Transforming growth factor-β1 (TGF-β1) is known to induce epithelial-mesenchymal transition in the kidney, a process involved in tubulointerstitial fibrosis. We hypothesized that a coactivator of the serum response factor (SRF), megakaryoblastic leukemia factor-1 (MKL1), stimulates α-smooth muscle actin (α-SMA) transcription in primary cultures of renal tubular epithelial cells (RTC), which convert into myofibroblasts on treatment with TGF-β1. Herein, we study the effect of MKL1 expression on α-SMA in these cells. We demonstrate that TGF-β1 stimulation of α-SMA transcription is mediated through CC(A/T)6-rich GG elements known to bind to SRF. These elements also mediate the MKL1 effect that dramatically activates α-SMA transcription in serum-free media. MKL1 fused to green fluorescent protein localizes to the nucleus and induces α-SMA expression regardless of treatment with TGF-β1. Using proteasome inhibitors, we also demonstrate that the proteolytic ubiquitin pathway regulates MKL1 expression. These data indicate that MKL1 overexpression is sufficient to induce α-SMA expression. Inhibition of endogenous expression of MKL1 by small interfering RNA abolishes TGF-β1 stimulation of α-SMA expression. Therefore, MKL1 is also absolutely required for TGF-β1 stimulation of α-SMA expression. Western blot and immunofluorescence analysis show that overexpressed and endogenous MKL1 are located in the nucleus in non-stimulated RTC. Chromatin immunoprecipitation assay demonstrates that TGF-β1 induces binding of endogenous SRF and MKL1 to the α-SMA promoter in chromatin. Since MKL1 constitutes a potent factor regulating α-SMA expression, modulation of endogenous MKL1 expression or activity may have a profound effect on myofibroblast formation and function in the kidney.
- epithelial-mesenchymal transition
renal fibrosis is a common feature of various kidney diseases leading to end-stage renal failure (15, 44). This process is characterized by the accumulation of myofibroblasts defined by the expression of α-smooth muscle actin (α-SMA). These cells are major contributors to the increased extracellular matrix deposition seen in kidney fibrosis (16, 69). A number of studies demonstrate that renal tubular cells (RTC) can convert to myofibroblasts on epithelial-mesenchymal transition (EMT) stimulated by transforming growth factor-β (TGF-β) (9, 11, 24, 45, 69).
The regulation of α-SMA transcription has been extensively studied in smooth muscle cells and in cells from the myocardium and skeletal muscle, which express α-SMA in adults and embryos, respectively (66). Studies on the α-SMA promoter from chickens, rats, mice, and humans highlight the importance of cell context and species differences for α-SMA transcriptional regulation (1, 5, 28, 56). For example, a distinct region of the rat α-SMA promoter modulates differentially the promoter activity in smooth muscle cells compared with non-smooth muscle cells (56). While the segment −125 bp downstream of rat α-SMA promoter is necessary and sufficient to mediate TGF-β activation in rat smooth muscle cells and fibroblasts (17), the mouse α-SMA promoter requires the region −191 to −150 bp for full basal and TGF-β1-stimulated activity in mouse fibroblasts (5). These studies clearly emphasize that α-SMA promoter activities are regulated through common and distinct mechanisms in different species and cell types.
The overall α-SMA gene structure in different species is similar. The human gene contains a TATA box, located 23 bp upstream from the transcription initiation (+1) site, followed by a 41-bp sequence that encompasses the noncoding exon 1, a 3,812-bp intron, followed by 34 bp of exon 2 and then the start codon (51). Also, the α-SMA promoter of different species contains a number of highly conserved regions with consensus binding motifs for transcription factors. The 5′-flanking sequence upstream of the TATA box contains several cis-acting elements including two different CC(A/T-rich)6GG (CArG) boxes (named A and B), a MCAT motif, an E box, a TGF-β1 control element, a TGF-β1 hypersensitivity region, and a conserved TGTTTATC sequence (5, 17, 19, 25, 28, 35). Recent studies indicate that different regulatory mechanisms induce α-SMA promoter activity in renal epithelial cells that convert to myofibroblasts during EMT (12, 54, 62).
Myocardin has been identified as a serum response factor (SRF), myocyte-specific coactivator essential for heart development and sufficient to induce smooth muscle differentiation (63–65). Myocardin is homologous to two other factors characterized as myocardin-related factors, MRTF-A and MRTF-B (64). These factors are also known as MAL/MKL1 (megakaryocytic acute leukemia and megakaryoblastic leukemia factor-1) originally identified from a gene rearranged and expressed in t(1; 22)(p13;q13) acute megakaryoblastic leukemia (34, 38), BSAC and MKL2 (34, 38, 53, 55). In contrast to the specific tissue expression of myocardin, these two homologous genes are ubiquitously expressed (8) and also act as SRF coactivators. SRF-dependent transcription is mediated through CArG consensus sequences (65) that are found in the promoter of several immediate-early response genes as well as muscle-specific genes including α-SMA (18, 35). Rho signaling pathways are known to regulate SRF activity (30, 57). Rho family members are small GTPases involved in actin dynamics mediating actin polymerization and promoting assembly of globular actin (G-actin) to filamentous actin (F-actin) organized in stress fibers (2, 42, 57). TGF-β1 stimulates RhoA activity and activates EMT of renal proximal epithelial cells (36, 59). As a result of Rho activity, it is thought that MKL1, which is tethered to monomeric G-actin and continuously shuttles between the cytosol and the nucleus, accumulates in the nucleus, where it activates transcription (61). Recent studies indicate that MKL1 plays an important role during EMT in various renal epithelial cell lines (12, 40, 54). We also recently hypothesized that MKL1 stimulates α-SMA transcription in primary renal tubular epithelial cells (RTC) (9). Herein, we analyze the effect of MKL1 expression in TGF-β1-induced α-SMA expression during transition of RTC to myofibroblasts.
Cell culture and treatment.
RTC were isolated from human kidney cortex and maintained as previously described by our laboratory (10, 60). These cells have been extensively characterized and found capable of forming tubules in three-dimensional gels (10). RTC were propagated in MEMα medium (Invitrogen, Carlsbad, CA) supplemented with 10 ng/ml EGF, 5 μg/ml hydrocortisone (Sigma, St. Louis, MO), 6.25 μg/ml insulin, 6.25 μg/ml transferrin, 6.25 ng/ml selenous acid (Biowhittaker, Walkersville, MD), and 10% (vol/vol) heat-inactivated FBS (Atlanta Biologicals, Norcross, GA), penicillin, and streptomycin in a 37°C-5% CO2 incubator. Cell cultures were split 1:3 or 1:4 when confluent using a 0.05% Trypsin-EDTA solution (Invitrogen) in T75 flasks. RTC were used for experiments between passages 3 and 5 at 80–100% confluence in six-well polystyrene tissue culture plates. EMT was induced by incubation in MEMα medium (Invitrogen) supplemented with 10 ng/ml recombinant human TGF-β1 (R&D Systems, Minneapolis, MN), which was changed every 2–3 days (9). Inhibitors E64, MG132, Z-LLF, and chloroquine were purchased from Sigma. Our study protocols were approved by the Institutional Review Board of The University of Oklahoma Health Sciences Center.
The −2580/+51 sequence of the α-SMA gene was subcloned from BAC Homo sapiens clone number RP11-115N17 (GenBank accession no. AC015461, purchased from BACPAC Resources Center, Children's Hospital Oakland Research Institute, Oakland, CA) by inserting the XbaI/KasI restriction fragment into the PGL3 basic vector containing the luciferase gene (Promega, Madison, WI). The cloning was confirmed by DNA sequencing. This sequence contains the promoter sequence, exon 1, and 7 bp of intron 1. The −893/+51 and −259/+51 sequences were generated using EcoRI and PstI restriction sites, respectively. The −259/+51 α-SMA promoter sequence was also subcloned into the pmaxFP-Red-PRL vector (Amaxa, Gaithersburg, MD) to drive the expression of red fluorescence protein (RFP). Site-directed mutagenesis was performed using a QuickChange kit (Stratagene, La Jolla, CA) using primers containing EcoRI sites and confirmed by DNA sequencing. CArG A and B wild-type CCTTGTTTGG and CCCTATATGG sequences were replaced by CCTTGAATTC and CCCGAATTCG sequences, respectively. MKL1 and MKL2 cDNAs in the p3xflag-CMV-7.1 expression vector (4, 55) were kindly provided by Dr. R. Prywes (Columbia University, New York, NY). The MKL1/green fluorescent protein (GFP) fusion gene was cloned by directed site mutagenesis of the MKL1 STOP codon and subcloning of the mutated MKL1 gene in-frame in the pEGFP-N1 vector (Clontech, Mountain View, CA). In this construct, the flag tag sequence expressed at the NH2 terminus of the protein was replaced by a HA tag.DNA. Plasmids were purified using a maxiprep kit (Qiagen, Valencia, CA).
α-SMA promoter activity.
RTC in suspension were transiently transfected with DNA (1 μg/well α-SMA promoter construct with or without 0.5 μg/well MKL1 expression vector) using Fugene 6 (Roche, Indianapolis, IN). RTC were then plated in six-well plates at ∼50% confluence. After 6 h, cells were incubated with or without 10 ng/ml TGF-β1. Luciferase activity was determined after 2 days using a luciferase assay system (Promega) and detected with a multimode fluorescence/luminescence plate reader (Ultra 384-Tecan, Research Triangle Park, NC). Luciferase activity was normalized to protein content assessed using a protein assay reagent (Bio-Rad, Hercules, CA). Each experiment was performed in triplicate at least two times.
RNA interference (RNAi) analysis of human MKL1 was performed using small interfering RNA (siRNA) duplexes from Qiagen. The control siRNA sense sequence was UUCUCCGAACGUGUCACGUdTdT, and the siRNA MKL1 sense sequence was GUGUCUUGGUGUAGUGUAAdTdT. RTC were transfected with 3 μg siRNA by nucleofection using a Nucleofector basic kit with “primary mammalian epithelial cell solution” and program T-23 (Amaxa) and then plated at 60–80% confluence in six-well plates or 150-mm plates (for nuclear/cytoplasmic extraction) overnight. Then, cells were incubated for 3–5 days in the presence or absence of TGF-β1 for real-time PCR analysis and Western blotting. Similarly, RTC were transfected by nucleofection with 3 μg siRNA and 2 μg of the −259/+51 α-SMA promoter/luciferase vector, plated in six-well plates, and incubated in growth medium for 7 h. Then, cells were treated for 1 day in the presence or absence of TGF-β1, and promoter activity was determined as indicated above.
Quantitative PCR analysis.
RNA from RTC was quantified using real-time RT-PCR analysis as previously described (9). Expression of mRNA was calculated using the formula 2−(ΔΔCT), where ΔΔCT represents the cycle difference corrected for 18S or cyclophilin used as internal controls. Values were then related to the sample from cells transfected with the control siRNA and cultured in the absence of TGF-β1 (32).
MKL1/GFP fusion protein fluorescence microscopy.
RTC were transfected with 3 μg MKL1/GFP or GFP (pmaxGFP, Amaxa) expression vectors by nucleofection as described above. Cells were plated at ∼50% confluence in six-well plates or 150-mm plates (for nuclear/cytoplasmic extraction) for 6 h until attachment. Then cells were incubated in the presence or absence of TGF-β1, and MKL1/GFP and GFP expression was determined at the indicated times. Colocalization of MKL1/GFP expression and α-SMA promoter activity was determined by cotransfection of 2 μg of plasmid vectors encoding MKL1/GFP or GFP and the −259/+51 α-SMA promoter sequence driving the expression of RFP. Nuclei were then stained by incubating in a 10 μg/ml 4′-6-diamidino-2-phenylindole (Sigma) solution for 3 min. The cells were viewed with an IX50 inverted fluorescent microscope (Olympus, Melville, NY).
Expression of α-SMA, MKL1, E-cadherin, and β-actin was detected by Western blotting as previously described (9). Nuclear and cytoplasmic extraction was performed using NE-PER reagents according to the manufacturer's instructions (Pierce, Rockford, IL). RTC were transfected by nucleofection with Flag/MKL1 or MKL1/GFP expression plasmids as described above. p3xflag-CMV-7.1 empty vector, pmaxGFP or pEGF-N1 expression was used as negative controls. Anti-Flag M2 (Sigma) and HA.11 (Covance, Berkeley, CA) monoclonal antibodies were used at 1:2,000 dilution to detect Flag/MKL1 and MKL1/GFP, respectively. Goat polyclonal antiserum against MKL1 (Santa Cruz Biotechnology, Santa Cruz, CA) was used at 1:500 dilution.
Fluorescence microscopic analysis.
Cells were fixed with either ice-cold 100% methanol on ice for 5 min or with 4% (wt/vol) paraformaldehyde in PBS at room temperature for 10–20 min or at indicated times in plastic dishes and then stained and viewed as previously described (9, 10). Phalloidin conjugated to Alexa Fluor 555 (Invitrogen) was diluted 1:250. Rabbit polyclonal antibody against MKL1 (Santa Cruz Biotechnology) or rIgG was used at concentration of 1 μg/ml followed by incubation with donkey anti-rabbit IgG conjugated to Cy3 (Jackson, West Grove, PA.). Nuclei were then stained by incubating in a 10 μg/ml 4′-6-diamidino-2-phenylindole (Sigma) solution for 5 min. Fluorescence was also directly observed in live transfected cells expressing MKL1/GFP fusion protein, GFP, and RFP.
Chromatin immunoprecipitation assay.
Chromatin immunoprecipitation (CHIP) assay was performed using ChIP-IT kit from Active Motive (Carlsbad, CA) and analyzed using real-time PCR. Briefly, cells treated with or without TGF-β1 for 3 days were fixed with 1% formaldehyde and homogenized using a Dounce homogenizer. After centrifugation, the nuclear pellet was sheared by enzymatic digestion for 15 min to ∼200-bp DNA fragments and submitted to immunoprecipitation. We used rabbit polyclonal antibody against MKL1 and SRF (Santa Cruz Biotechnology), rabbit IgG (rIgG as negative control, Jackson), and mouse monoclonal RNA pol II antibody and mouse IgG (mIgG) that were provided with the kit as positive and negative controls, respectively. Antibodies and control IgGs were used at a concentration of 2 μg/reaction. Recovered and input DNA was quantified by real-time PCR to amplify the proximal CArG-containing region and upstream distal region (used as negative control) of the α-SMA promoter. The primers were designed using PrimerExpress Software (Applied Biosystems) as follows: −136/−58 α-SMA promoter sequence, 5′-AGTTTTGTGCTGAGGTCCCTATATG-3′ and 5′- TTCCCAAACAAGGAGCAAAGA-3′; −952/−862 α-SMA promoter sequence, 5′- CCCCCATCGATCCAGTCA-3′ and 5′- ACGTCCCCACCCAAATCTC-3′.
Data from three different experiments in Table 1 were analyzed using paired t-test with GraphPad software (San Diego, CA).
Functional analysis of the α-SMA promoter.
To determine the promoter regions that modulate human α-SMA gene expression in human RTC during EMT, various deletions and mutations of the α-SMA promoter were generated. Three different 5′-deletion constructs of the α-SMA promoter were transfected into primary cultures of human RTC. Figure 1 (top) demonstrates the basal and TGF-β1-stimulated activity of the −2580/+51-bp promoter sequence, along with the activity of the −893/+51- and −259/+51-bp deletion sequences. The promoter efficiency of these deletions indicate that removal of the regions spanning −2580 to −893 and −893 to 259 bp does not affect basal promoter activity substantially. Furthermore, TGF-β1 stimulates promoter activity of all three deletion constructs by two- to threefold. Thus the bulk of the promoter activity resides in the −259/+51 region of the α-SMA gene. Figure 1 (bottom) demonstrates that mutation of the CArG A or CArG B elements in the α-SMA −259/+51-bp promoter construct abolishes the TGF-β1 stimulatory effect. Therefore, CArG cis-acting elements are essential for TGF-β1-induced α-SMA promoter activity.
MKL1/2 expression induces α-SMA promoter activity.
Since MKL1 and MKL2 are known to mediate their effect through CArG elements in other systems, we assessed the possible role of MKL1 and MKL2 expression in regulating α-SMA transcription. Table 1 shows the activity of the −259/+51-bp human α-SMA promoter fragment when cotransfected with MKL1, MKL2, and empty expression vectors in three different experiments. MKL1 and MKL2 strongly transactivate the α-SMA promoter by an average of 149 ± 6.9- and 53 ± 23-fold, respectively. In RTC transfected with MKL1 or MKL2 and treated with TGF-β1, this effect tended to be higher, by an average of 695 ± 295- and 113 ± 47-fold respectively, but this TGF-β1 effect was not statistically significant. These results indicate that exogenous expression of MKL1 or MKL2 is sufficient to promote α-SMA transcriptional activity in RTC.
We next determined the relative importance of the two cis-acting CArG elements in MKL1-induced α-SMA promoter activity in the presence and absence of TGF-β1. Figure 2 (bottom left) demonstrates that single CArG A and CArG B mutations inhibit the MKL1-stimulated α-SMA promoter activity by ∼50 and 90%, respectively. Furthermore, double mutation of both the CArG A and CArG B elements leads to almost 100% inhibition of MKL1-induced activity. We also analyzed the activity of various 5′ α-SMA promoter deletion constructs, which may provide information about whether sequences adjacent to the CArG elements contribute to MKL1 activity. Figure 2 (bottom right) shows that MKL1 activity is closely correlated only with the presence of the CArG elements in the different constructs. In summary, our results indicate that both CArG A and CArG B elements are required for MKL1 activity.
Overexpressed MKL1 is located in the nucleus and drives α-SMA expression.
To determine the cellular localization of MKL1 in RTC, we generated a construct containing MKL1 fused to GFP. This construct (MKL1/GFP) was transfected into RTC by nucleofection. As a control, we also used a vector encoding GFP. Cells were incubated in growth medium containing serum for the initial 6 h after transfection to allow cell attachment; then, cells were incubated in serum-free medium. Figure 3 demonstrates that cells transfected with MKL1/GFP vector express MKL1/GFP in their nuclei even without TGF-β1 stimulation. In contrast, GFP not fused to MKL1 is expressed intensely in the cytoplasm. TGF-β1 treatment also does not change the localization of MKL1/GFP fusion protein in the nucleus, nor does it alter the cytoplasmic localization of GFP. While MKL1/GFP expression is mainly abolished after 2 days, GFP is sustained. The percentage of cells, depicted in Table 2 expressing MKL1/GFP in the nucleus was 97 and 90% 6 h and 1 day after transfection, respectively. Both nuclear and cytoplasmic localization is observed 2 days after transfection; however, the minute number of cells that express MKL1/GFP after 2 days did not allow accurate quantification. TGF-β1 treatment has no effect on the percentage of cells expressing MKL1/GFP in their nuclei after 1 or 2 days (data not shown). In addition to nuclear MKL1/GFP distribution, Fig. 3 also shows diffuse localization of MKL1/GFP in the cytoplasm. Magnification depicted in Fig. 4 demonstrates that expression of MKL1/GFP is nuclear but also perinuclear, with vesicles containing MKL/GFP around the nucleus.
As shown in Fig. 3, we also assessed how actin dynamics control nuclear localization of MKL1. Cells were stained for F-actin with phalloidin after transfection with MKL1/GFP or GFP. Little difference in F-actin staining is observed in cells transfected with MKL1/GFP compared with cells transfected with GFP, even after 2 days. However, as expected, TGF-β1 treatment induced a significant increase in F-actin by 2 days of treatment as previously shown in this cell type (59).
Therefore, MKL1/GFP localization to the nucleus in RTC occurs without a requirement for substantial TGF-β1 stimulated F-actin formation.
To directly demonstrate that cells transfected with MKL1/GFP are the same cells in which α-SMA expression is induced, we cotransfected vectors for MKL1/GFP or GFP with the α-SMA promoter linked to RFP. After 6-h transfection, MKL1/GFP is expressed (Fig. 3) but α-SMA promoter activity, highlighted by red fluorescence, is not yet expressed (results not shown). Figure 5 shows that RFP expression driven by the α-SMA promoter, highlighted by red fluorescence, coincides with nuclear MKL1/GFP expression in some cells after 1-day transfection. In cells expressing MKL1/GFP and/or RFP, coexpression was observed in 65–83% of cells; MKL1/GFP and RFP do not coexpress in 17–35 and 3–19% of cells, respectively. Cells cotransfected with the α-SMA promoter/RFP construct and GFP expression vector (used as a negative control) manifest no RFP expression, suggesting that MKL1/GFP is specifically responsible for the activation of the α-SMA promoter/RFP construct. The presence of red fluorescence in some cells without green fluorescence, especially after 2 days, most likely is a result of initial induction of the α-SMA promoter/RFP construct by MKL1/GFP, followed by a rapid loss of MKL1/GFP expression. Therefore, the expression of MKL1/GFP, which induces α-SMA promoter activity, represents a dynamic process.
Degradation by the ubiquitin pathway modulates MKL1 expression.
Quantification of MKL1/GFP and GFP expression in transfected cells is reported in Table 2. The number of nuclei per field is similar in cells transfected with MKL1/GFP and GFP, which indicates that cell loss is not induced by MKL1/GFP compared with GFP expression. While a similar percentage of cells (between 50 and 61%) expressed GFP 2 days after transfection, MKL1/GFP expression decreased from 30% 6 h after transfection to 4–6% after 1 day and 0.3–0.5% after 2 days, whereas GFP expression remained stable, indicating that MKL1/GFP has a relatively short half-life.
Determination of MKL1 expression by Western blotting in Fig. 6 demonstrates that MKL1 expression is downregulated 3 days after transfection compared with 1 day. While a band corresponding to the expected molecular weight of MKL1, ∼160 kDa (39), is observed (Fig. 6, arrow), numerous lower molecular bands also appeared, which suggests MKL1 degradation. Therefore, we determined the major protein degradation pathway for MKL1 by comparing the level of MKL1/GFP protein in RTC after overnight treatment (added 6 h after transfection) with E-64, a cysteine protease inhibitor; Z-LLF and MG132, inhibitors of the ubiquitin-proteasome system, or chloroquine, a lysosome inhibitor (23). Figure 6 (right) demonstrates that the two proteasome inhibitors sustained MKL1 expression, which indicates that the ubiquitin degradation pathway regulates MKL1 expression.
Inhibition of MKL1 expression by siRNA inhibits TGF-β1 stimulation of α-SMA.
To determine the effect of MKL1 on the regulation of α-SMA expression in RTC, we used the RNAi technique to reduce the level of endogenous MKL1. Expression of MKL1 and α-SMA mRNA was then assessed by real-time quantitative RT-PCR. Figure 7 (left) shows that MKL1 siRNA reduces the MKL1 mRNA expression level. The basal level of α-SMA expression is similar in the presence or absence of MKL1 siRNA. However, α-SMA mRNA expression stimulated by TGF-β1 treatment is significantly reduced in RTC transfected with MKL1 siRNA. We demonstrated previously that TGF-β1 treatment of RTC cells modulates the expression of reference genes (9); therefore, we validated our results by using cyclophilin A and 18S as two different reference genes that show similar results (results not shown). We further confirmed the effect of MKL1 siRNA by Western blot analysis of protein expression. Figure 7 (right) shows that MKL1 expression analyzed by Western blotting is inhibited in the presence of MKL1 siRNA. TGF-β1 stimulation of α-SMA expression does not occur when MKL1 expression is inhibited. In contrast, TGF-β1 abrogation of E-cadherin expression is not affected by inhibition of MKL1 expression. These results indicate that MKL1 expression is necessary for TGF-β1 stimulation of α-SMA expression.
MKL1 is constitutively localized in the nucleus.
To analyze localization of endogenous MKL1, we used MKL1 antibodies. Figure 7 demonstrates specific detection of endogenous MKL1 by Western blot analysis using lysates of RTC transfected with control and MKL1 siRNAs. Figure 8, top left, also shows the level of expression of MKL1 in lysates of RTC transfected with GFP (as a control) and MKL1/GFP. MKL1 antibody recognizes a band (arrow) of expected molecular weight that is much stronger in MKL1-transfected cells, which indicates a low level of endogenous MKL1 expression compared with that found in MKL1/GFP-transfected cells. Then, we analyzed nuclear and cytoplasmic extracts from RTC transfected with control and MKL1 siRNA. Analysis of histone H1 and β-actin was used for control of nuclear and overall protein loading, respectively. Figure 8 (top right) shows that endogenous MKL1 is located in both nuclear and cytoplasmic compartments. As expected, MKL1 expression is inhibited in both nuclear and cytoplasmic extracts of cells transfected with MKL1 siRNA. We also analyzed nuclear and cytoplasmic extracts from RTC transfected with GFP and MKL1/GFP. Figure 8 (bottom) shows expression of endogenous MKL1 and MKL1/GFP mainly in the nucleus. Although the relative ratio between the amount of MKL1 in cytoplasmic and nuclear extracts is variable in different experiments (Fig. 8 and Supplementary Fig. S1), our results indicate that MKL1 constitutively was located in the nucleus. TGF-β1 treatment did not have a major effect on MKL1 expression and MKL1 distribution in the nuclear and cytoplasmic compartments. Supplementary Fig. S1 also compared expression of (Flag)MKL1 and (HA)MKL1/GFP. The difference in molecular weights represents the difference in the size of tags and also probably partial degradation of transfected MKL1 as observed, as shown in the analysis of endogenous MKL1 and MKL1/GFP in Fig. 8. The higher expression of (HA)MKL1/GFP than (Flag)MKL1 is consistent with higher activation on α-SMA promoter activity (Supplementary Fig. S1, bottom), which demonstrates that GFP fusion to MKL1 protein does not affect MKL1 activity.
We next determined MKL1 localization by immunofluorescence. To determine the efficiency of the MKL1 antibody, RTC were transfected with GFP and MKL1/GFP, and two different fixatives were used, 4% paraformaldehyde at room temperature for 20 min and ice-cold methanol on ice for 5 min. Figure 9 reveals that GFP fluorescence disappears when cells are fixed in methanol but not paraformaldehyde. MKL1/GFP is seen in the nucleus of RTC fixed by each one of these fixatives. However, fixations by formaldehyde and methanol have very different outcomes on the detection of MKL1 by immunofluorescence. Surprisingly, strong staining of MKL1/GFP is detected in cytoplasm of RTC by MKL1 antiserum (red fluorescence) even though green fluorescence appears only in the nucleus. Moreover, MKL1/GFP staining in the nucleus is visible only when cells were fixed in methanol but not paraformaldehyde. In addition, endogenous MKL1 is stained in the cytoplasm and nucleus when cells are fixed by paraformaldehyde and methanol, respectively. rIgG used as a negative control shows no staining. These results indicate major differences of detection of MKL1 depending on the methodology used in fluorescence microcopy. To further determine the effect of paraformaldehyde fixation on nuclear localization of MKL1, RTC were fixed during different times before staining by immunofluorescence. Supplementary Fig. S2 reveals that nuclear localization of MKL1 is visible when cells were fixed for 1–5 min, but this staining is impaired when cells were fixed for longer times. Figure 10 shows immunofluorescence staining of RTC transfected with GFP and MKL1/GFP and treated with or without TGF-β1 and fixed in paraformaldehyde for 1 min. As expected, green fluorescence from MKL1/GFP is localized in the nucleus while green fluorescence from GFP was much less intense than in cells fixed with paraformaldehyde for longer times. Immunofluorescence reveals that MKL1/GFP and endogenous MKL1 are located in the nucleus and the cytoplasm in cells treated or not treated with TGF-β1. Altogether, our results indicate that MKL1 is constitutively located in both the cytoplasm and the nucleus in RTC.
TGF-β1 induced binding of SRF and MKL1 to the α-SMA promoter in vivo.
To further characterize endogenous MKL1 binding to the α-SMA promoter, we performed a CHIP assay. This assay utilizes immunoprecipitation to determine factors that bind to the α-SMA promoter within the context of native chromatin in RTC. Figure 11 (top) shows that TGF-β1 induced MKL1, SRF, and RNA pol II binding (used as a positive control) to the proximal region of the α-SMA promoter. Comparison of the level of α-SMA promoter DNA detected when specific antibodies and nonspecific rIgG and mouse IgG in the immunoprecipitation procedure were used demonstrates that α-SMA promoter DNA detection is not due to nonspecific binding of antibodies to chromatin. We also confirmed specificity of binding of these factors to the proximal region of the α-SMA promoter by analyzing binding of these factors in a more distal region of the α-SMA promoter (Fig. 11, bottom). Binding of specific and nonspecific antibodies was similar in this region of the α-SMA promoter, which further indicates specificity of binding in the α-SMA promoter proximal region containing the CArG boxes. Therefore, our data suggest that TGF-β1 induces SRF/MKL1 binding to CArG boxes in the α-SMA promoter.
Our study provides evidence for an important role of MKL1 in α-SMA expression in RTC. MKL1 overexpression is sufficient to induce α-SMA expression in RTC. We showed that MKL1 activates promoter activity in serum-free medium in the absence of TGF-β1. Consistent with these findings, transfected and endogenous MKL1 is expressed in its “active location,” the nucleus. Our findings also provide evidence for the first time that MKL1 expression is required for TGF-β1-induced α-SMA expression in RTC. Inhibition of MKL1 expression by siRNA inhibits TGF-β1's stimulating effect on α-SMA expression. Both TGF-β1- and MKL1-induced promoter activity is mediated through CArG response elements, which indicates that MKL1/SRF signaling is necessary for the TGF-β1 stimulatory effect. Our results also indicate that TGF-β1 stimulates SRF/MKL1 and RNA pol II binding to the proximal region of the α-SMA promoter, which suggest that TGF-β1 induces α-SMA transcriptional activity mediated through activation of MKL1/SRF and transcription factors of the basal transcription machinery, respectively. Although the specific mechanisms interconnecting TGF-β1 stimulation, MKL1 activity, and α-SMA transcriptional regulation need to be investigated in further detail, our study indicates that MKL1 is a key component in the regulation of TGF-β1-induced α-SMA expression in renal epithelial cells during EMT.
TGF-β1-induced α-SMA expression is mediated by CArG elements on the α-SMA promoter. MKL1 expression is probably involved in CArG activation since transfection of MKL1 siRNA inhibits the TGF-β1 effect. CArG motifs are known to confer serum and growth factor-induced transcriptional activation through the binding of SRF as well as other transcription factors and accessory proteins (18, 35). Myocardin, MKL1, and MKL2 associate with SRF to coactivate transcription of genes, such as α-SMA, SM-22, and SM-myosin heavy chain, that require promoter pairs of CArG boxes for activation. However, some promoters of genes, such as telokin and c-fos, contain only a single CArG box (65, 70–72). Cotransfection of MKL1 or MKL2 had no effect on smooth muscle gene markers that do not contain CArG boxes, such as smothelin-B, ACLP, and FRNK (70). Our study indicates that both CArG A and CArG B are necessary to mediate α-SMA promoter activity in RTC (Fig. 1). However, when MKL1 was cotransfected (Fig. 2), a single CArG box is capable of mediating α-SMA promoter activity independently. The mechanism by which MKL1 activates single or double CArG-containing promoter genes is not known. Wang et al. (65) proposed that myocardin formed SRF/myocardin complexes that bind to the two adjacent CArG boxes on promoters, which allows dimerization of myocardin and induction of the transactivation domain into an active conformation. This process requires stoichiometric amounts of SRF and MKL1. It is possible that cotransfection of MKL1 allows formation of MKL1 dimers that bind with low affinity even in the presence of only one functional CArG element. Zaromytidou et al. (72) recently suggested that interaction of MKL1 with SRF induced SRF-mediated DNA bending that facilitates MKL1-DNA contact in the c-fos promoter. The central AT-rich consensus region is present in the c-fos promoter but contains G or C substitutions in the α-SMA promoter that reduce SRF binding activity (18). Substitution of the CArG in the α-SMA promoter with the c-fos consensus sequence resulted in relaxed specificity in cultured cells but exhibited no effect on α-SMA expression during development and maturation in the vasculature of transgenic mice (20). However, this substitution has a profound effect on injury-induced downregulation of α-SMA expression under conditions inducing an increase in SRF and a decrease in myocardin expression (20). Taken together, these results indicate that MKL1 expression modulates the activity of CarG boxes in the α-SMA promoter.
A possible mechanism for complex binding is that myocardin and MKL regulate chromatin relaxation, which may permit SRF accessibility to CArG boxes by regulating histone acetylation and methylation (37, 70). Our results demonstrated that TGF-β1 induced binding of endogenous SRF and MKL1 to the α-SMA promoter in chromatin. Our results also indicate that RNA pol II binding, a transcription factor which interacts with a multisubunit complex that contains TATA-binding protein and initiates transcription on the α-SMA promoter, is also induced by TGF-β1 treatment. Therefore, binding of the MKL1/SRF complex is likely to be linked to activation of the basal transcription machinery mediated by the TATA box located at −27 on the α-SMA promoter, which is close to the CArG A and B boxes located at −70 and −120, respectively. TGF-β1 stimulates α-SMA expression through induction of SRF/MKL1 complex binding to CArG boxes. Whether MKL1 facilitates binding of SRF and other transcription factors from basal transcription machinery on the α-SMA promoter through an epigenetic component of chromatin (e.g., histone modification) needs to be clarified.
Our data indicate that the expression of MKL1 is rapidly downregulated in transfected RTC (Table 2). Consistent with this result, the half-life of MKL1 is between 2 and 4 h in transfected HeLa cells (41). We found that the ubiquitin proteolytic pathway regulates MKL1 degradation in RTC independently of TGF-β1 signaling. The effect of TGF-β1 can also be dampened by E3 ubiquitin ligase that specifically targets Smads for proteolytic destruction (6). Since TGF-β1 signaling is critical in a variety of biological processes and cell types, sensitivity to TGF-β1 may be achieved through ubiquitin-mediated degradation of different proteins involved in TGF-β1 signaling (13, 14, 67). Therefore, ubiquitin-mediated proteolysis may play an important role by regulating the level of MKL1 expression in RTC. α-SMA promoter activation only correlates with MKL1/GFP expression in some cells, which possibly reflects the rapid loss of the bulk of MKL1/GFP expressed. It is also probable that posttranslation modification regulates MKL1 activity. Preliminary results indicate that despite the effect on stabilizing MKL1 expression, inhibition of ubiquitination by MG132 impedes MKL1-induced α-SMA promoter activity (results not shown). Consistent with these results, ubiquitin-dependent mechanisms induce heterochromatin relaxation, facilitate assembly of transcription complexes on promoters, and enhance transcriptional activation capacity of transcription factors but also target transcriptional coactivators and corepressors for degradation (7, 22, 33).
Myocardin is often described as a master regulator of expression of smooth muscle genes that contain two or more essential binding sites for SRF in their promoter regions and conveys smooth muscle-specific differentiation (63–65). Our results demonstrate that inhibition of MKL1 expression by siRNA inhibits TGF-β1-induced α-SMA expression but does not affect TGF-β1's effect on E-cadherin downregulation. In addition, MKL1 overexpression leading to α-SMA expression is not associated with stress fiber formation. Therefore, MLK1 expression in RTC is likely to be specifically involved in α-SMA and other muscle-related gene expressions that are stimulated during EMT. Our findings apparently differ from a recent study which suggested that MKL1 and MKL2 compensate for each other and affect a wide spectrum of the EMT process through a mechanism that involves slug induction and actin remodeling in Madin-Darby canine kidney (MDCK), human kidney-2 (HK-2) proximal tubular, and mouse mammary gland (NMuMG) epithelial cell lines (40). However, that study also revealed variations in the effect of MKL1 expression on Snail and Twist transcription regulators and epithelial/mesenchymal markers such as vimentin and β-catenin in the different cell lines. Therefore, the extent of the effect on EMT triggered by MKL1 may be different in the various epithelial cell lines and primary human RTC. We recently reported phenotypic variability in these cells, especially when they were cultured in a three-dimensional matrix (10) Various transcription regulators are thought to have a wide effect on EMT. Snail/Slug family and Twist transcription factors have been described as direct repressors of E-cadherin and inducers of EMT and invasion when overexpressed in epithelial cells (3, 68). Notably, Snail may have an important role in TGF-β1-induced EMT in the MDCK cell line (27, 46). Recently, Venkov et al. (62) described formation of CArG box binding factor-A and KRAB-associated protein 1 complex binding to fibroblast transcription site-1 that is distinct from CArG box binding elements and is possibly involved in the expression of multiple gene EMT markers in renal epithelial cells. Therefore, TGF-β1 induced a process of EMT that is integrated in coordinated events, such as stimulation of α-SMA expression, E-cadherin downregulation, and stress fiber formation in RTC. The precise spectrum of changes and mechanisms of regulation occurring during EMT is still unclear and involves multiple signaling networks, which have been extensively reviewed (26, 31, 58).
MKL1 nuclear translocation does not necessarily account for its activity and constitutive MKL1 nuclear localization has been observed in some cell types (8, 21, 48, 70). Our study indicates that methodology leads to different outcomes regarding MKL1 localization. MKL1 expression has been recently studied in LLC-PK1 cells (porcine renal proximal tubular epithelial cells) (12). Similar to the results of our study using RTC, transfected MKL1 is located in the nucleus in the absence of TGF-β1 stimulation in LLC-PK1 cells, whereas endogenous MKL1 is mainly cytoplasmic and TGF-β1 and cell contact inhibition induce MKL1 translocation to the nucleus via activation of the Rho pathway (12). TGF-β1 also induces MKL1 accumulation into the nucleus of MDCK cells (40). The original model proposed by Treisman's group (39, 48) suggested that MKL1 binds to monomeric G-actin in the cytoplasm through the NH2-terminal RPEL domain. Stimulation of Rho activity by serum induces G-actin into F-actin and subsequently releases MKL1, which translocates into the nucleus, where it activates transcription (39). This model has been recently revised, and MKL1 is now considered to be constitutively targeted to the nucleus. In this model, serum-induced Rho signaling and MKL1 interaction with actin regulate MKL1 nuclear export; MKL1 binds to SRF in the nucleus but remains inactive unless its binding to nuclear actin is disrupted (61). Our results clearly indicate that endogenous MKL1 is located in both the nucleus and cytoplasm in RTC (Figs. 8–10). Possibly, detection of MKL1 in the nucleus is dependent on interaction of MKL1 and actin. Interestingly, methanol fixation impairs F-actin detection by phalloidin (43) and allows endogenous MKL1 detection solely in the nucleus while overexpressed MKL1 is still detectable in the cytoplasm, which further suggests that detection of MKL1 in the cytoplasm is limited by its ratio to intracellular components, possibly G/F-actin, that may be affected during the fixation process.
While both serum and TGF-β1 stimulation is thought to activate the Rho pathway and MKL1 nuclear accumulation, α-SMA expression is induced by TGF-β1 but not serum, which indicates that additional signaling pathways triggered by TGF-β1 are necessary for α-SMA expression. Smad2 and Smad3 inhibition by siRNA indicates their essential role in TGF-β1-induced α-SMA expression in HKC-8 (proximal tubular cell line from human kidney) (47), and we obtained similar results in RTC (unpublished observations). Smad signaling activation triggered by TGF-β can possibly be integrated with the SRF pathway to either mediate or inhibit TGF-β1-induced gene transcription (29, 49, 50). Smad3 has been shown to interact with MKL1 whereas the MKL1/Smad3 complex activates slug expression (40). Our study indicates that TGF-β1 induces MKL1 and SRF binding to CArG response elements in the α-SMA promoter (Fig. 11). Therefore, our results imply that TGF-β1 induced MKL1 activity through induction of MKL1/SRF complex binding to the α-SMA promoter responsive elements. Overexpression of MKL1 is sufficient to induce α-SMA expression in RTC incubated in serum-free medium in the absence of TGF-β1 treatment (Figs. 2 and 5). The high levels of overexpressed MKL1 compared with the endogenous level found in RTC may induce activity that does not occur at physiological levels or may overwhelm particular cellular constituents present in limited amounts. A similar example of this effect is Smad3; when Smad 3 is overexpressed, the collagen Iα2 promoter is also induced even in the absence of TGF-β1 treatment (52). In conclusion, future identification of the mechanisms by which MKL1 regulates TGF-β1 signaling in RTC may help delineate cell-specific processes responsible for the formation of renal myofibroblasts.
We gratefully acknowledge the financial support of the Children's Medical Research Institute and Children's Medical Network.
We thank Dr. R. Prywes for kindly providing MKL1 and MKL2 expression vectors.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society