|Home | About | Journals | Submit | Contact Us | Français|
Rho GTPases are small monomeric G-proteins that play key roles in many cellular processes. Due to their widespread expression and broad functions, analyses of Rho GTPase function during late development require tissue-specific modulation of activity. The GAL4/UAS system provides an excellent tool for investigating the function of Rho GTPases in vivo. With this in mind, we created a transgenic toolkit enabling spatial and temporal modulation of Rho GTPase activity in zebrafish.
Transgenic constructs were assembled driving dominant-negative, constitutively-active, and wild-type versions of Cdc42, RhoA and Rac1 under 10XUAS control. The self-cleaving viral peptide F2A was utilized to allow bicistronic expression of fluorescent reporter and Rho GTPase. Global heat shock of hsp70l:gal4+ transgenic embryos confirmed GAL4-specific construct expression. Western blot analysis indicated myc-tagged Rho GTPases were expressed only in the presence of GAL4. Construct expression was confined to proper cells when combined with pou4f3:gal4 or ptf1a:gal4. Finally, transgene expression resulted in reproducible defects in lens formation indicating that the transgenes are functional in vivo.
We generated and validated ten transgenic lines, creating a versatile toolkit for the temporal-spatial modulation of Cdc42, RhoA and Rac1 activity in vivo. These lines will enable systematic analysis of Rho GTPase function in any tissue of interest.
Rho GTPases, a subfamily of Ras GTPases, are small monomeric G-proteins that play key roles in a myriad of cellular processes, including cell cycle progression, cytoskeletal dynamics, cellular polarity, and membrane trafficking (Etienne-Manneville and Hall, 2002; Takai et al., 2001). Rho GTPase activity depends upon a binary molecular switch: when bound to GTP, Rho GTPases are active, when bound to GDP, they are inactive. This switch is tightly regulated within the cell (Boguski and McCormick, 1993; Jaffe and Hall, 2005), as Rho proteins regulate numerous downstream processes through their interactions with a diverse array of effector proteins. Most studies of Rho GTPases have focused on the Rho subfamily proteins: Cdc42, Rac1 and RhoA. Cdc42 was first discovered in Saccharomyces cerevisiae as a protein required for proper cell polarity during budding (Adams et al., 1990). Since its discovery, Cdc42 has been shown to regulate membrane trafficking, actin filament polymerization to form filopodia, and numerous other cellular processes (Erickson and Cerione, 2001). Rac1 stimulates the assembly of lammellipodia and mediates the formation of cell adhesion structures (Bosco et al., 2009; Ridley et al., 1992). RhoA activity leads to the formation of actin stress fibers (Hall, 1998), maturation of focal adhesions (Luo, 2002) and contraction of the cytokinesis furrow (Lai et al., 2005; Piekny et al., 2005). All three proteins are required for cell cycle progression (Olson et al., 1995). Rho GTPases are also thought to be involved in diverse developmental and pathological processes, including axon pathfinding (Bashaw and Klein, 2010; Jin et al., 2005), cell migration (Kardash et al., 2010; Raftopoulou and Hall, 2004), and oncogenesis (Ellenbroek and Collard, 2007; Sahai and Marshall, 2002).
However, these analyses have predominantly utilized cell culture or in vitro methods, limiting insight into how Rho GTPases function in vivo. Indeed, in vivo investigations of the molecular functions of Rho GTPases during animal development have been relatively rare, owing to the need for tissue specific approaches to manipulate their activity. For example, knockout of Cdc42 (Chen et al., 2000) or Rac1 (Sugihara et al., 1998) in mice results in severe pleiotropic defects and early embryonic lethality. Analysis of Rho GTPase activity and function in vivo therefore requires experimental approaches that allow modulation of activity in specific tissues or cell populations, and at specific time points (e.g. Chew et al., 2014; Govek et al., 2005; Heasman and Ridley, 2008; Heynen et al., 2013; Jackson et al., 2011; Luo et al., 1996; Ruchhoeft and Ohnuma, 1999; Wong and Faulkner-Jones, 2000; Xiang and Vanhoutte, 2011).
Zebrafish provide an excellent model for investigation of the molecular function of vertebrate Rho GTPases in vivo (Kardash et al., 2010; Lai et al., 2005; Salas-Vidal et al., 2005; Zhu et al., 2006). Previous studies of Rho GTPase function in developing zebrafish employed microinjection of mRNA to drive global overexpression of wild-type, constitutively active or dominant negative versions (Hsu et al., 2012; Xu et al., 2014; Yeh et al., 2011; Zhu et al., 2008, 2006), or morpholino oligos for transient disruption of Rho GTPase expression (Hsu et al., 2012; Srinivas et al., 2007). Due to the central role of Rho GTPases in early embryogenesis, approaches that modulate global Rho GTPase activity must focus on events occurring very early during zebrafish development, and thus, have not been effective in analyzing the functions of these proteins during later developmental events. Following the development of highly efficient transgenesis techniques in zebrafish (Kwan et al., 2007), transgenic lines have been generated in which the expression of constitutively active and dominant negative versions of different Rho GTPase family members are driven by cell-type-specific promoters (Chew et al., 2014; Choe et al., 2013; Jung and Leem, 2013). Although these tools restrict Rho GTPase construct expression to specific cell and tissue types and, dependent on the promoter, allow functional investigations at later stages in development, systematic comparisons of functional roles of different Rho GTPase members in different tissues is limited to a small number of extant cell-type-specific promoter Rho GTPase transgenic lines (Chew et al., 2014; Choe et al., 2013; Jung and Leem, 2013). Moreover, this strategy necessitates the generation and validation of a new line for each desired promoter and/or Rho GTPase combination, an inefficient and time-consuming approach. The GAL4/UAS system is a powerful transgenic system for enabling the temporal-spatial control of transgene expression (Brand and Perrimon, 1993; Fischer et al., 1988; Ornitz et al., 1991; Scheer, 1999). In this system, a promoter fragment drives expression of the yeast transcription factor GAL4. GAL4 then binds an upstream activating sequence (UAS) to drive transgene expression. Gal4/UAS is widely used in zebrafish, and there are now hundreds of published GAL4 drivers and UAS constructs listed on www.zfin.org. The strength of this system lies in its flexibility: a single GAL4 transgene can be used to drive expression of multiple UAS constructs, enabling researchers to express multiple transgenes within a defined cellular or tissue context.
Taking advantage of this approach, we have created a versatile transgenic toolkit that enables spatial and temporal modulation of Rho GTPase activity in zebrafish. We generated and validated ten GAL4-inducible transgenic lines that express dominant negative, constitutively active, and wild type versions of Cdc42, RhoA, and Rac1, as well as a fluorescent protein marker to highlight expressing cells. We have confirmed GAL4-specific expression of these transgenes, and have demonstrated transgene functionality by reporting reproducible lens phenotypes in induced embryos. These lines now enable systematic tissue-specific investigation of the molecular function of Rho GTPases in vivo.
To generate transgenic lines for GAL4-driven expression of Rho GTPases, we first designed and assembled transgenic constructs encoding wild-type (WT), dominant negative (DN) and constitutively active (CA) human Cdc42, RhoA, and Rac1 (Figure 1). An additional DN cdc42 was generated using Xenopus Cdc42F37A which has recently been utilized in vivo (Kieserman and Wallingford, 2009) and is 98% identical to human Cdc42. We selected the 10X UAS element (Kwan et al., 2007) for our constructs due to this promoter’s strong, GAL4-specific expression (Gabriel et al., 2012; Kwan et al., 2007). Although recent research has raised concern over multigenerational silencing of the repeat-heavy UAS element (Akitake et al., 2011; Goll et al., 2009), other reports have postulated that silencing may be reduced when UAS lines are maintained separately from GAL4 drivers (Choe et al., 2012). Cdc42 and RhoA constructs were assembled with the monomeric and highly photostable mCherry fluorescent protein (Shaner et al., 2004), while EGFP was chosen as the fluorescent reporter for Rac1 transgenics. A myc-tag was also added to enable detection of each Cdc42 and RhoA isoform, and there is a significant amount of literature reporting the use and effective function of C-terminal tags of Rho GTPases (Choe et al., 2012; Disanza et al., 2006; Hussain et al., 2001; Kim et al., 2008; Kroschewski et al., 1999; Lee et al., 2004; Sakurai-Yageta et al., 2008). Furthermore, our constructs were engineered using cDNAs that have been shown to be expressed and functional (Kieserman and Wallingford, 2009; Nobes and Hall, 1995). Nucleotide sequence encoding the self-cleaving viral peptide F2A was inserted between fluorescent reporters and Rho GTPase cDNA sequences to allow bicistronic expression of the fluorescent protein and the GTPase. Rather than requiring a second ribosomal binding event, as is the case for IRES, F2A peptides lead to two protein products via a ribosomal skipping mechanism (Donnelly and Luke, 2001; Szymczak et al., 2004), and when compared to IRES, F2A leads to more efficient bicistronic expression (Chan et al., 2011; Wang et al., 2011). With this rationale, we created three types of constructs: 10xuas:mCherry-F2A-myc-Cdc42XX for the expression of mCherry-F2A and myc-Cdc42, 10xuas:EGFP-F2A-Rac1XX to express EGFP-F2A and Rac1, and 10xuas:mCherry-F2A-myc-RhoAXX, to express mCherry-F2A and myc-RhoA (Figure 1A). In all, we created a total of ten new transgenic constructs as elaborated in Figure 1B. All constructs include the cmlc2:egfp transgenesis marker to allow easy visualization of the presence of constructs. Germline transmission, scored by the appearance of offspring with ubiquitously GFP-positive hearts, occurred in outcrosses of roughly 25% of injected F0 fish.
To determine the spatial expression of F1 alleles while minimizing the potential for severe phenotypes arising from early induction, we crossed F0 adults with hsp70l:gal4 zebrafish, globally heat shocked all transgenic embryos at ~54 hours post fertilization (hpf), and imaged reporter expression at 72hpf. Each line displayed robust and ubiquitous transgene expression following global heat shock (data not shown). No leakiness of transgene expression was observed before heat shock. PCR genotyping confirmed that fluorescent marker-positive embryos were also GAL4-positive and non-fluorescent embryos were GAL4-negative (data not shown), indicating that UAS-driven transgene expression was restricted to fish expressing GAL4 and that there was no observable leaky activation of the UAS promoter.
Non-equimolar expression of the second cistron following F2A ribosomal skipping has been reported (Chan et al., 2011). Because Rho GTPase function is dependent upon prenylation at the C-terminus, Rho GTPase sequence could not be placed upstream of the F2A element, as that would have resulted in the C-terminal fusion of the F2A peptide and the disruption of protein trafficking. Thus, to confirm that Gal4-mediated transgene induction results in expression of both the Rho GTPase and reporter proteins, we performed Western blot analyses of heat shocked F1 embryos of Cdc42 and RhoA constructs (Figure 2). GAL4-induced embryos displayed high levels of myc-tagged protein while no myc-tagged protein was observed in GAL4-negative embryos. This confirms that the F2A peptide is functioning properly by enabling bicistronic expression of mCherry and myc-tagged Rho GTPase protein, both of which are only inducible by GAL4. Western blot analysis of heat shocked F1 embryos of Rac1 constructs using human Rac1 antibody showed strong Rac1 staining in both GFP+ and GFP− siblings, indicating that the Rac1 antibody also detects endogenous Rac1, an unsurprising result considering that human Rac1 protein sequence is 93% identical to zebrafish (data not shown). While expression of all constructs was induced, Cdc42F37A displayed a higher level of expression than the other Cdc42 alleles. One potential explanation for this result is that Xenopus Cdc42F37A has greater stability in zebrafish compared to human Cdc42. However, there also appears to be higher expression of mCherry in Cdc42F37A embryos (Figure 5). Our bicistronic constructs result in the expression of two separate proteins; mCherry should not display any species-dependent enhancement of stability. Tol2 transgenesis often leads to allelic quality differences between founders due to random integration events into the genome; therefore, a more likely explanation is that uas:mCherry-f2a-myc-Cdc42F37A is a particularly well-expressed allele due to its genomic location.
To confirm that the transgenes are spatially and temporally responsive to GAL4, F0 uas:mcherry-f2a-cdc42WT transgenics were crossed to pou4f3:gal4, which express GAL4 in a subset (<20%) of retinal ganglion cells (RGCs), (Xiao and Baier, 2007). To determine if Cdc42 expression is confined to RGCs in the retina, we cryosectioned 4dpf transgenic embryos and imaged the retina for mCherry fluorescence (Figure 3). mCherry fluorescence is confined to the RGC layer in a pattern consistent with previous studies utilizing the pou4f3:gal4 driver (Xiao et al., 2005), indicating our construct expression is restricted to RGCs, and that our transgenic constructs are amenable to cell-specific expression. To determine if cells expressing mCherry also express myc-Rho GTPase, we immunostained for myc in uas:mCherry-f2a-myc-Cdc42WT embryos both with and without the hsp70l:gal4 transgene. Following heat shock, immunostaining reveals strong induction of both mCherry and myc throughout the retina (Figure 4A), which is entirely absent in mCherry-negative siblings (Figure 4B). Next, we crossed uas:GFP-f2a-Rac1WT to ptf1a:gal4, a driver expressing GAL4 in amacrine cells and horizontal cells (Parsons et al., 2009), and stained retinal sections for human Rac1 (Penzes et al., 2000). Rac1 staining was detected throughout the retina, further supporting our hypothesis that this antibody detects endogenous Rac1. However, there does appear to be a higher degree of Rac1 signal in GFP-positive cells (Figure 4C), supporting ptf1a:gal4-driven cell specific induction of the construct. Similarly, immunostaining of hsp70l:GAL4;UAS:mCherry-f2a-RhoAWT showed mCherry and myc colocalization only in mCherry-expressing cells (Figure 4D,E).
Finally, we sought to validate transgene function by determining whether transgene expression leads to distinct morphological phenotypes. The majority of transgenics heat shocked at 3dpf did not display an overt phenotype by ~18 hours post heat shock, with two notable exceptions: induced Cdc42CA and RhoACA embryos displayed widespread tissue disorganization and cardiac defects (data not shown). Such phenotypes are not surprising, as RhoA is required for cardiac development in chick (Kaarbø et al., 2003), and has been shown to be critical for proper heart morphogenesis and contractile function (Phillips et al., 2005; Sah et al., 1999), while cdc42 regulates sarcomere assembly during cardiomyocyte development (Nagai et al., 2003). Because heat shock in these cases occurred late during embryonic development, we reasoned that heat shocking embryos earlier during development will likely lead to more pronounced phenotypes. Therefore, we heat shocked F1 transgenics outcrossed to hsp70l:GAL4, at ~26hpf. At 50hpf, embryos were examined for overt embryonic phenotypes and transgene expression, and then were fixed, sectioned and stained for F-actin and DAPI to examine the structure of the eye. (Figures 5–7).
While no overt phenotype was detected in Cdc42WT embryos, the lens epithelium was thicker and lens fibers appeared mildly disorganized (Figure 5A). Cdc42CA-expressing embryos displayed heart defects and edema, and they were microphthalmic. Cryosectioning revealed severe lens fiber disorganization (Fig. 5B). Despite the lack of gross morphological defects in embryos expressing either of the Cdc42DN isoforms, cryosections revealed severe lens fiber disorganization in both (Fig. 5C,D). RacWT-expressing embryos were microphthalmic but lens formation appeared largely normal (Figure 6A). RacCA-expressing embryos displayed obvious morphological defects, including microphthalmia and cardiac edema, and severe disruption of lens fiber organization (Figure 6B). RacDN-expressing embryos display lower levels of cardiac edema, but sections revealed notable lens fiber disorganization (Figure 6C). Finally, overexpression of RhoAWT resulted in mild lens fiber disorganization (Figure 7A). RhoCA-expressing embryos were microphthalmic and possessed heart defects, mild cardiac edema and severe lens fiber disorganization (Figure 7B). Embryos expressing RhoDN were also microphthalmic and lens fibers were disorganized (Figure 7C).
Cdc42, Rac1 and RhoA are all expressed in the lens (Chen et al., 2006), and have been implicated to play a critical role during lens formation from several studies. For example, knockout of RhoA and Rac1 in the mouse lens disrupts lens development, and defects include disorganization of the lens fibers actin cytoskeleton (Maddala et al., 2011, 2004), and Cdc42 is required for lens pit invagination and early lens development (Chauhan et al., 2009; Muccioli et al., 2016). Thus, the lens phenotypes detailed here provide strong evidence that our transgenes express functional Rho GTPase proteins. However, it is important to note that numerous additional experiments must still be performed to definitively establish the role of Rho GTPases in lens formation during these developmental windows, as it is equally plausible that the defects reported here arise indirectly from pleitropic disruption of general embryonic or ocular development following widespread modulation of GTPase function throughout the embryo and thus do not reflect a direct function of these RhoGTPase proteins during lens development.
These validated transgenic lines represent a versatile toolkit for the temporal-spatial modulation of Cdc42, RhoA and Rac1 activity. To our knowledge, these are the first UAS-inducible transgenic lines for the bicistronic expression of Rho GTPases and a fluorescent reporter. Furthermore, myc tags on Cdc42 and RhoA allow direct determination of protein expression, and potential experiments assaying the altered cellular localization and behaviors of mutant Rho GTPases. However, due to the requirement of Gal4 induction for construct expression and the inherent time delay therein, these transgenic lines are not optimal for studying Rho GTPase function during early developmental processes; their utility lies in modulating Rho GTPase activity during later development, during time points and in tissues that have been inaccessible using previous approaches.
Zebrafish were maintained at 28.5°C on a 14-hour light/10 hour dark cycle. Embryos were obtained from the natural spawning of transgenic or wild-type parents in pairwise crosses. According to established protocols (Westerfield, 1995), embryos were collected and raised at 28.5°C in the dark until they reached appropriate ages for experimentation. hsp70l:gal4kca4 (Scheer et al., 2001) transgenic embryos were obtained from the Zebrafish International Resource Center (ZIRC) and were propagated by outcrosses to AB-strain wild type fish. pou4f3:gal4s311t transgenic (Xiao and Baier, 2007) embryos were provided by Dr. Chris Chang and propagated by outcrosses to AB-strain fish. ptf1a:gal4 transgenic embryos were provided by Dr. Michael Parsons (Johns Hopkins University) and propagated by outcrosses to AB-strain wild-type fish (Parsons et al., 2009). All animals were treated in accordance with provisions established by the University of Texas at Austin and University of Pittsburgh School of Medicine Institutional Animal Care and Use Committees. The following transgenic lines were generated in this study: au66 (uas:mCherry-f2a-myc-Cdc42WT), au67 (uas:mCherry-f2a-myc-Cdc42T17N), au68 (uas:mCherry-f2a-myc-Cdc42CA), au69 (uas:mCherry-f2a-myc-Cdc42F37A), au70 (uas:mcherry-f2a-Rac1WT), au71 (uas:mcherry-f2a-Rac1DN), au72 (uas:mcherry-f2a-Rac1CA), au73 (uas:mCherry-f2a-myc-RhoAWT), au74 (uas:mCherry-f2a-myc-RhoADN), and au75 (uas:mCherry-f2a-myc-RhoACA), and these will be deposited at ZIRC for distribution.
To create pME-mcherry-f2a-cdc42wt (pME-mCWT), a pUCIDT-attL1-mcherry-f2a-myc-cdc42wt-attL2 (pUCIDT-mCWT) oligo was purchased from IDT Gene Synthesis and used to create a pME-mCherry-f2a-Cdc42WT via Gateway cloning (Invitrogen). To build pME-mCherry-f2a-myc-Cdc42CA (pME-mCCA), pME-mCherry-f2a-myc-Cdc42DN(T17N) (pME-mCDN), and pME-mCherry-f2a-myc-Cdc42DN(F37A) (pME-mCF37A), each Cdc42 isoform was PCR amplified from pCS2+cdc42XX plasmids using primers 3 and 4 (Table 1) to create Cdc42xx-attL2 PCR fragments. attL1-mCherry-f2a was PCR amplified from pUCIDT-attL1-mCherry-f2a-Cdc42WT-attL2 using primers 4 and 5 (Table 1) These PCR fragments were then cloned into pME via Gibson Assembly (New England BioLabs).
LR Clonase II Plus was used to carry out all Multisite Gateway assembly reactions using protocols established previously (Kwan et al., 2007). Tg(uas:mCherry-f2a-myc-Cdc42XX) was created using Tol2Kit vectors #302 (p3E-pA), #327 (p5E-UAS), #395 (pDestTol2CG) and pME-mCXX.
To build pME-mCherry-f2a-myc-RhoAWT (pME-mRhWT), pME-mCherry-f2a-myc-RhoACA (pME-mRhCA), and pME-mCherry-f2a-myc-RhoADN (pME-mRhDN), each RhoA isoform was PCR amplified from pCS2+RhoAXX plasmids using primers 5 and 6 (Table 1) to create RhoAxx-attL2 PCR fragments. attL1-mcherry-f2a was PCR amplified from pUCIDT-mCWT using primers 4 and 5 (Table 1) These PCR fragments were then cloned into pME via Gibson Assembly (New England BioLabs).
Tg(uas:mcherry-f2a-myc-RhoA2XX) was created using Tol2Kit vectors #302 (p3E-pA), #327 (p5E-UAS), #395 (pDestTol2CG) and pME-mRhXX.
To create pME-gfp-f2a-Rac1WT (pME-gRWT) and pME-gfp-f2a-Rac1DN (pME-gRDN), Rac1WT and Rac1DN were PCR amplified using Vent polymerase (New England Biolabs) using primers 7 and 8 (Table 1). gfp-f2a was amplified from T2Kactb2:gfp-f2a-creERT2 (Wang et al., 2011), a gift from Dr. Michael Parsons, using primers 9 and 10 (Table 1) and cloned into a 3’ entry vector via InFusion recombination technology (Clontech). Finally, attB1-egfp-2a-Rac1WT-attB2 and attB1-egfp-2a-Rac1DN-attB2 were amplified via Phusion polymerase using primers 11 and 14 (Table 1) and recombined with pDONR221 to create middle entry vectors pME-gRWT and pME-gRDN. To create pME-gfp-f2a-Rac1CA (pME-gRCA), Rac1CA-attB2 was PCR amplified with Phusion polymerase using primers 11 and 12 (Table 1) and ligated to attB1-gfp-f2a (primers 13 and 14 (Table 1)) via overlap PCR to form attB1-gfp-f2a-Rac1XX-attB2 PCR fragments. These fragments were then cloned into middle entry vectors by BP reaction to create pME-gRCA. Tg(uas:mcherry-f2a- Rac1XX) constructs were created using Tol2Kit vectors #302 (p3E-pA), #327 (p5E-UAS), #395 (pDestTol2CG) and pME-gRXX.
All plasmids were sequence confirmed via sequencing on Applied Biosystems 3730 DNA Analyzers at the University of Texas at Austin Institute for Cellular and Molecular Biology DNA Sequencing Facility.
Capped Tol2 mRNA was synthesized from pCS2FA-transposase using the Ambion mMessage mMachine SP6 in vitro transcription kit. Between 50–75pg Tol2 mRNA and between 20–25pg cDNA were microinjected into single-cell embryos. Embryos displaying acceptable levels of mosaic cmlc2:egfp expression were raised to adulthood, and outcrossed to screen for founders. F1 embryos displaying ubiquitous cmlc2:egfp expression were isolated and reared to generate stable lines.
Embryos from hsp70l:gal4 outcrosses of transgenic founders were raised in system water supplemented with Phenylthiourea (PTU). Embryos were individually placed in ~120uL system water in PCR tubes and heat shocked for 30 minutes at 39.5°C in a PCR thermocycler. They were then immediately returned to 28.5°C fish medium for recovery and imaging on either a Leica MZ16F or a Zeiss Axio Zoom V16 fluorescent stereoscope.
Heat shocked and dechorionated embryos were collected at 3dpf. To deyolk embryos, a borosilicate injection needle was used to mechanically disrupt yolks. Embryos were next washed in deyolking buffer without calcium (Link et al., 2006), spun at 300rcf and washed in wash buffer (110mM NaCl, 3.5mM KCl, 2.7mM CaCl2, 10mM Tris/Cl) containing Complete Mini protease inhibitor mixture (Roche Diagnostics). Deyolked embryos were lysed with modified LeMeer’s Lysis Buffer (50mM Tris pH 7.5, 150mM NaCl, 1mM EDTA, 1% IGEPAL, 0.1% sodium deoxycholate) supplemented with Complete Mini protease inhibitor cocktail (Lemeer et al., 2007) before being centrifuged at low speed and sonicated by a Sonic Dismembranator Model 300 (Fisher Scientific). Using manufacturer’s protocol, protein samples were gel electrophoresed using 4–12% Bis-Tris gel and transferred onto PVDF membrane (Invitrogen NuPage system). Blots were incubated 1:5000 anti-myc (abcam ab9106), followed by 1:5000 horse anti-mouse HRP secondary (Cell Signaling Technology 7076). Blots were imaged via the Super Signal West Femto visualization system (Life Sciences) on an ImageQuant LAS4000 machine (GE Life Sciences). Following imaging of myc antibody labeling, blots were stripped for 15 minutes in Restore Western Blot Stripping Buffer (Thermo Scientific 21059) and reprobed with 1:5000 anti-actin (calbiochem cp01).
Embryos were fixed in 4% paraformaldehyde overnight at 4°C, sucrose-protected, and embedded in OCT tissue-freezing medium (TBS, Inc.) before being sectioned at 14µm on a Leica CM1850 crysotat. Sections were rehydrated in 1XPBS for 5 minutes, and blocked in 5% normal goat serum in PBS for 2 hours at room temperature. Sections were stained with 1:500 TOPRO or 1:500 DAPI (Life Technologies) for 9 minutes at room temperature, washed 3× with PBS, and mounted with Vectashield (Vector Laboratories). Images were obtained with a 63× objective on a Leica SP5 confocal microscope. Antibodies used in this study include Rac1 (Milipore, 05-389) and myc (Abcam, ab9106). Phalloidin (Thermo Fisher, A22284) was used at a 1:33 dilution.
We thank members of the Gross lab and Dr. Steve Ekker for helpful comments and suggestions on this work, Ryoko Minowa for fish maintenance and Phil Anselmo and Nick Cave for technical assistance. This work was supported by NIH RO1 EY18005 to JMG, NIH CORE Grant P30 EY08098 to the Department of Ophthalmology at the University of Pittsburgh School of Medicine and a Knights Templar Pediatric Ophthalmology Fellowship to EM. Zebrafish were obtained from ZIRC, which is supported by NIH-NCRR Grant P40 RR012546. We acknowledge additional support from the Eye and Ear Foundation of Pittsburgh and from an unrestricted grant from Research to Prevent Blindness, New York, NY.