Effect of destrin mutations on the gene expression profile in vivo

Angela M. Verdoni, Natsuyo Aoyama, Akihiro Ikeda, Sakae Ikeda

Abstract

Remodeling of the actin cytoskeleton through actin dynamics (assembly and disassembly of filamentous actin) is known to be essential for numerous basic biological processes. In addition, recent studies have provided evidence that actin dynamics participate in the control of gene expression. A spontaneous mouse mutant, corneal disease 1 (corn1), is deficient for a regulator of actin dynamics, destrin (DSTN, also known as ADF), which causes epithelial hyperproliferation and neovascularization in the cornea. Dstncorn1 mice exhibit an actin dynamics defect in the corneal epithelial cells, offering an in vivo model to investigate cellular mechanisms affected by the Dstn mutation and resultant actin dynamics abnormalities. To examine the effect of the Dstncorn1 mutation on the gene expression profile, we performed a microarray analysis using the cornea from Dstncorn1 and wild-type mice. A dramatic alteration of the gene expression profile was observed in the Dstncorn1 cornea, with 1,226 annotated genes differentially expressed. Functional annotation of these genes revealed that the most significantly enriched functional categories are associated with actin and/or cytoskeleton. Among genes that belong to these categories, a considerable number of serum response factor target genes were found, indicating the possible existence of an actin-SRF pathway of transcriptional regulation in vivo. A comparative study using an allelic mutant strain with milder corneal phenotypes suggested that the level of filamentous actin may correlate with the level of gene expression changes. Our study shows that Dstn mutations and resultant actin dynamics abnormalities have a strong impact on the gene expression profile in vivo.

  • corn1
  • microarray
  • cornea
  • serum response factor

actin is one of the most abundant proteins in eukaryotic cells and is best known for its role as a cytoskeletal component. Within cells, actin exists either in a monomeric form (globular actin, or G-actin) or in a filamentous form (F-actin). The process by which G-actin is assembled into F-actin and F-actin is depolymerized into G-actin, which is referred to as “actin dynamics,” is tightly regulated and responds to extracellular signals. Remodeling of the actin cytoskeleton through actin dynamics is essential for numerous basic biological processes, including cell polarization, contractile force generation, cell migration, cell division, endocytosis, and exocytosis (5, 7, 39). In addition to its structural role, the status of the actin cytoskeleton has been shown to influence a diverse array of intracellular signaling processes. For example, the disruption of proper actin dynamics has been shown to result in the altered expression of cell surface molecules (34, 42), loss of cyclin D expression (10), and changes in histone acetylation (25). Moreover, recent research has provided convincing evidence that cytoskeletal actin dynamics also plays a direct role in the control of gene expression (33). Although the role of actin dynamics in cell signaling and transcriptional regulation has been investigated in vitro using the cell culture system, it has not been thoroughly explored in vivo.

It has been shown that members of the actin depolymerizing factor (ADF)/cofilin family of proteins are the primary regulators of actin dynamics in vivo (11, 23, 41). These proteins, which include destrin (DSTN; also known as ADF), cofilin 1 (CFL1), and cofilin 2 (CFL2) in mammals, enhance the depolymerization of F-actin into its monomeric form and promote filament severing (11, 28). DSTN shows the strongest depolymerization activity out of all family members and is expressed in epithelial and endothelial cells of multiple tissues (55, 57). CFL1 is ubiquitously expressed, while CFL2 is a muscle-specific isoform with the weakest depolymerization activity (55). Animal models with functional mutations in the ADF/cofilin genes have provided opportunities to examine the effect of actin cytoskeletal abnormalities in vivo. Mice homozygous for a targeted null allele of Cfl1 display an embryonic lethal phenotype with defective neural crest cell migration and a lack of neural tube closure (17). Neuronal cell specific targeting of Cfl1 further revealed its function in neuronal migration and cell cycle control in the cerebral cortex (8). Corneal disease 1 (corn1) mice are homozygous for a spontaneous null allele of the Dstn gene and show corneal abnormalities, including epithelial cell hyperproliferation and neovascularization in the stroma (19, 46). An allelic missense mutation in Dstn, Dstncorn1-2J, which causes milder corneal epithelial hyperproliferation without corneal neovascularization was also discovered (19). Although DSTN is expressed in many tissues, the phenotypes in Dstn mutant mice appear to be restricted to the cornea, where the main ADF/cofilin molecule expressed is DSTN (19). In other tissues, the loss of DSTN is likely compensated by other ADF/cofilin family members (19). The cornea of Dstn mutant mice, therefore, presents a unique model in which the effect of the loss of DSTN can be investigated.

In this study, we first examined the state of actin in the cornea of Dstn mutant mice and show that actin dynamics are affected in these mice. Then, we sought to determine the possible cellular mechanisms affected by Dstn mutations and resultant actin dynamics abnormalities in vivo through the identification of differentially expressed genes. We also compared the effects of allelic Dstn mutations on the gene expression profile. Allelic mutants show different levels of F-actin accumulation and gene expression changes, indicating that the level of F-actin may correlate with the level of change in gene expression patterns.

MATERIALS AND METHODS

Mice.

A.BY H2bc H2-T18f/SnJ (A.BY wild type), A.BY H2bc H2-T18f/SnJ-Dstncorn1/J (Dstncorn1), C57BL/6J (B6 wild type), and C57BL/6JSmn-Dstncorn1-2J/J (Dstncorn1-2J) mice were obtained from The Jackson Laboratory (Bar Harbor, ME) and bred in the animal facility at the University of Wisconsin-Madison. All mouse procedures were performed in accordance with the protocols approved by the Animal Care and Use Committee at the University of Wisconsin-Madison and conform to the ARVO statement for the use of animals in Ophthalmic and Vision Research and APS's Guiding Principals in the Care and Use of Animals.

Analysis of F/G-actin ratio.

The ratio of F-actin and G-actin in the cornea was analyzed using an F-actin/G-actin in vivo assay kit (Cytoskeleton, Denver, CO) based on the manufacturer's protocol. Briefly, pooled corneas from five mice of each genotype were homogenized in 500 μl F-actin stabilization buffer [50 mM PIPES pH 6.9, 50 mM KCL, 5 mM MgCl2, 5 mM EGTA, 5% (vol/vol) glycerol, 0.1% Nonidet P40, 0.1% Triton X-100, 0.1% Tween 20, 0.1% 2-mercaptoethanol, 0.001% AntifoamC, 1 mM ATP, 0.004 mM tosyl arginine methyl ester, 0.015 mM leupeptin, 0.01 mM pepstatin A, and 0.01 M benzamidine]. To pellet unbroken cells, lysates were centrifuged at 2,000 rpm for 5 min at 37°C, and supernatants were collected. To separate the F-actin from the G-actin pool, supernatants were centrifuged at 100,000 g for 1 h at 37°C. Supernatants were immediately collected, while pellets were resuspended in ice-cold molecular grade H2O plus 1 μM cytochalasin D and incubated on ice for 1 h to dissociate F-actin. The resuspended pellets were gently mixed every 15 min. To measure F/G-actin ratio, equal amounts of both the supernatant (G-actin) and the resuspended pellet (F-actin) for each genotype were subjected to immunoblot analysis with the use of an actin antibody (Cytoskeleton). Fractionation was performed for three separate groups for each genotype. The F/G-actin ratio was determined by scanning densitometry using ImageJ software (http://rsb.info.nih.qjgov/ij). Two-tailed, unpaired t-tests were performed comparing mutant and wild-type ratio values.

RNA isolation from cornea.

Total RNA was isolated from corneas with TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA) and further purified with an RNeasy kit (Qiagen, Valencia, CA). For each biological replicate, 20 corneas (10 mice) were pooled for microarray analysis and 2 corneas (1 mouse) were used for quantitative real-time PCR (qPCR). Total RNA quality was assessed on the basis of the A260/A280 ratio, the integrity of 18S and 28S rRNA, and chromatograms from the Agilent 2100 Bioanalyzer RNA Nanochip. Samples that had A260/A280 ratios between 1.9–2.1, gel profiles exhibiting distinct bands for 18S and 28S rRNA, and no obvious degradation as observed by Bioanalyzer were used for microarray and qPCR analyses.

Microarray analysis.

Microarray experiments were designed to comply with minimum information about microarray experiment (MIAME) guidelines. Microarray analysis for each genotype was performed with three individually derived RNA samples, each being hybridized to one Affymetrix MG 430 2.0 array (Affymetrix, Santa Clara, CA) according to the manufacturer's instructions (1). Briefly, 5.5 μg of total RNA sample was used to synthesize double stranded cDNA with the Super Script II double-stranded cDNA synthesis kit (Invitrogen Life Technologies) according to the manufacturer's instructions. Labeled cRNA was synthesized using the Affymetrix IVT amplification and labeling kit (Affymetrix). We hybridized 10 μg of purified, fragmented labeled cRNA with an array at 45°C for 16 h in an Affymetrix 640 hybridization oven. The posthybridization process was performed in an Affymetrix 450 fluidic station according to the manufacturer's instructions. All gene chips were scanned on an Affymetrix GC3000 G7 scanner, and data were extracted from scanned images using AFX GCOS 1.4 software. All technical microarray procedures were carried out by the Gene Expression Center at the University of Wisconsin-Madison Biotechnology Center (UWBC).

Microarray data analysis.

Background intensities were adjusted and normalized using GCRMA (version 2.0.0) (56) implemented in R (version 2.4.1). After the normalization process, only those probe sets for which at least one gene chip for each genotype had a “P” (present) call according to the Affymetrix detection algorithm and had sequences unique to single genes were included for further analysis (2). Two-tailed, unpaired t-tests were performed comparing mutant (n = 3) and wild-type (n = 3) log2 expression values. The probe sets with a P value of <0.01 were considered differentially expressed. Functional annotation was performed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) Functional Annotation Tool (http://david.abcc.ncifcrf.gov/) (13) with all probe sets for the mouse 430 2.0 chip as the background. This tool allowed us to identify Gene Ontology (GO) terms (4) that are associated with differentially expressed genes, measure overrepresentation of GO terms in differentially expressed gene lists using a modified Fisher exact test, and calculate the P value. The P value is referred to as the Expression Analysis Systematic Explorer (EASE) score (18). Euclidean distance complete linkage clustering was performed for probe sets commonly up- and downregulated between Dstncorn1 and Dstncorn1-2J using Genesis (version 1.7.2) (14, 50). The complete data set generated from this study is available from the National Center for Biotechnology Information Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/) with GEO accession number GSE9743.

qPCR analysis.

Total RNA was extracted as described above. For each biological replicate, one animal (two corneas) was analyzed. We converted 20 ng of total RNA to cDNA using Superscript III first-strand synthesis system (Invitrogen Inc, Carlsbad, CA) according to the manufacturer's instructions. qPCR reactions (20 μl) contained 2 μl of template cDNA, 10 μM of each primer, 1× Takara Premix Ex Taq (Takara Bio, Madison, WI), 1× ROX reference dye, and 5% dimethyl sulfoxide. Amplification was performed using an ABI Prism 7300 Real-Time PCR system (Applied Biosystems, Foster City, CA). Each cDNA sample was assayed in triplicate. The primer sequences used are listed in Supplemental Table S1.1 Thermal cycle conditions consisted of an initial denaturation step at 95°C for 10 s, followed by 40 cycles of a denaturing step at 95°C for 5 s and an annealing/extension step at 60°C for 45 s. Fluorescence was measured at the beginning of each annealing/extension step. To check the specificity of each primer pair, the predicted PCR amplicon melting temperature was confirmed by a dissociation curve analysis. PCR products were denatured at 95°C for 60 s and gradually reannealed while the temperature was adjusted to 60°C. The fluorescence was measured at every 0.5°C interval. Amplicon size was determined by electrophoresis on an agarose gel [1.5% (wt/vol)]. All PCR amplicons were sequenced at the UWBC, and the correct products were confirmed for all the primer pairs.

qPCR data analysis.

DART-PCR software (version 1.0) (http://www.gene-quantification.de/DART_PCR_version_1.0.xls) (37) was used to compute the amplification efficiency for each primer set and to compute the relative expression value. The amplification efficiency of each primer set is listed in Supplemental Table S2. Relative expression values for genes of interest were normalized to the geometric mean of the relative expression values of the two most stable reference genes (Fmr and Ubc). Two-tailed, unpaired t-tests were performed comparing mutant and wild-type relative expression values. Reference genes were selected by the following procedure. Five genes stably expressed in both mutant and wild-type mice were selected as candidate reference genes based on our microarray data (Cat, Fmr, Rgs10, Gapdh, and Ubc). The relative expression values of these genes were analyzed by geNorm (version 3.4) (53). After the analysis, a gene with the highest M value (a stability measure) was eliminated. This procedure was repeated until elimination of the gene with the highest M value had no significant effect on the newly calculated normalization factor.

Immunohistochemistry.

For immunohistochemistry on frozen sections, mice at postnatal day (P) 14 of age were killed by decapitation and the eyes were immediately removed and immersion fixed in 4% paraformaldehyde (PFA) for 2 h at 4°C, then cryoprotected at 4°C in a graded series of sucrose. Eyes were embedded in optimal temperature cutting compound and sectioned at 12 μm thickness. Sections were blocked in phosphate-buffered saline (PBS) with 0.5% Triton X-100 and 2% normal donkey serum for 20 min at room temperature. Sections were then incubated at 4°C overnight with primary antibody against serum response factor (SRF, 1:400; Santa Cruz Biotechnology, Santa Cruz, CA). Sections were rinsed in PBS and incubated with Alexa Fluor 488 conjugated secondary antibody (1:400, Invitrogen) and Alexa Fluor 568 conjugated phalloidin (1:50; Invitrogen) for 45 min at room temperature. Slides were counterstained with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI, 1:200; Sigma-Aldrich, St. Louis, MO).

For whole mount immunohistochemistry, mice at P14 of age were killed by decapitation, the eyes were immediately removed and placed in PBS, and the corneas were isolated. Corneas were fixed overnight in 4% PFA at 4°C. The next day, corneas were rinsed twice in PBS for 10 min, postfixed in 100% acetone for 20 min, and rinsed twice in PBS for 10 min. Corneas were blocked overnight in PBS with 0.8% Triton X-100 and 2% normal donkey serum, transferred to block solution with beta actin (1:50; Cell Signaling Technology, Danvers, MA) or MAL (MRTF-A, 1:250; Santa Cruz Biotechnology, Santa Cruz, CA) antibodies, and incubated overnight. Corneas were rinsed in PBS six times at 30 min each. They were transferred to block solution containing phalloidin (1:50, Invitrogen) and an Alexa Fluor 488 conjugated secondary antibody (1:400, Invitrogen) and incubated overnight. The following day, corneas were rinsed in PBS six times at 30 min each and counterstained with DAPI (1:200, Sigma-Aldrich). All steps following fixation in 4% PFA were performed at room temperature.

Images were captured on a Zeiss 510 Confocal Laser Scanning System and Axio Imager Microscope using LSM 510 Software (release 4.2) (Carl Zeiss MicroImaging, Thornwood, NY). Immunohistochemistry was performed on Dstncorn1 and Dstncorn1-2J mice, as well as on wild-type controls with matched genetic background (A.BY and B6, respectively).

RESULTS

Perturbation of actin dynamics in Dstn mutant mice.

We previously reported the increase of phalloidin signals in the corneal epithelium of Dstncorn1 and Dstncorn1-2J mice (19). Here, we further tested whether this increase in phalloidin signals reflects the increase of F-actin. We first performed double labeling of the whole mount cornea from mice at P14 of age using phalloidin and an antibody for beta actin. Consistent with our previous data, phalloidin produced increased signals in the corneal epithelial cells of Dstncorn1 and Dstncorn1-2J mice (Fig. 1). The signal was much higher in Dstncorn1 mice compared with Dstncorn1-2J mice (Fig. 1). In both mutants, we observed overlapping of phalloidin and beta actin signals (Fig. 1), suggesting that the increase of phalloidin signals in Dstn mutants indeed reflects the increase of F-actin. We further examined actin dynamics in these mutants by a biochemical analysis. Corneal lysates from Dstn mutant and wild-type control mice at P14 of age were subjected to ultracentrifugation, and fractionated cell extracts containing free G-actin and F-actin were analyzed using Western blotting. As shown in Fig. 2, there is a highly significant difference in the F/G-actin ratio between the Dstncorn1 mutant and A.BY wild-type control (0.796 ± 0.032 vs. 0.271 ± 0.018, respectively; P = 0.0001) and a significant difference between the Dstncorn1-2J mutant and B6 wild-type control (0.422 ± 0.096 vs. 0.117 ± 0.006, respectively; P = 0.0333). These results are consistent with our data for the phalloidin/beta actin staining in Fig. 1. It is also to note that the F-actin level in the A.BY wild-type cornea is considerably higher than that in the B6 wild-type cornea, as shown by both phalloidin staining (Fig. 1) and the fractionation experiment (Fig. 2). Our results demonstrate that Dstn mutations cause actin dynamics abnormalities in the corneal epithelial cells in the physiological setting and highlight the functional importance of DSTN in vivo.

Fig. 1.

Immunohistochemical staining for beta actin and phalloidin in the corneal epithelial cells of wild-type (WT) and destrin (Dstn) mutant mice. Single slice confocal images of beta actin (green) and phalloidin (red) demonstrate the colocalization of beta actin with F-actin only in Dstn mutant cornea, albeit to a significantly greater level in Dstncorn1 mice. Corneas were counterstained with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI, blue). The scale bar represents 20 μm.

Fig. 2.

Analysis of actin polymerization in the cornea of WT and Dstn mutant mice. The F/G-actin ratio is significantly higher in Dstn mutant mice compared with WT controls, demonstrating an increase in the amount of polymerized actin in these mice (top). Error bars represent SE. Statistical significance by t-test: *P < 0.05, ***P < 0.001. A representative immunoblot image of relative F-actin and G-actin levels in WT and Dstn mutant mice is shown (bottom).

Identification of genes differentially expressed in the cornea of Dstn mutant mice.

To examine the effect of the Dstn mutation on the gene expression profile, we first performed a microarray analysis using Dstncorn1 and A.BY wild-type cornea isolated from mice at P14. After normalization of the data, our analysis identified 1,226 annotated genes that are differentially expressed (P < 0.01) out of >14,500 genes represented on the array. Of these genes, 599 genes were upregulated and 627 genes were downregulated in the Dstncorn1 cornea.

To compare the effect of allelic Dstn mutations with different severity (null in Dstncorn1, missense in Dstncorn1-2J), we also performed a microarray analysis of Dstncorn1-2J and B6 wild-type cornea at P14. We identified 202 annotated genes differentially expressed in Dstncorn1-2J mice compared with B6 wild-type mice (P < 0.01), out of which 150 genes are upregulated and 52 genes are downregulated. As shown in the Venn diagrams in Fig. 3, a significant portion of these 202 genes (60 upregulated and 21 downregulated genes) is also differentially expressed in the Dstncorn1 cornea.

Fig. 3.

Venn diagrams showing the number of genes significantly up- or downregulated (P < 0.01) in the cornea of Dstncorn1 and Dstncorn1-2J mice. Sixty genes were upregulated in both Dstncorn1 and Dstncorn1-2J mice, while 21 genes were downregulated in both mutants.

Clustering analysis of differentially expressed genes.

The overlapping but less severe effect of the Dstncorn1-2J mutation could be further visualized when hierarchical clustering analysis was performed using the expression data for 81 genes (83 probe sets) identified as differentially expressed (upregulated or downregulated) in both Dstncorn1 and Dstncorn1-2J mice. In Fig. 4, the dendrogram showing clustering of the arrays (Fig. 4, top) indicates that arrays were successfully grouped by the genotype. Clustering of the probe sets based on the similarity of their expression patterns (dendrogram shown on left) yielded five distinct clusters (A–E). A box plot representation of log2 fold changes between mutant and wild-type cornea was calculated for each cluster (Fig. 4, right). Cluster A includes genes that are highly upregulated in both Dstncorn1 and Dstncorn1-2J cornea. The box plot shows that upregulation of these genes is more pronounced in the Dstncorn1 cornea. As observed in the heat map, most genes in clusters B and C are also upregulated to a greater extent in Dstncorn1 compared with Dstncorn1-2J mice. We also observed that most genes in clusters D and E are downregulated to a greater extent in Dstncorn1 compared with Dstncorn1-2J mice. In summary, for the majority of genes that are commonly disregulated in both Dstn mutants, the level of change (upregulation or downregulation) tends to be greater in Dstncorn1 mice compared with Dstncorn1-2J mice.

Fig. 4.

Hierarchical clustering of the 83 probe sets (81 genes) that are significantly up- or downregulated (P < 0.01) in both Dstncorn1 and Dstncorn1-2J mice. All of the probe sets are grouped into 5 distinct clusters (labeled A–E) based on their expression patterns. Each row corresponds to a single probe set, and each column to a single array. Affymetrix Probe IDs and gene symbols are indicated at right. Relative gene expression signal levels with WT as the baseline are indicated by color as shown in a scale at the top (green represents decreased expression, and red represents increased expression). Box plots at right show the average log2 fold changes (y-axis) for probe sets in the corresponding cluster using WT as the baseline for each mutant. The dendrogram at the top of the figure shows clustering of the experiments (arrays).

Identification of significantly affected functional gene classes.

In an attempt to determine which functional categories of genes are significantly affected by the Dstncorn1 and Dstncorn1-2J mutations, we performed functional annotation of the differentially expressed genes using the DAVID Functional Annotation Tool (http://david.abcc.ncifcrf.gov/) (13). Through this analysis, 67 GO terms were found to be significantly (EASE score < 0.01) overrepresented (enriched) (Table 1) in the list of genes upregulated in the Dstncorn1 cornea, which included 599 genes. Among these, one of the most significantly overrepresented terms was “cytoskeleton,” which included 55 genes (EASE score = 2.4 × 10−7, Table 2). Six other terms (“actin cytoskeleton,” “cytoskeletal protein binding,” “actin binding,” “cytoskeletal organization and biogenesis,” “actin filament-based process,” and “actin cytoskeletal organization and biogenesis”) are also associated with actin and/or cytoskeleton (Table 1). Consistent with the epithelial hyperproliferation phenotype caused by the Dstncorn1 mutation, cell cycle-associated terms (“cell cycle,” “regulation of progression through cell cycle,” “regulation of cell cycle,” “cell cycle regulator,” “mitotic cell cycle,” “mitosis,” “M phase of mitotic cell cycle,” and “cell division”) are also overrepresented. We also observed that inflammation-associated terms (“chemokine receptor binding,” “chemokine activity,” “response to biotic stimulus,” and “myeloid cell differentiation”) are overrepresented.

View this table:
Table 1.

List of GO terms overrepresented (EASE score < 0.01) in the upregulated gene list for Dstncorn1mice

View this table:
Table 2.

Genes that are associated with the GO term “cytoskeleton” in the upregulated gene list for Dstncorn1 mice

The same analysis performed on the microarray data from Dstncorn1-2J and B6 wild-type cornea yielded 20 GO terms overrepresented in the upregulated gene list, which included 150 genes (Table 3). Interestingly, the most significantly overrepresented term was “cytoskeleton,” which included 22 genes for this mutant (EASE score = 4.3 × 10−6, Table 4). Other GO terms associated with actin and/or cytoskeleton were also identified (Table 3). In contrast, GO terms associated with cell cycle progression and inflammation were not identified as overrepresented in the upregulated gene list for the Dstncorn1-2J cornea. These data suggest the existence of a threshold for the induction of gene expression due to actin perturbation. The increased expression of genes associated with actin and/or cytoskeleton is observed in both Dstn mutants, which shows that genes falling into these categories may be upregulated due to actin perturbation occurring even at mild levels.

View this table:
Table 3.

List of GO terms overrepresented (EASE score < 0.01) in the upregulated gene list for Dstncorn1-2J mice

View this table:
Table 4.

Genes that are associated with the GO term “cytoskeleton” in the upregulated gene list for Dstncorn1-2J mice

Functional annotation of the downregulated gene list for Dstncorn1 mice (627 genes) identified 10 overrepresented GO terms (Table 5) including “GTPase regulator activity,” “guanyl-nucleotide exchange factor activity,” “oxidoreductase activity,” and “alcohol dehydrogenase activity”. Just two were identified from analysis of the list of genes downregulated in Dstncorn1-2J (52 genes), and these were “small GTPase-mediated signal transduction” and “guanyl-nucleotide exchange factor activity” (Table 6).

View this table:
Table 5.

List of GO terms overrepresented (EASE score < 0.01) in the downregulated gene list for Dstncorn1 mice

View this table:
Table 6.

List of GO terms overrepresented (EASE score < 0.01) in the downregulated gene list for Dstncorn1-2J mice

Expression of SRF, its target genes, and coactivator MAL.

Out of 55 genes identified as upregulated in the cornea of Dstncorn1 mice that are associated with the GO term “cytoskeleton” (Table 2), 24 of them are known downstream targets of a transcription factor, SRF (6, 29, 38, 45, 52, 58). This phenomenon is of particular interest since it has been shown that the activity of SRF is influenced by actin dynamics in cultured cells (47). Moreover, Srf itself is one of the genes included in the list of differentially expressed genes in Dstncorn1 mice. By real-time PCR using the cDNA of P14 Dstncorn1 and A.BY wild-type cornea, we verified that the expression of Srf is significantly upregulated at the mRNA level (Fig. 5A). To examine the spatial expression pattern of the SRF protein in the Dstncorn1 cornea, we performed immunohistochemistry. The SRF signal was detected in the nuclei of corneal epithelial cells in Dstncorn1, but not in A.BY wild-type mice, at P14 (Fig. 5B), demonstrating the presence of SRF at the site of its potential activity in the Dstncorn1 cornea. SRF expression was observed primarily in the superficial layers of the corneal epithelium, where strong accumulation of F-actin is also observed (Fig. 5B).

Fig. 5.

Expression of serum response factor (SRF), its target genes, and coactivator MAL in the Dstncorn1 cornea. A: quantitative real-time PCR (qPCR) analysis demonstrates a significantly higher mRNA level for Srf in the Dstncorn1 cornea compared with A.BY WT. RE, relative expression. B: immunohistochemistry shows the nuclear accumulation of SRF in Dstncorn1 but not in WT cornea. SRF expression (green) is detectable in the nuclei (DAPI, blue) of the superficial layers in mutant cornea, where F-actin (phalloidin, red) accumulation occurs. The scale bar represents 20 μm. C: qPCR analysis of SRF target genes confirms their upregulation in the Dstncorn1 cornea. Relative expression values for Tpm2 are not observable for WT and mutant due to the scale used but are 0.0317 ± 0.0024 and 0.0641 ± 0.0098, respectively. Error bars represent SE. Statistical significance by t-test: *P < 0.05, **P < 0.01, ***P < 0.001. D: immunohistochemical staining for MAL and phalloidin in the corneal epithelial cells of WT and Dstncorn1 mice. Single slice confocal images of MAL (green), phalloidin (red), and DAPI (blue) demonstrate an increase in the MAL signal in the cytoplasm and nucleus of Dstncorn1 mice compared with WT. The scale bar represents 20 μm.

We expanded our analysis and examined the expression data for ∼400 known SRF target genes [Refs. 6, 29, 38, 45, 52, 58; 180 validated, 212 hypothetical; a list compiled by Miano et al. (30)] from Dstncorn1 and A.BY wild-type controls and found that 53 of them are upregulated in Dstncorn1 at a statistically significant level (P < 0.01, Supplemental Table S3). To confirm the upregulation of these SRF target genes in the Dstncorn1 cornea, we performed real-time PCR analyses for 17 genes shown in Fig. 5C. All of the genes tested are upregulated in Dstncorn1 mice, validating the microarray result. We also found 24 SRF target genes in the upregulated gene list for Dstncorn1-2J mice (Supplemental Table S4), which includes the Dstn gene. The microarray experiment did not detect Srf signals above the baseline in either Dstncorn1-2J or B6 wild-type cornea (no “P” call in any of the chips). Consistent with the microarray data, the real-time PCR analysis showed that the Srf expression level was extremely low in both Dstncorn1-2J and B6 wild-type cornea without a statistically significant difference between them (data not shown). The SRF protein was not detected in the Dstncorn1-2J cornea by immunohistochemistry (data not shown).

Since in vitro studies have shown that actin dynamics control SRF activity by regulating the subcellular localization of myocardin-related SRF coactivator, MAL, we examined whether the expression pattern of the MAL protein is affected in the Dstncorn1 cornea. The whole mount corneas from Dstncorn1 and A.BY wild-type mice were subjected to immunohistochemistry using an anti-MAL antibody. The MAL signal was detected in both the cytoplasm and nucleus in the corneal epithelial cells of A.BY wild-type mice (Fig. 5D). In the cornea of Dstncorn1 mice, the increased level of MAL signal was observed both in the cytoplasm and nucleus (Fig. 5D). The MAL signal in B6 wild-type and Dstncorn1-2J cornea was similar to that observed in A.BY wild-type cornea (data not shown). These results show that, while its subcellular localization pattern was not affected, the protein expression level of MAL is higher in the Dstncorn1 cornea compared with that in the A.BY wild-type cornea.

DISCUSSION

Dstn mutations lead to dramatic alteration of the gene expression profile in vivo.

In this study, we demonstrated that mutations in the Dstn gene, a regulator of actin dynamics, leads to dramatic alteration of the gene expression profile in the cornea of Dstncorn1 and Dstncorn1-2J mice. DSTN is known to regulate actin dynamics by enhancing the depolymerization of F-actin into G-actin and promoting filament severing (11, 28). This is accomplished by binding of DSTN to actin subunits in F-actin, which results in the conformational change of F-actin (27). This conformational change is believed to be responsible for the increase in the dissociation rate for actin subunits from the pointed (minus) end of F-actin and for the weak filament-severing activity. Through these functions, DSTN regulates the balance between two pools of cytoplasmic actin, F-actin and G-actin. Consistent with this notion, the F/G-actin ratio is markedly increased in the cornea of Dstn mutant mice (Fig. 2). Our results suggest that actin dynamics and the status of actin (G- or F-actin) have a strong influence on the gene expression profile in vivo.

Allelic Dstn mutants show different levels of F-actin accumulation and gene expression changes.

We previously identified that Dstncorn1 is a null mutation with complete deletion of the entire coding sequence of the Dstn gene, while Dstncorn1-2J is a missense mutation that results in a proline-to-serine amino acid change (19). In the current study, we showed that these two mutations both lead to the increase of F-actin. However, the level of F-actin is much higher in Dstncorn1 mice compared with Dstncorn1-2J mice (Figs. 1 and 2). There are two possibilities underlying this difference. First, Dstncorn1-2J may be a hypomorphic allele with reduced DSTN activity, thus causing milder F-actin accumulation. The Dstncorn1-2J mutation occurred at a proline residue that is conserved across species (7, 26) and is located in the region implicated in both F- and G-actin binding (24). It is possible that the Dstncorn1-2J mutation reduces the binding of DSTN to F- and G-actin but does not abolish DSTN activity completely. Alternatively, the difference in the level of F-actin may be due to the genetic background difference between Dstncorn1 and Dstncorn1-2J. In double labeling and actin fractionation experiments (Figs. 1 and 2), we noted that the F-actin level in the A.BY wild-type cornea is significantly higher compared with that of the B6 wild-type cornea. It is possible that this basal difference in the F-actin level between inbred strains is reflected to the F-actin levels in Dstncorn1 and Dstncorn1-2J mice.

The current microarray study showed that the number of differentially expressed genes identified in the Dstncorn1 cornea (1,226 genes) is much larger than that in the Dstncorn1-2J cornea (202 genes). Moreover, the level of change for commonly disregulated genes tends to be higher in the Dstncorn1 cornea (Fig. 4). These findings indicate that the level of F-actin may correlate with the level of change in the gene expression profile.

Perturbation of SRF target genes by Dstn mutations in vivo.

The functional gene categories that are most significantly affected by the Dstncorn1 mutation are those associated with the cytoskeleton, which include many SRF target genes. Further analysis of our microarray data identified 53 SRF target genes to be upregulated (P < 0.01) in the Dstncorn1 cornea, strongly suggesting that SRF-dependent transcription is activated in this mutant. Through studies using the cell culture system, SRF has been found as a transcription factor that links actin dynamics and gene expression. It was shown that depletion of G-actin is required for the activation of SRF and transcription of its target genes (40, 47). In the proposed model, cytoplasmic G-actin binds to myocardin-related SRF coactivator, MAL, and prevents its nuclear translocation (32). A more recent study showed that nuclear G-actin also plays an important role in this mechanism by promoting actin-dependent nuclear export of MAL and by preventing SRF activation (54) through its binding to MAL. In response to actin polymerization, depletion of cytoplasmic G-actin promotes nuclear import of MAL, and depletion of nuclear G-actin (and, therefore, diminished interaction between nuclear G-actin and MAL) results in decreased nuclear export of MAL and allows SRF activation (54). Based on this model, the observed accumulation of F-actin with concomitant depletion of cellular G-actin level that is expected to occur in the Dstncorn1 cornea may activate this pathway and promote SRF-dependent gene transcription. However, the result of our immunohistochemical analysis for MAL indicated that the molecular mechanism of the phenomenon observed in the Dstncorn1 cornea may not be as simple. While we did not observe the nuclear translocation of MAL in the Dstncorn1 cornea, a significant increase in the MAL signal was observed in both the cytoplasm and nucleus. More MAL proteins are available in the nucleus of Dstncorn1 corneal epithelial cells compared with that in the wild-type mice, which could account for the upregulation of SRF-dependent transcription. It remains to be investigated how the Dstncorn1 mutation leads to the observed increase of MAL, and whether this increase is solely responsible for the activation of SRF-dependent transcription. Nevertheless, our results strongly suggest that the link between actin dynamics and SRF-dependent transcription exists in vivo.

In contrast to the Dstncorn1 cornea where upregulation of Srf and nuclear localization of SRF are observed, the expression level of Srf was very low and the SRF protein was not detectable by immunohistochemistry in the Dstncorn1-2J cornea. Nevertheless, 24 SRF target genes are upregulated in the Dstncorn1-2J cornea (P < 0.01), indicating that SRF-dependent transcription may be activated in this mutant as well. These results suggest that a relatively low level of SRF expression is sufficient for the transcription of its target genes to be activated, although the level of activation appears to be much higher (53 vs. 24 SRF target genes upregulated in Dstncorn1 and Dstncorn1-2J cornea, respectively) with increased SRF expression.

Based on the possible difference in SRF activation between allelic Dstn mutant mice, SRF target gene expression may contribute to the development and severity of the corneal abnormalities in each mutant. SRF activation has been shown to induce the expression of genes upstream of growth factor and cell cycle genes (15, 20). It has also been shown to mediate angiogenesis (12) as well as cell proliferation (16). Additionally, SRF-dependent transcription is essential for the normal expression of contractile genes (31, 36) and the organization of normal cytoskeleton and contractile generation systems (reviewed in Ref. 30). It will be of benefit to examine the effect of SRF inactivation in the cornea of Dstn mutants, since it may determine the in vivo effects of SRF-dependent transcription as regulated by alterations in actin dynamics.

Affected functional categories and phenotypes.

In addition to the molecular mechanisms that are directly affected by Dstn mutations and abnormal actin dynamics, our microarray data should also reflect the physiological changes that have occurred to the corneal tissue originally due to the mutations. Although these changes may not be the direct effect of the mutations on gene expression, they could indicate the molecular pathways through which Dstn mutations lead to observed phenotypes in vivo. Considering the phenotypes observed in Dstn mutant mice, there are some functional categories of genes identified by our analysis that are of particular interest. At the age (P14) used in this study, Dstncorn1 and Dstncorn1-2J mice both exhibit an accumulation of F-actin in the superficial layers of the corneal epithelium. Through the microarray analysis, genes that are associated with the GO term “cytoskeleton” and terms related to actin and/or cytoskeleton were identified as upregulated in the Dstn mutants. While the biological significance of this upregulation is unknown, this does show that a disruption in proper actin dynamics in vivo can modify gene expression to accommodate cellular changes in the structure of the cytoskeleton. Further analyses of the genes associated with cytoskeletal structure are necessary to determine the biological roles for these proteins and whether downstream signaling is possibly affected by their upregulation.

At P14, Dstncorn1 mice show epithelial cell hyperproliferation, and the neovascularization phenotype is just at the onset. Dstncorn1-2J mice display a milder hyperproliferation phenotype at this time point and do not display neovascularization at any age. “Cell cycle”-associated categories that are enriched in the upregulated gene list for Dstncorn1 mice may include genes that are responsible for the hyperproliferation of the corneal epithelial cells. In particular, “M phase of mitotic cell cycle” is overrepresented in Dstncorn1 mice, indicating that this particular phase of the cell cycle may be affected. While GO terms associated with the cell cycle were not found to be enriched for the Dstncorn1-2J mutant, this may be because only a small number of genes associated with this change are upregulated in this mutant. Additionally, this may reflect a change occurring at the protein level that contributes to the hyperproliferation phenotype or the fact that genes causing this phenotype may not have been previously associated with an increase in cell proliferation.

Inflammation associated terms were also overrepresented in the Dstncorn1, but not the Dstncorn1-2J, mutant. Since inflammatory cells and molecules are known to participate in the angiogenic process (35), this group of genes could be involved in the induction of the neovascularization phenotype in Dstncorn1 mice. The list of genes associated with these functional categories obtained in this study would serve as an entry point for further investigation toward the identification of the molecular bases for the Dstncorn1 and Dstncorn1-2J phenotypes.

Functional categories of downregulated genes.

One group of GO terms overrepresented in the list of genes downregulated in the Dstncorn1 and Dstncorn1-2J cornea is associated with the GTPase regulator activity. Among them, the GO term “guanyl-nucleotide exchange factor (GEF) activity” was enriched in both mutants. This term represents molecules that stimulate the exchange of GDP for GTP in association with the GTPases (http://www.geneontology.org/), which act as binary switches by cycling between an inactive (GDP-bound) and an active (GTP-bound) state (21, 48). GEFs promote the generation of activated forms of GTPases that are capable of recognizing downstream targets or effectors (49). Therefore, downregulation of GEFs in Dstn mutants likely results in an alteration of cell signaling through GTPases. Particularly in the downregulated gene list for Dstncorn1 mice, this GO term category included three Rho GEFs. This group was of interest since Rho GEFs stimulate the generation of the activated form of the Rho GTPase (43), which is the upstream negative regulator of DSTN and other ADF/cofilin family members (9, 22, 44, 51). The activated Rho phosphorylates Rho-associated coiled-coil containing protein kinases (3), which in turn phosphorylates the LIM kinase that phosphorylates ADF/cofilin family members (9, 44). The phosphorylation of ADF/cofilin family members by the LIM kinase prevents them from binding actin. Downregulation of Rho GEFs in the Dstncorn1 cornea should decrease the amount of activated Rho (an upstream negative regulator of ADF/cofilins), which would result in the increase of the binding of ADF/cofilins to actin. Therefore, it appears that there may be a feedback mechanism that detects the loss of DSTN function or the abnormality in actin dynamics and compensates this by increasing the activity of the ADF/cofilin family of proteins.

Summary

In summary, this study demonstrated that Dstn mutations and resultant aberrant actin dynamics lead to dramatic alterations of the gene expression profile in the cornea. In particular, we have shown that the expression of genes related to the structure of the cytoskeleton are strongly affected. We also have identified genes that may be responsible for the development of Dstn mutant phenotypes and provided strong evidence that the change in actin dynamics results in the activation of SRF-dependent transcription in vivo. Dstn mutant mice should serve as in vivo models to further investigate cell signaling mechanisms and transcriptional regulation affected by actin dynamics.

GRANTS

This work was supported by National Institutes of Health (NIH) Grants R01 EY-016108 and P30 EY-016665. Support for A. M. Verdoni was partially provided by the NIH Predoctoral Training Program in Genetics (T32 GM-07133).

Acknowledgments

The authors thank Drs. Patsy Nishina and Richard Smith for continuous support and scientific discussions, Britt Johnson and Xinjie Xu for critical review of this manuscript, Dr. Ivan Rayment and Bob Smith for the use of the ultracentrifuge and technical advice, Satoshi Kinoshita for generating frozen sections, Lance Rodenkirch and Nelson Wayne Davis for the technical support, Sandra Splinter BonDurant for scientific discussion, and the University of Wisconsin-Madison Genetics Confocal Facility for the use of the confocal microscope.

Footnotes

  • * A. M. Verdoni and N. Aoyama contributed equally to this work.

  • 1 The online version of this article contains supplemental material.

  • Addresses for reprint requests and other correspondence: S. Ikeda, Dept. of Medical Genetics, Univ. of Wisconsin-Madison, 425-G Henry Mall, Rm. 5350 Genetics/Biotech, Madison, WI 53706 (e-mail: ikeda{at}wisc.edu); A. Ikeda, Dept. of Medical Genetics, Univ. of Wisconsin-Madison, 425-G Henry Mall, Rm. 5322 Genetics/Biotech, Madison, WI 53706 (e-mail: aikeda{at}wisc.edu).

  • 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.

REFERENCES

View Abstract