRF peak locations that are cobound in neurons were significantly enriched in the PU.1-independent subset of SRF binding sites (P [double less-than sign] 10−10, Fisher exact test) (Fig. (Fig.8E).8E). Conversely, those SRF sites that were also cobound by PU.1 were significantly enriched in the PU.1-dependent subset. Collectively, these data suggest that SRF associates with cell type-restricted factors primarily at distal sites to mediate expression of cell type-restricted genes.
SRF and PU.1 regulate the expression of hematopoietic-system-specific cytoskeletal genes.
Several of the SRF target genes that were identified by the expression microarray analysis (Fig. (Fig.3C)3C) of macrophages are cytoskeletal genes whose expression is restricted to hematopoietic cells (Lsp1, Coro1a, and Lcp1) (10, 27). Further analysis of the ChIP-seq data (described in the legend to Fig. Fig.44 and in reference 15) for these target genes showed that there are distal, but not promoter-proximal, SRF and PU.1 peaks and H3K4me1 signal associated with these genes, exemplified by Lsp1 in Fig. Fig.9A9A (the distal SRF peaks associated with Coro1a and Lcp1 are listed in Table S3 in the supplemental material). In order to address whether the association between SRF and PU.1 is of functional significance, the expression and regulation of these target genes were examined. In the absence of PU.1 activation in PUER cells, Lsp1, Coro1a, and Lcp1 mRNAs were detected at very low levels, but following macrophage differentiation induced by PU.1 activation with tamoxifen, transcript levels were dramatically upregulated (Fig. (Fig.9B).9B). Relative Srf mRNA expression levels were not affected by tamoxifen treatment of these cells (data not shown). These results suggest that expression of these SRF-dependent target genes is also PU.1 dependent.
FIG. 9.
FIG. 9.
SRF and PU.1 regulate hematopoietic-cell-specific cytoskeletal gene expression in macrophages. (A) UCSC genome browser images of H3K4me1, PU.1, and SRF ChIP-seq peaks found near the Lsp1 gene in PU.1−/−, differentiated PUER cells and primary (more ...)
To determine whether both SRF and PU.1 regulate these cytoskeletal genes in primary macrophages, siRNA knockdown experiments were performed. Partial reduction of the expression of either PU.1 or Srf in primary macrophages decreased the levels of Lsp1, Coro1a, and Lcp1 mRNAs compared to levels in cells transfected with a nonspecific siRNA control (Fig. 9C and D). Taken together, these data suggest that SRF and PU.1 functionally collaborate to regulate the expression of lineage-restricted, cytoskeletal target genes in macrophages.
Many specialized cell types require both general and lineage-restricted cytoskeleton components to execute their various biological functions, but the molecular mechanisms that establish cell-specific programs of cytoskeletal gene expression remain poorly understood. Here, we provide evidence that SRF regulates both general and cell type-specific programs of cytoskeletal gene expression in macrophages, which results in defects in specialized functions, such as cell spreading, invasion, and phagocytosis (Fig. (Fig.2).2). These results are consistent with recently published work from Ragu et al. (39) that showed that SRF regulates the expression of cytoskeletal genes in hematopoietic stem cells and that Srf-deficient hematopoietic stem cells lack the ability to properly adhere to and be retained in the bone marrow. In contrast with what we found, they also observed an increase in CD11b+ GR1+ progenitors, suggesting that proper myeloid differentiation is also SRF dependent. These differences may be due to the shorter time course of our studies (harvest at 3 to 6 days [1 injection] versus 5 weeks [3 injections] after pIpC injection). This shorter time course and single-injection protocol was used because the Srffl/fl Mx-Cre+ mice used in our studies became visibly ill following a three-injection protocol and more-extended studies could not be performed (data not shown).
SRF has previously been considered to function primarily at promoters, in part because many of the SRF target genes that have previously been described have CArG boxes in the proximal promoter regions (i.e., c-Fos, Tagln, Srf, etc.) (25, 49, 51). Additionally, bioinformatic studies have suggested that SRF binding occurs mainly at proximal sites because almost all CArG boxes shown to be functional based on RNA interference (RNAi) and luciferase reporter assays were found to be located within 4 kb of the TSS (46, 50). The current ChIP-seq analysis indicates that SRF binding is highly enriched at promoters in comparison to binding at a random distribution of sites (Fig. (Fig.4C)4C) but that the majority of binding sites (60%) in macrophages are more than ±4 kb from the TSS. SRF binding in the vicinity of promoters was highly correlated with binding motifs for ubiquitously expressed transcription factors, such as SP1, and a profile of histone modifications typical of active promoters. These findings are thus consistent with the well-established roles of SRF in regulating the expression of ubiquitous cytoskeletal components, such as β-actin and vinculin.
In contrast, binding of SRF to distal genomic locations was highly associated with and dependent upon the macrophage- and B cell-specific transcription factor PU.1. Recent studies provided evidence that PU.1, in concert with limited sets of other lineage-determining transcription factors, establishes the majority of the cell-specific enhancer-like elements in macrophages and B cells (15). These sites have been proposed to provide subsequent access to signal-dependent factors, such as nuclear receptors, to enable acquisition of cell-specific responses to incoming signals. The present studies extend this concept to SRF-dependent programs of gene expression in which the binding of PU.1 is a prerequisite for the establishment of the distal DNA binding program of SRF at a subset of macrophage-specific enhancers. These sites in turn are suggested to regulate the expression of hematopoietic-cell-restricted cytoskeletal genes, such as Lsp1, Coro1a, and Lcp1, which are required for specialized macrophage functions (Fig. (Fig.9).9). Thus, we hypothesize that SRF regulates both general and cell type-specific programs of gene expression in macrophages through the use of two distinct mechanisms: a promoter-based strategy that controls expression of ubiquitously expressed genes and a PU.1-dependent enhancer-based strategy that controls the expression of cell-restricted genes.
Consistent with this model, SRF has previously been shown to bind to a distal enhancer region for the myocyte-specific transcription factor MyoD in skeletal muscle cells (24), while in macrophages, MyoD is not expressed and SRF binding to this site was not observed (data not shown). In addition, our analysis of recently published SRF ChIP-seq data from neurons shows that the majority of both macrophage- and neuron-specific SRF binding sites are located distally (more than ±1 kb from the TSS) but that the majority of binding sites common to both cell types occur in the proximal promoter region (Fig. (Fig.8A).8A). Furthermore, motif analysis of the distal SRF peak sequences in the neuron data set showed enrichment of a binding site for RFX factors, many of which are highly expressed in the brain and play important roles in development (1, 5, 28). Based on our analysis, it is likely that SRF collaborates with lineage-determining factors analogously to PU.1 in other cell types to establish corresponding cell-specific programs of gene expression.
Despite our identification of 1,055 SRF binding sites in the genome, relatively few (6%) were actually associated with genes that were differentially regulated by the microarray analysis. However, this percentage is similar to that observed for other sequence-specific transcription factors (e.g., LXRs [15]) and is consistent with the possibility that many SRF DNA binding events are nonproductive in the absence of a stimulus or that the normal functions of SRF at a subset of its binding sites can be compensated for by other transcription factors. Consistent with the former possibility, we have observed that a number of genes with vicinal SRF binding sites that are not affected by loss of SRF under basal conditions are highly dependent on SRF for transcriptional activation by pattern recognition receptors (A. L. Sullivan and L. Xie, unpublished data). It is also possible that SRF acts at some of these sites to exert context-dependent functions or to regulate the expression of noncoding RNAs that were not represented on the microarrays used for transcriptome analysis. In addition, due to the large distances over which enhancers may exert their functions, it is likely that the simple linkage of SRF binding sites to the nearest transcription units misses functionally important interactions. In this case, employment of more advanced genomic techniques, such as genome-wide chromosome conformation studies, may be informative.
In conclusion, we demonstrate that SRF is required for normal macrophage migration and phagocytosis and that it functions to regulate both ubiquitous and hematopoietic-cell-specific cytoskeletal gene targets. By mapping and analyzing the SRF cistrome, we identified and validated gene targets that suggest that cell-specific target gene expression is conferred by collaborative interactions with PU.1 at distal enhancer-like regions, providing insights into how a ubiquitously expressed transcription factor mediates concurrent programs of both cell type- and non-cell type-specific gene expression.
Supplementary Material
[Supplemental material]
A.L.S. was supported by a grant from the American Heart Association. These studies were also supported by NIH grants P01-HC088093 and R01 CA52599 to C.K.G. and HL091168 to J.M.M. C.K.G. acknowledges support from a Leducq Transatlantic Network grant. We also acknowledge the UCSD Neuroscience Microscopy Shared Facility (grant P30 NS047101).
We thank James Sprague, Roman Sasik, and Colleen Ludka of the UCSD Biogem Core facility for their microarray and ChIP sequencing services and M. G. Farquhar for microscope access. Mx-CRE mice were a kind gift of L. Goldstein. We thank Lynn Bautista for her assistance with preparation of the figures.
[down-pointing small open triangle]Published ahead of print on 6 December 2010.
Supplemental material for this article may be found at
1. Aftab, S., L. Semenec, J. S. Chu, and N. Chen. 2008. Identification and characterization of novel human tissue-specific RFX transcription factors. BMC Evol. Biol. 8:226. [PMC free article] [PubMed]
2. Alberti, S., et al. 2005. Neuronal migration in the murine rostral migratory stream requires serum response factor. Proc. Natl. Acad. Sci. U. S. A. 102:6148-6153. [PubMed]
3. Arsenian, S., B. Weinhold, M. Oelgeschlager, U. Ruther, and A. Nordheim. 1998. Serum response factor is essential for mesoderm formation during mouse embryogenesis. EMBO J. 17:6289-6299. [PubMed]
4. Bailey, T. L., and C. Elkan. 1994. Fitting a mixture model by expectation maximization to discover motifs in biopolymers. Proc. Int. Conf. Intell. Syst. Mol. Biol. 2:28-36. [PubMed]
5. Blackshear, P. J., et al. 2003. Graded phenotypic response to partial and complete deficiency of a brain-specific transcript variant of the winged helix transcription factor RFX4. Development 130:4539-4552. [PubMed]
6. Cen, B., et al. 2003. Megakaryoblastic leukemia 1, a potent transcriptional coactivator for serum response factor (SRF), is required for serum induction of SRF target genes. Mol. Cell. Biol. 23:6597-6608. [PMC free article] [PubMed]
7. Delanote, V., J. Vandekerckhove, and J. Gettemans. 2005. Plastins: versatile modulators of actin organization in (patho)physiological cellular processes. Acta Pharmacol. Sin. 26:769-779. [PubMed]
8. Dennis, G., Jr., et al. 2003. DAVID: Database for Annotation, Visualization, and Integrated Discovery. Genome Biol. 4:P3. [PubMed]
9. Ernst, W. H., R. Janknecht, M. A. Cahill, and A. Nordheim. 1995. Transcriptional repression mediated by the serum response factor. FEBS Lett. 357:45-49. [PubMed]
10. Ferrari, G., H. Langen, M. Naito, and J. Pieters. 1999. A coat protein on phagosomes involved in the intracellular survival of mycobacteria. Cell 97:435-447. [PubMed]
11. Ghisletti, S., et al. 2010. Identification and characterization of enhancers controlling the inflammatory gene expression program in macrophages. Immunity 32:317-328. [PubMed]
12. Han, Z., X. Li, J. Wu, and E. N. Olson. 2004. A myocardin-related transcription factor regulates activity of serum response factor in Drosophila. Proc. Natl. Acad. Sci. U. S. A. 101:12567-12572. [PubMed]
13. Heintzman, N. D., et al. 2009. Histone modifications at human enhancers reflect global cell-type-specific gene expression. Nature 459:108-112. [PMC free article] [PubMed]
14. Heintzman, N. D., et al. 2007. Distinct and predictive chromatin signatures of transcriptional promoters and enhancers in the human genome. Nat. Genet. 39:311-318. [PubMed]
15. Heinz, S., et al. 2010. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell 38:576-589. [PMC free article] [PubMed]
16. Huang, Q. Q., et al. 2008. Role of H2-calponin in regulating macrophage motility and phagocytosis. J. Biol. Chem. 283:25887-25899. [PubMed]
17. Johansen, F. E., and R. Prywes. 1995. Serum response factor: transcriptional regulation of genes induced by growth factors and differentiation. Biochim. Biophys. Acta 1242:1-10. [PubMed]
18. Jongstra-Bilen, J., and J. Jongstra. 2006. Leukocyte-specific protein 1 (LSP1): a regulator of leukocyte emigration in inflammation. Immunol. Res. 35:65-74. [PubMed]
19. Kent, W. J., et al. 2002. The human genome browser at UCSC. Genome Res. 12:996-1006. [PubMed]
20. Kim, T. K., et al. 2010. Widespread transcription at neuronal activity-regulated enhancers. Nature 465:182-187. [PMC free article] [PubMed]
21. Kuhn, R., F. Schwenk, M. Aguet, and K. Rajewsky. 1995. Inducible gene targeting in mice. Science 269:1427-1429. [PubMed]
22. Laslo, P., et al. 2006. Multilineage transcriptional priming and determination of alternate hematopoietic cell fates. Cell 126:755-766. [PubMed]
23. Lattin, J. E., et al. 2008. Expression analysis of G protein-coupled receptors in mouse macrophages. Immunome Res. 4:5. [PMC free article] [PubMed]
24. L'honore, A., et al. 2003. MyoD distal regulatory region contains an SRF binding CArG element required for MyoD expression in skeletal myoblasts and during muscle regeneration. Mol. Biol. Cell 14:2151-2162. [PMC free article] [PubMed]
25. Li, L., Z. Liu, B. Mercer, P. Overbeek, and E. N. Olson. 1997. Evidence for serum response factor-mediated regulatory networks governing SM22alpha transcription in smooth, skeletal, and cardiac muscle cells. Dev. Biol. 187:311-321. [PubMed]
26. Li, S., et al. 2005. Requirement for serum response factor for skeletal muscle growth and maturation revealed by tissue-specific gene deletion in mice. Proc. Natl. Acad. Sci. U. S. A. 102:1082-1087. [PubMed]
27. Lin, C. S., T. Park, Z. P. Chen, and J. Leavitt. 1993. Human plastin genes. Comparative gene structure, chromosome location, and differential expression in normal and neoplastic cells. J. Biol. Chem. 268:2781-2792. [PubMed]
28. Ma, K., S. Zheng, and Z. Zuo. 2006. The transcription factor regulatory factor X1 increases the expression of neuronal glutamate transporter type 3. J. Biol. Chem. 281:21250-21255. [PubMed]
29. McLean, C. Y., et al. 2010. GREAT improves functional interpretation of cis-regulatory regions. Nat. Biotechnol. 28:495-501. [PubMed]
30. Mellor, J., P. Dudek, and D. Clynes. 2008. A glimpse into the epigenetic landscape of gene regulation. Curr. Opin. Genet. Dev. 18:116-122. [PubMed]
31. Miano, J. M. 2003. Serum response factor: toggling between disparate programs of gene expression. J. Mol. Cell. Cardiol. 35:577-593. [PubMed]
32. Miano, J. M., X. Long, and K. Fujiwara. 2007. Serum response factor: master regulator of the actin cytoskeleton and contractile apparatus. Am. J. Physiol. Cell Physiol. 292:C70-C81. [PubMed]
33. Miano, J. M., et al. 2004. Restricted inactivation of serum response factor to the cardiovascular system. Proc. Natl. Acad. Sci. U. S. A. 101:17132-17137. [PubMed]
34. Miralles, F., G. Posern, A. I. Zaromytidou, and R. Treisman. 2003. Actin dynamics control SRF activity by regulation of its coactivator MAL. Cell 113:329-342. [PubMed]
35. Mosser, D. M., and J. P. Edwards. 2008. Exploring the full spectrum of macrophage activation. Nat. Rev. Immunol. 8:958-969. [PMC free article] [PubMed]
36. Niu, Z., et al. 2005. Conditional mutagenesis of the murine serum response factor gene blocks cardiogenesis and the transcription of downstream gene targets. J. Biol. Chem. 280:32531-32538. [PubMed]
37. Parlakian, A., et al. 2004. Targeted inactivation of serum response factor in the developing heart results in myocardial defects and embryonic lethality. Mol. Cell. Biol. 24:5281-5289. [PMC free article] [PubMed]
38. Pollard, J. W. 2009. Trophic macrophages in development and disease. Nat. Rev. Immunol. 9:259-270. [PubMed]
39. Ragu, C., et al. 2010. The transcription factor SRF regulates hematopoietic stem cell adhesion. Blood 116:4464-4473. [PubMed]
40. Ramanan, N., et al. 2005. SRF mediates activity-induced gene expression and synaptic plasticity but not neuronal viability. Nat. Neurosci. 8:759-767. [PubMed]
41. Robertson, G., et al. 2007. Genome-wide profiles of STAT1 DNA association using chromatin immunoprecipitation and massively parallel sequencing. Nat. Methods 4:651-657. [PubMed]
42. Sasazuki, T., et al. 2002. Identification of a novel transcriptional activator, BSAC, by a functional cloning to inhibit tumor necrosis factor-induced cell death. J. Biol. Chem. 277:28853-28860. [PubMed]
43. Schratt, G., et al. 2002. Serum response factor is crucial for actin cytoskeletal organization and focal adhesion assembly in embryonic stem cells. J. Cell Biol. 156:737-750. [PMC free article] [PubMed]
44. Schratt, G., et al. 2001. Serum response factor is required for immediate-early gene activation yet is dispensable for proliferation of embryonic stem cells. Mol. Cell. Biol. 21:2933-2943. [PMC free article] [PubMed]
45. Selvaraj, A., and R. Prywes. 2003. Megakaryoblastic leukemia-1/2, a transcriptional co-activator of serum response factor, is required for skeletal myogenic differentiation. J. Biol. Chem. 278:41977-41987. [PubMed]
46. Shen, X., et al. 2009. The correlations of the function and positional distribution of the cis-elements CArG around the TSS in the genes of Mus musculus. Genome 52:217-221. [PubMed]
47. Shore, P., and A. D. Sharrocks. 1995. The MADS-box family of transcription factors. Eur. J. Biochem. 229:1-13. [PubMed]
48. Shore, P., and A. D. Sharrocks. 1994. The transcription factors Elk-1 and serum response factor interact by direct protein-protein contacts mediated by a short region of Elk-1. Mol. Cell. Biol. 14:3283-3291. [PMC free article] [PubMed]
49. Spencer, J. A., and R. P. Misra. 1996. Expression of the serum response factor gene is regulated by serum response factor binding sites. J. Biol. Chem. 271:16535-16543. [PubMed]
50. Sun, Q., et al. 2006. Defining the mammalian CArGome. Genome Res. 16:197-207. [PMC free article] [PubMed]
51. Treisman, R. 1986. Identification of a protein-binding site that mediates transcriptional response of the c-fos gene to serum factors. Cell 46:567-574. [PubMed]
52. Treisman, R., R. Marais, and J. Wynne. 1992. Spatial flexibility in ternary complexes between SRF and its accessory proteins. EMBO J. 11:4631-4640. [PubMed]
53. Valledor, A. F., et al. 2004. Activation of liver X receptors and retinoid X receptors prevents bacterial-induced macrophage apoptosis. Proc. Natl. Acad. Sci. U. S. A. 101:17813-17818. [PubMed]
54. Walsh, J. C., et al. 2002. Cooperative and antagonistic interplay between PU. 1 and GATA-2 in the specification of myeloid cell fates. Immunity 17:665-676. [PubMed]
55. Wang, D., et al. 2001. Activation of cardiac gene expression by myocardin, a transcriptional cofactor for serum response factor. Cell 105:851-862. [PubMed]
56. Wang, D. Z., et al. 2002. Potentiation of serum response factor activity by a family of myocardin-related transcription factors. Proc. Natl. Acad. Sci. U. S. A. 99:14855-14860. [PubMed]
57. Wang, Z., et al. 2004. Myocardin and ternary complex factors compete for SRF to control smooth muscle gene expression. Nature 428:185-189. [PubMed]
58. Watson, D. K., et al. 1997. FLI1 and EWS-FLI1 function as ternary complex factors and ELK1 and SAP1a function as ternary and quaternary complex factors on the Egr1 promoter serum response elements. Oncogene 14:213-221. [PubMed]
59. Wei, G. H., et al. 2010. Genome-wide analysis of ETS-family DNA-binding in vitro and in vivo. EMBO J. 29:2147-2160. [PMC free article] [PubMed]
60. Weinhold, B., et al. 2000. Srf(−/−) ES cells display non-cell-autonomous impairment in mesodermal differentiation. EMBO J. 19:5835-5844. [PubMed]
61. Wickramasinghe, S. R., et al. 2008. Serum response factor mediates NGF-dependent target innervation by embryonic DRG sensory neurons. Neuron 58:532-545. [PMC free article] [PubMed]
62. Wu, C., et al. 2009. BioGPS: an extensible and customizable portal for querying and organizing gene annotation resources. Genome Biol. 10:R130. [PMC free article] [PubMed]
63. Xie, X., et al. 2005. Systematic discovery of regulatory motifs in human promoters and 3′ UTRs by comparison of several mammals. Nature 434:338-345. [PMC free article] [PubMed]
Articles from Molecular and Cellular Biology are provided here courtesy of
American Society for Microbiology (ASM)