A Novel Mutation in β Integrin Reveals an Integrin-Mediated Interaction between the Extracellular Matrix and cki-1/p27KIP1

The cell-extracellular matrix (ECM) interaction plays an essential role in maintaining tissue shapes and regulates cell behaviors such as cell adhesion, differentiation and proliferation. The mechanism by which the ECM influences the cell cycle in vivo is poorly understood. Here we demonstrate that the β integrin PAT-3 regulates the localization and expression of CKI-1, a C. elegans homologue of the cyclin dependent kinase inhibitor p27KIP1. In nematodes expressing wild type PAT-3, CKI-1::GFP localizes primarily to nucleoli in hypodermal cells, whereas in animals expressing mutant pat-3 with a defective splice junction, CKI-1::GFP appears clumped and disorganized in nucleoplasm. RNAi analysis links cell adhesion genes to the regulation of CKI-1. RNAi of unc-52/perlecan, ina-1/α integrin, pat-4/ILK, and unc-97/PINCH resulted in abnormal CKI-1::GFP localization. Additional RNAi experiments revealed that the SCF E3 ubiquitin-ligase complex genes, skpt-1/SKP2, cul-1/CUL1 and lin-23/F-box, are required for the proper localization and expression of CKI-1, suggesting that integrin signaling and SCF E3 ligase work together to regulate the cellular distribution of CKI-1. These data also suggest that integrin plays a major role in maintaining proper CKI-1/p27KIP1 levels in the cell. Perturbed integrin signaling may lead to the inhibition of SCF ligase activity, mislocalization and elevation of CKI-1/p27KIP1. These results suggest that adhesion signaling is crucial for cell cycle regulation in vivo.


Introduction
Integrins are ab heterodimeric receptors that mediate bidirectional interactions between cells and extracelluar matrix (ECM) [1]. In mammals, 13a and 8b chains comprise more than twenty heterodimers and play important roles in controlling cell behaviors such as cell adhesion, migration and proliferation [2]. Among the b subunits, b1 integrin is broadly expressed and has multiple splice variants. For example, four b1 splice variants, b1A, b1B, b1C and b1D differing in their cytoplasmic tails, are expressed in many tissues [3]. b1A is the dominant splice form and is expressed ubiquitously [3]. b1D is produced by alternative splicing and is found in striated muscle cells only [4]. Both b1A and b1D forms are localized to focal adhesions and retain the conserved NPxY phosphorylation motif [5,6]. However, the b1B variant, expressed in keratinocytes and hepatocytes, fails to localize to focal adhesions and exhibits dominant negative activity to b1A-paired integrins [7,8]. b1B is the result of mis-splicing of intron 7, and retains intronic sequence in its mRNA [9]. b1C is expressed in normal tissues, such as the prostatic epithelium, and is downregulated in cancer cells [10]. b1C integrin is produced from an alternative splicing event in the cytoplasmic tail of b1 integrin, usually includes exon C and results in a protein 27 amino acids longer than the regular b1A splice form [5,11].
In many cancerous conditions, integrins lose their connection to the ECM or change their expression patterns [3,12,13]. The ECM also undergoes remodeling, resulting in abnormal deposition of proteins or increased ECM stiffness. A change in ECM composition or mechanical properties may upregulate integrin signaling, which promotes cell survival, adhesion and proliferation [14,15]. For example, cell detachment from the ECM increases the level of cyclin dependent kinase (CDK) inhibitors thereby preventing advancement to S phase of the cell cycle [16].
In some cases, integrin signaling can promote cell cycle arrest [17]. For example, the expression of integrin b1C in mammalian cells increases the level of p27 KIP1 , a CDK inhibitor [18]. In contrast, lowering the level of p27 KIP1 allows activation of the CDK/cyclin complex and promotes the cell cycle transition from G1 to S [19][20][21][22]. In these cells, adhesion to the ECM activates an E3 ubiquitin ligase that is essential for the degradation of p27 KIP1 . The expression of SKP2, an important component of the SCF ubiquitin ligase (E3) complex, is also dependent on cell adhesion at the G1 to S transition [23]. However, there is little information available about how integrin signaling regulates the level of cell cycle inhibitors like p27 KIP1 in vivo.
The nematode Caenorhabditis elegans expresses only two integrins, PAT-3/INA-1 [24] and PAT-3/PAT-2 [25], which simplifies the analysis of genetic interactions between integrin and cell cycle control genes. Overexpression of C. elegans p27 KIP1 /CKI-1 has been found to induce growth arrest and the cki-1 null mutation results in hyperplasia of tissues such as the hypodermis, the vulva and the intestine [26][27][28]. The disruption of cki-1 also results in the production of extra distal tip cells from the somatic gonad lineage [29,30]. The activity of cki-1 is regulated by the coordination of stage-specific cellular events such as binding to the CDK/cyclin complex and a sharp increase in expression at the late larval and young adult stages [26,[31][32][33].
This study the role of pat-3(sp) (previously known as pat-3(b1C)) builds on our previous work on the function of PAT-3 signaling in C. elegans [34]. pat-3(sp) is a frameshift mutation in the splice acceptor region (ag to aa) that abolishes conserved interaction domains such as the NPxY motifs and creates a splice variant with an extra 19 amino acids. The pat-3(sp) animals not only produce mutant pat-3, but also express the regular splice form due to utilization of an unusual splice acceptor [34]. pat-3(sp) has similarities to the human b1B and b1C integrins. The pat-3(sp) mutant is similar to b1B because it retains intron sequence in the mRNA and the transgenic line expresses mutant and normal splice forms simultaneously. The mutant is similar to the b1C variant in that it is expected to produce a longer PAT-3 lacking the NPxY motifs [34].
In this study, we assessed the role of pat-3(sp) in cell cycle regulation. Briefly, in transgenic pat-3 rescued lines carrying pat-3(sp) [34], cki-1::GFP [26] was upregulated and exhibited a distinct sub-nuclear localization compared to wild type animals. RNA interference analyses revealed that the localization and level of CKI-1 are mediated by focal adhesion molecules and the SCF E3 ubiquitin ligase complex [35,36]. Taken together, our findings suggest that integrin signaling, in conjunction with SCF E3 ligase complex activity, plays a crucial role in the localization and level of CKI-1 in vivo.

bpat-3(sp) Increases CKI-1 Levels and Exhibits a Distinct Localization in Nuclei
PAT-3 b integrin is expressed in virtually all tissues in the nematode C. elegans [37], and is required for muscle development and function. Null mutations in pat-3 cause a fully penetrant embryonic arrest due to defective muscle elongation [25,38]. Previously, we created a mutation at the intron 7 splice junction in the cytoplasmic tail of PAT-3 integrin ( Figure 1) [3,34]. The transgenic rescued line, pat-3(sp), is viable but exhibits coldsensitive larval arrest with gonad and muscle defects. We found that pat-3(sp) expresses the non-spliced as well as the spliced pat-3 mRNA, suggesting that mutant pat-3 might inhibit the function of wild type pat-3. In this study, we have expanded our analysis to investigate the molecular function of pat-3(sp) [34].
Studies of the mammalian b1B and b1C integrins in mammalian cells revealed that expression of these integrins suppresses cell adhesion and proliferation [5] and upregulates the expression of p27 KIP1 [39]. To investigate a potential linkage between PAT-3 and CKI-1, the C. elegans homolog of mammalian p27 KIP1 [26,29,40], we created rescued transgenic lines containing pat-3(+) or pat-3(sp) genomic DNA by co-injecting with DNA encoding a CKI-1::GFP fusion protein [26,34]. As previously described [26,29], CKI-1::GFP is expressed in hypodermal cells of late L4 and young adult animals. Nuclear expression is observed within an ER meshwork, typical of the hypodermal syncytium (hyp 7) (Figure 2A and 2B) [41]. In pat-3(+) rescued animals, the appearance of CKI-1::GFP is a distinct fluorescent spot within a round green nucleus, suggesting nuclear and nucleolar localization (Figures 2A and 2C). To substantiate our interpretation that CKI-1::GFP localizes to the nucleolus, the ncl-1 gene, disruption of which results in enlarged nucleoli, was depleted using RNAi [42,43]. In the pat-3(+) background, ncl-1 (RNAi) significantly increased the size of the CKI-1::GFP spot. ImageJ analysis showed that the size of the green spot was increased by 2.4 fold when compared to that of control RNAi animals ( Figure S1, Table  S1). Therefore, we conclude that the observed spots on the nuclei are likely to represent nucleolar localization.
In the pat-3(sp) rescued animals, CKI-1::GFP localization was visibly different from that seen in pat-3(+) animals. In contrast to the compact, nucleolar staining seen in pat-3(+) animals, CKI-1::GFP in pat-3(sp) was clumped and accumulated in a ring around a dark center in the nucleus ( Figures 2B and 2D), suggesting mislocalization and possible exclusion from the nucleolus. In addition, the intensity of green fluorescence in pat-3(sp) was increased compared to pat-3(+). In order to test for a possible correlation between the level of CKI-1::GFP and the integrin (pat-3(+) or pat-3(sp)) expressed, we first analyzed the amount of CKI-1::GFP in the pat-3 rescued lines. Protein lysates were prepared from an equal number of L4/young adult transgenic animals and tested for CKI-1::GFP protein levels. CKI-1::GFP level in pat-3(sp) was ten fold more intense than that seen in pat-3(+) lysates ( Figure 3A), suggesting that PAT-3 signaling may control CKI-1 levels.
Because the immunoblot results revealed that pat-3(sp) animals produced more CKI-1::GFP protein than pat-3(+), we next assessed the effect on cki-1 transcription. RNA from each rescued line was isolated and analyzed for the amount of pat-3 or cki-1 mRNA using RT-PCR ( Figure 4A). We also measured the cki-1 mRNA level in BU7221, a pat-3(sp) rescued line without cki-1::GFP [34]. No significant differences were seen in any of the experiments, suggesting that pat-3(sp) does not significantly increase the level of cki-1 mRNA compared to controls ( Figure 4B). Integrin Signaling Regulates CKI-1 Localization In order to define genetic pathways that link PAT-3 integrin to CKI-1, a series of RNAi experiments were performed. We hypothesized that the integrin effect on the localization of CKI-1::GFP in pat-3(sp) would be mediated by genes that interact with the cytoplasmic domain of b integrin. Thus, candidate genes were selected from known focal adhesion components [44]. Previous analysis of embryonic muscle development identified 20 essential genes, mostly encoding components of dense bodies and M-lines, which are analogous to focal adhesions [44][45]. Integrins are located in the base of these structures and anchor the sarcomeres to the basement membrane. Data from the SAGE database [37] indicated ten of the integrin signaling genes were expressed in the hypodermis. To screen for the genes involved in CKI-1 localization, we tested if RNAi depletion of these integrin signaling components [44] in pat-3(+) would result in mislocalization of CKI-1::GFP in the nucleoplasm, similar to that seen in pat-3(sp) animals (Results summarized in Table 1).
In pat-3(+) animals, pat-3(RNAi) resulted in CKI-1::GFP accumulation in the nucleoplasm similar to that seen in pat-3(sp) ( Figures 5B and 2D). Next, integrin a subunits were depleted. Depletion of ina-1 in the pat-3(+) animals also resulted in abnormally clumped CKI-1::GFP (Figures 5C), suggesting that the CKI-1 localization is integrin dependent. Among the focal adhesion genes, pat-4/ILK [46], unc-97/PINCH [47] and pat-6/ parvin [48] together form an IPP complex, which is implicated in the control of signaling pathways by the phosphorylation of downstream targets [49]. RNAi of pat-4/ILK or unc-97/PINCH in pat-3(+) resulted in the expected uncoordinated phenotypes ( Figure S2) and CKI-1 mislocalization in hypodermal nuclei ( Figure 5D and 5E). However, in pat-6/parvin RNAi animals, CKI-1 maintained its wild-type localization, possibly suggesting that the CKI-1 localization is independent of parvin ( Figure 5F). Because pat-6 RNAi did not result in a strong uncoordinated phenotype ( Figure S2), it is possible that the pat-6 RNAi is not as effective as the RNAi to pat-4 and unc-97. However, our data is consistent with the interpretation that ILK and PINCH are mediating integrin signals to control CKI-1 localization in the nucleus.
Our RNAi screen also found that unc-52/perlecan, a basement membrane component and presumptive integrin ligand [50][51][52], is required for the proper localization of CKI-1. RNAi of unc-52 in pat-3(+) affected the CKI-1 localization pattern ( Figure 5G). In contrast, depletion of other basement membrane components, such as let-2/collagen IV [53], failed to affect the localization ( Figure 5H), suggesting that a subset of ECM components is required for CKI-1 localization. Ubiquitin-mediated Protein Degradation Regulates Localization of CKI-1::GFP Next, we investigated the mechanism by which integrin regulates CKI-1 protein levels without affecting RNA levels. One plausible explanation is that integrin signaling leads to the degradation of CKI-1 [23]. Integrin-triggered p27 KIP1 degradation has been observed in mammalian cells. For example, integrin crosstalk with receptor tyrosine kinase (RTK) induces the production of SCF SKP2 [54], a member of the SCF E3 ubiquitin ligase complex, which binds to the SKP1 [55], CUL1 [56] and FBX-1 [23] E3 ligase complex. This SCF complex targets CDK/cyclin inhibitors such as p27 KIP1 and p21 CIP1 [57].

Immunoblot Analysis Demonstrates a Correlation between CKI-1 Overexpression and Mislocalization
We next investigated whether depletion of the focal adhesion and SCF complex genes would affect expression levels of CKI-1/ p27 KIP1 in addition to affecting nuclear localization patterns (Figures 5; Table 1). Immunoblot analyses of CKI-1::GFP were performed using protein extracts from RNAi-treated adult animals ( Figure 5A and 5B). The amount of CKI-1::GFP generally increased in protein extracts of focal adhesion and SCF complex RNAi treated animals. Depletion of focal adhesion genes such as pat-3, ina-1, unc-97, unc-52, pat-6 and let-2 resulted in an up to 3fold increase in CKI-1::GFP levels ( Figure 3B, Table S1). RNAi of SCF E3 ligase genes such as skpt-1, lin-23, and cul-1 also produced significant increases in CKI-1::GFP levels. (Figure 3C, Table S1). Interestingly, even though nucleolar localization was not affected by pat-6 and let-2 RNAi, protein levels were significantly increased ( Figure 3B). Our analysis suggested that disruption of integrin signaling or SCF-mediated protein degradation can result in the mislocalization as well as increased expression of CKI-1::GFP. However, the let-2 and pat-6 RNAi results suggest there may not be a direct relationship between protein levels and localization.

Discussion
In this study, integrin regulation of CKI-1 was assessed in vivo. Our analysis revealed that CKI-1/p27 KIP1 had an abnormal localization pattern in the nucleoplasm of animals expressing a mutant integrin, pat-3(sp). In the mutant animals, CKI-1::GFP was overexpressed and clumped in the nucleoplasm, while in animals expressing wild type integrin, CKI-1::GFP was localized predominantly to the nucleolus (Figure 2). Further studies revealed that the amount of CKI-1::GFP protein was increased in the pat-3 mutant. To delineate the genetic pathway responsible for the upregulation of CKI-1/p27 KIP1 , we depleted focal adhesion and SCF E3 ubiquitin ligase genes in pat-3(+) animals and found that these genes are essential for the proper localization and expression of CKI-1::GFP. We conclude that the inhibition of integrin signaling and protein degradation significantly affects CKI-1 protein localization and expression level in vivo.
Previous studies have suggested a link between the mammalian integrin splice variant b1C and p27 KIP1 . For example, increased expression of b1C integrin elevated p27 KIP1 in prostate cancer cell lines [13,21,61]. Our pat-3(sp) mutant splice form is an artificial variant [34], however, our study found that pat-3(sp) behaves similarly to the b1C variant of mammalian b1. pat-3(sp) increases CKI-1/p27 KIP1 expression, possibly leading to a cell cycle arrest.

CKI-1 is Localized to the Nucleolus
Our observations of nuclear morphology and experiments using ncl-1 (RNAi) strongly suggest a nucleolar localization for CKI-1::GFP in pat-3(+). In contrast, CKI-1::GFP appeared clumped and was irregularly distributed in the nucleoplasm of pat-3(sp) animals, suggesting ectopic accumulation of CKI-1. This nucleolar to nuclear transition in CKI-1::GFP localization suggests a possible link between CKI-1 and nucleolar function.
The nucleolus is the main site of ribosome biogenesis and ribosomal RNA (rRNA) synthesis [62]. The rRNAs are synthesized and assembled into a ribosomal complex in the nucleolus. In The GAPDH (glyceraldehyde-3-phosphate dehydrogenase) gene, gpd-1, was used as an mRNA loading control. The level of each transcript was measured using ImageJ software. Quantification of cDNA using ImageJ software revealed that cki-1 mRNA levels were similar in all lines compared (Table S1). Panel B: cki-1 mRNA was measured in bpat-3 rescued lines lacking the CKI-1::GFP construct. Lanes 1-3 are N2, JE443 pat-3(+), and BU7221 pat-3(sp). The GAPDH gene, gpd-1, was used as an mRNA loading control. Quantification of cDNA levels using ImageJ software revealed that cki-1 mRNA levels were similar in all lines compared (Table S1) mammals, stressed ribosomal synthesis leads to cell cycle arrest via increased p27 KIP1 levels [63]. In addition, mutations in the cytoplasmic tail of b4 integrin or p27 b4 binding protein (p27 BBP/ eIF6 ) result in an inability of b4 to localize to hemidesmosomes and a defect in assembly of the 80S ribosomal subunit, suggesting a connection between integrin signaling and ribosome biogenesis  [64][65][66]. Although further studies are required, we speculate that the splice defective pat-3(sp) integrin may cause a decrease in ribosome biosynthesis.

Proper CKI-1 Localization is Linked to Integrin and Integrin-associated Molecules
Integrins are ab heterodimers and both subunits are necessary for integrin function. Therefore, depletion of the a subunit would have the same effect as disruption of the b subunit, PAT-3. Indeed, CKI-1::GFP mislocalization and increased expression were observed when the a integrin gene ina-1 was depleted by RNAi, suggesting that CKI-1 upregulation is a result of disrupted integrin function and that integrin activity is normally required for proper CKI-1 localization and protein levels. In addition, RNAi analysis of the IPP complex showed that pat-4/ILK [46,67], unc-97/ PINCH [47] and pat-6/parvin [48] are also required for correct expression and/or localization of CKI-1. Some studies have suggested that these components act together and are degraded if pat-4/ILK or unc-97/PINCH is not present [68] but other studies, including ours, suggest these molecules may have independent roles [69,70]. In addition, PINCH or parvin binding to ILK is mutually exclusive, suggesting that PINCH might provide a different mode of signaling than parvin [71,72]. ILK has been shown to play an important role in cell proliferation in tissue culture [73] and PINCH is frequently upregulated in human cancers [74]. Disruption of ILK or PINCH inhibits cell proliferation and increases the expression level of p27 KIP1 and pRb [75], suggesting that cells might generally respond to perturbed cell adhesion by inducing p27 KIP1 expression and halting the cell cycle. In addition, our analysis links the ECM ligand, unc-52/ perlecan, to CKI-1 localization, suggesting that perlecan-bound integrin might regulate CKI-1 by activating pat-4/ILK and unc-97/PINCH.

Integrin Signaling May Influence the Cell Cycle by Regulating CKI-1 Level via SCF Ubiquitin Ligases
We propose a potential model for the role of integrin signaling in CKI-1 regulation. Our preferred model assumes the presence of functional integrin in the hypodermal cells. Integrin is activated by binding to ECM ligands and signaling is initiated and propagated by molecules such as pat-4/ILK and unc-97/PINCH to SCF ligase which degrades CKI-1 ( Figure 6). pat-3(sp) may interfere with the formation of the SCF complex and the degradation of CKI-1 by inhibiting the function of wild-type PAT-3 integrins or by acting as a non-functional b subunit that significantly dilutes integrin signaling [8]. SAGE analysis indicates that pat-3 and ina-1 are expressed in hypodermal cells [79,80] consistent with a previously identified role for ina-1/a integrin function in hypodermis [24]. Although we have identified a role for the SCF complex in CKI-1 degradation, our work does not specifically address whether SCF activation is directly linked to integrin signaling in a linear manner, as displayed in the model (Figure 6). Future genetic studies should determine the cell autonomy and epistatic relationships of the genes in the pathway from integrin to cell cycle control.
In addition to muscle and gonad morphogenesis, our work identifies another important function of integrin in C. elegans, the regulation of CKI-1/p27 KIP1 . This finding brings new insight onto cell cycle control. Integrins appear to regulate the level of p27 KIP1 via signaling mediators such as pat-4/ILK and unc-97/PINCH and to maintain SCF ubiquitin ligase activity. Importantly, unc-52 RNAi produces a similar phenotype, suggesting that the cellmatrix interaction balances the amount of CKI-1 in the cell. This information will be useful in understanding the mechanism on how the cell-ECM interaction regulates cell cycle progression.

Phenotype Characterization and Nuclei Expression Pattern Identification
To characterize rescued lines, young adult worms were mounted in a drop of M9 buffer containing 1% NaN 3 (Sigma Chemical Co., St. Louis, MO) or 0.5 mM levamisole on a 24660 mm coverslip coated with 4% agarose and examined on a Nikon TE2000-U Diaphot epifluorescence microscope. Images were captured using a CoolSnap cf monochrome camera (Roper Scientific, Tucson, AZ) and analyzed with Metavue imaging software (version 7.5, Molecular Devices Co., Downingtown, PA). Typically, CKI-1::GFP became visible at the late L4 stage. In the wild-type rescue pat-3(+) animals, GFP was apparently nucleolar: a strong green spot on a larger green nucleus. To photograph the image, the camera was set at the exposure time of 2000 milliseconds. However, in pat-3(sp) animals, the CKI-1::GFP expression was much brighter than in the pat-3(+). For pat-3(sp), images were taken at the exposure time of 300 milliseconds. About 20 hypodermal nuclei in the midbody, an area including vulva at the ventral midline, of each animal was observed for nuclear morphology characterization. Displayed images are patterns seen most commonly under the described conditions.
To analyze the size of nucleoli, the area was measured using ImageJ software (version 1.33, National Institute of Health, Rockville, MD) [85]. To determine the nucleolar to nuclear ratio, the area value of nucleolus, a brighter spot on a nucleus, was divided by that of nucleus. Five measurements for each rescued line were averaged for comparison.

Reverse Transcription Polymerase Chain Reaction (RT-PCR)
To analyze cki-1, pat-3, and gpd-1/GAPDH mRNA levels, animals were partially synchronized by isolating embryos using 20% alkaline hypochlorite solution [86]. After 48 hours, forty L4 to young adult animals were picked into 10 mL of M9 buffer. RNA was extracted using 250 mL Tri-Reagent (Sigma-Aldrich, St. Louis, MO) and 50 mL chloroform (1/5 volume of Tri-Reagent) and RNA was precipitated from the extract with isopropanol. Approximately 1 mg of total RNA was used to synthesize cDNA with Transcriptor Reverse Transcription Kit (Roche, Carlsbad, CA) primed with random hexamers in a total 20 mL reaction volume. Total of 1 mL cDNA was used in PCR amplification with cki-1 primers, pat-3 primers, and with control gpd-1/GAPDH primers previously described [34]. Primer sequences listed below were used for amplification: CKI1 Forward 2: 59-GGAGTTCTACAGAACC-39 CKI1 Reverse 2: 59-CACCGGAGACAGCTTG-39 PAT3PT Forward1: 59-CTCAACGAAACTACACCCTGCC-39 PAT3PT Reverse 1: 59-TTAGTTGGCTTTTCCAGCGTA-TACTGG-39. Figure 6. Model for the integrin regulation of CKI-1. When ECM ligand binds to integrin on the cell surface, the IPP complex delivers signals to the SCF E3 ligase complex resulting in sequestration of CKI-1/p27 to the nucleolus. In pat-3(sp), the mutant b integrin interferes with the function of wild-type pat-3(+). Consequently, integrin does not signal to the IPP complex. We suggest the reduced integrin signals may leave the SCF complex inactive. This allows an increase in the amount of CKI-1/p27 and alters the location of the protein in the nucleus. doi:10.1371/journal.pone.0042425.g006 Table 2. Transgenic C. elegans used in this study.

Mutant designation
Constructs injected

Immunoblot Analysis
For quantitative analysis of CKI-1 protein levels in each strain, we first picked 30 young adults into 10 mL of M9 buffer and 10 mL of 2X Laemmli Sample Buffer (Bio-Rad Laboratories, Hercules, CA) premixed with 1:1000 b-mercaptoethanol. Sample solutions were then boiled for 10 minutes at 100uC and electrophoresed through a 10% SDS-polyacrylamide gel at 180 V for 1 hour. Isolated proteins bands were electrotransferred onto a nitrocellulose membrane (Whatman Ltd., Dassel, Germany) using wet transfer at 100 V for 75 minutes in BSN transfer buffer with no methanol or SDS. This nitrocellulose membrane was then blocked for 1 hour in 5% milk solution at room temperature. Rabbit polyclonal IgG anti-GFP antibody (ab290, Abcam Inc., Cambridge, MA) at 1:2000 was applied overnight at 4uC as primary antibody and goat anti-rabbit IgG HRP conjugated (ab6721, Abcam Inc., Cambridge, MA) at 1:5000 was applied for 1 hour at RT. For control blots, LS25, a monoclonal antibody against UNC-54, or MH33, a monoclonal antibody against a gut specific intermediate filament protein, were diluted at 1:1000 and 1:2000. The primary antibody solutions were applied and detected by the goat anti-mouse IgG HRP conjugated (Sigma Chemical, Mo) secondary antibody. ECL chemiluminescence reagents (Thermo Fisher Scientific, Rockford, IL) were added to the membrane for 1 minute before exposure to the ULTRA-LUM gel imager (Ultra-Lum Inc., Claremont, CA) and analyzed with UltraQuant software (Ultra-Lum Inc., Claremont, CA). Individual band intensity was quantified using ImageJ software (version 1.66, National Institute of Health, Rockville, MD) that measured the integrated density of each band to analyze the intensity of bands.

RNA-mediated Interference of Gene Expression (RNAi) Analysis
C. elegans RNA interference analysis was performed using the bacterial feeding method [87,88]. In addition to the standard RNAi protocol, we synchronized the stage of animals; embryos were collected using the standard 20% alkaline hypochlorite solution method [86]. After washes, collected embryos were placed onto RNAi plates, which were incubated in 20uC for 3 to 4 days until young adulthood before characterization. About 20 hypodermal nuclei in the midbody of transgenic worms were observed for the characterization of nuclear morphology using Metavue software. To verify the efficiency, all RNAi animals were examined for behavioral phenotypes linked to the RNAi gene. For example, focal adhesion gene RNAi was subjected to thrashing assays [89] because uncoordinated (Unc) phenotypes were previously reported for RNAi of these genes ( Figure S2). Each RNAi-inducing plasmid (Geneservice, Hinxton, UK) used in this study was isolated and sequenced in order to verify the gene targeted by the construct.