Transgenesis of the Wolffian duct visualizes dynamic behavior of cells undergoing tubulogenesis in vivo


  • Yuji Atsuta,

    1. Department of Zoology, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto, Japan
    2. Graduate School of Biological Sciences, Nara Institute of Science and Technology, Ikoma, Nara, Japan
    Search for more papers by this author
  • Ryosuke Tadokoro,

    1. Department of Zoology, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto, Japan
    2. Core Research for Evolutional Science and Technology (CREST), Japan Science and Technology Agency (JST), Chiyoda-ku, Tokyo, Japan
    Search for more papers by this author
  • Daisuke Saito,

    1. Graduate School of Biological Sciences, Nara Institute of Science and Technology, Ikoma, Nara, Japan
    2. Core Research for Evolutional Science and Technology (CREST), Japan Science and Technology Agency (JST), Chiyoda-ku, Tokyo, Japan
    Search for more papers by this author
  • Yoshiko Takahashi

    Corresponding author
    1. Graduate School of Biological Sciences, Nara Institute of Science and Technology, Ikoma, Nara, Japan
    2. Core Research for Evolutional Science and Technology (CREST), Japan Science and Technology Agency (JST), Chiyoda-ku, Tokyo, Japan
    • Department of Zoology, Graduate School of Science, Kyoto University, Sakyo-ku, Kyoto, Japan
    Search for more papers by this author

Author to whom all correspondence should be addressed.



Deciphering how the tubulogenesis is regulated is an essential but unsolved issue in developmental biology. Here, using Wolffian duct (WD) formation in chicken embryos, we have developed a novel method that enables gene manipulation during tubulogenesis in vivo. Exploiting that WD arises from a defined site located anteriorly in the embryo (pronephric region), we targeted this region with the enhanced green fluorescent protein (EGFP) gene by the in ovo electroporation technique. EGFP-positive signals were detected in a wide area of elongating WD, where transgenic cells formed an epithelial component in a mosaic manner. Time-lapse live imaging analyses further revealed dynamic behavior of cells during WD elongation: some cells possessed numerous filopodia, and others exhibited cellular tails that repeated elongation and retraction. The retraction of the tail was precisely regulated by Rho activity via actin dynamics. When electroporated with the C3 gene, encoding Rho inhibitor, WD cells failed to contract their tails, resulting in an aberrantly elongated process. We further combined with the Tol2 transposon-mediated gene transfer technique, and could trace EGFP-positive cells at later stages in the ureteric bud sprouting from WD. This is the first demonstration that exogenous gene(s) can directly be introduced into elongating tubular structures in living amniote embryos. This method has opened a way to investigate how a complex tubulogenesis proceeds in higher vertebrates.


In most organs in the vertebrate body, tubular tissues play central roles in physiological functions, including a delivery of nutrients and wastes, and barrier functions to protect the body from surrounding environment (Andrew & Ewald 2010). Massive tubulogenesis occurs both in embryos and adults, that is, during organogenesis and repairing from injury, respectively. In addition, a failure of tubular integrity would lead to carcinoma (Nguyen et al. 2009), emphasizing that studies of tubular formation are critical in developmental biology, regenerative biology, and cancer biology.

A primary tubular structure is composed of single-layered epithelial cells. Studies with three-dimensional cultured cells and also molecular genetics in flies put forward the model that the onset of tubulogenesis involves several distinct steps including tissue elongation (Shakya et al. 2005; Ewald et al. 2008; Friedl & Gilmour 2009) and cell epithelialization that eventually forms an internal lumen (Ghabrial & Krasnow 2006; Bryant et al. 2010). In the body of amniotes (birds and mammals), however, a complex architecture of tubular tissues has hampered in vivo analyses. Thus, it remains unexplored how individual cells contribute to tubulogenesis.

Here, using chicken embryos, we have established a novel in ovo experimental model that enables gene manipulation of Wolffian duct (WD; also called nephric duct). WD is the earliest basis for kidney formation, which arises exclusively from the anterior intermediate mesoderm, the presumptive pronephric region (Saxen & Sariola 1987). WD elongates caudally as a straight cord, and the cells undergo epithelialization to form a tubular structure. We have achieved gene manipulation of developing WD with the in ovo electroporation technique, where the pronephric region is targeted with EGFP. Resulting WD contains EGFP-positive epithelial cells lining the tubule. Time-lapse live imaging analyses further show cellular dynamics during WD elongation: some cells protrude numerous filopodia, and others repeatedly elongate and retract their tails. The small GTPase Rho is a critical factor to regulate such cellular dynamics, since electroporation with the C3 gene (Rho inhibitor) affected normal behavior of WD cells. Thus, transgenesis of WD in chicken embryos has opened a way to investigate the cellular and molecular mechanisms underlying tubular elongation in amniotes.

Materials and methods

Chicken embryos

Fertilized chicken eggs were commercially obtained from the poultry farm Shiroyama Keien (Sagamihara, Japan), and embryos were staged according to Hamburger Hamilton (Hamburger & Hamilton 1951).


The presumptive pronephric region of 10-somite embryos (10-sm) was labeled with PKH26 red fluorescent cell linker (Sigma) using a micro pipette prepared with the micropipette puller (Narishige).

DNA plasmids

pCAGGS-EGFP, pCAGGS-DsRed, pCAGGS-T2TP, pT2K-CAGGS-EGFP and pCAGGS-tTA were as previously described (Sato et al. 2007; Watanabe et al. 2007). pCAGGS-H2B-EGFP and pCAGGS-H2B-mCherry were gifted from Dr Nakaya (RIKEN), and pCAGGS-GAP43-EGFP from Dr Nakagawa (RIKEN). For pBI-GAP43-EGFP, GAP43-EGFP was subcloned into NotI-HindIII sites of pBI plasmid (Clontech). For pBI-GAP43-tdTomato, GAP43-tdTomato was subcloned into NotI-HindIII sites of pBI plasmid. cDNAs of C3 transferase and a constitutively active form mutant of RhoA (G14V), provided by Dr Kaibuchi (Nagoya University), were subcloned into MluI-NheI sites and MluI site of pBI-GAP43-EGFP, respectively. For pCAGGS-Lifeact-mCherry, Lifeact-mCherry, gifted from Dr Fukuhara (NCVC), was subcloned into MluI-NheI sites of pCAGGS plasmid.

In ovo electroporation

In ovo electroporation was carried out according to the method previously reported (Momose et al. 1999) with slight modifications; Anode and cathode were prepared with a platinum wire (diameter of 0.5 mm) and a sharpened tungsten needle (40 μm diameter at the tip), respectively. DNA plasmids were diluted in EB buffer (QIAGEN) containing 4% fast green FCF (Nacalai) at a final concentration of 4 μg/μL. Platinum electrode was inserted in between the yolk and in parallel to the antero-posterior (A–P) axis of the embryo. Using the micropipette, a DNA solution was administered in between the surface ectoderm and pronephric rudiment. Subsequently, electric pulses were applied six times including a prepulse (See Table 1) (CUY21EX, BEX).

Table 1. Conditions tested for in ovo DNA electroporation into Wolffian duct (WD)
Embryonic stagePre-pulse (V)Main pulsesNo. embryosNo. embryos without abnormality (Viability)No. embryos with EGFP-positive WD (Efficiency)
  1. 6-sm, 10-sm and 14-sm embryos were tested with main electric pulses ranging from 4 to 12 V, either with or without a pre-pulse (30 V or 50 V). See text for details. Assessed 12 h after electroporation.

6-sm 453028 (93%)0 (0%)
 753021 (70%)1 (3%)
 1053018 (60%)3 (10%)
 125308 (27%)3 (10%)
30753017 (57%)6 (20%)
50753017 (57%)15 (50%)
10-sm 453029 (97%)0 (0%)
 753029 (97%)7 (23%)
 1053028 (93%)7 (23%)
 1253023 (77%)11 (37%)
30753028 (93%)13 (43%)
50753027 (90%)22 (73%)
14-sm 453030 (I00%)0 (0%)
 753026 (87%)2 (7%)
 1053027 (90%)6 (20%)
 1253027 (90%)6 (20%)
30753030 (100%)5 (17%)
50753027 (90%)7 (23%)

Primary culture of WD-cells

Wolffian duct cells were dissected with tungsten needles, and placed in a fibronectin-coated glass-bottomed dish containing 10% fetal bovine serum/Dulbecco's modified eagle medium (FBS/DMEM) (Nissui). Cells were cultured for 12 h at 38°C. A solution of membrane-permeable C3 transferase (Cytoskeleton) or Y-27632 (Merck) was added to the culture medium at a final concentration of 0.5 μg/mL or 50 μmol/L, respectively.


Embryos were fixed with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) (137 mmol/L NaCl, 2.7 mmol/L KCl, 10 mmol/L Na2HPO4, 1.8 mmol/L KH2PO4, pH 7.4) overnight at 4°C, followed by preparation of histological sections by a cryostat (10 μm thick). After preblocking with 1% blocking reagent (Roche)/TNT (0.1 mol/L Tris-HCl [pH 7.5], 0.15 mol/L NaCl, 0.1% Tween20) for 1 h at room temperature (RT), the sections were incubated overnight at 4°C with the following primary antibodies: anti-laminin mouse monoclonal antibody (3H11; DSHB), 1:500 dilution; anti-E-cadherin mouse monoclonal antibody (BD Biosciences), 1:500 dilution; anti-Fibronectin mouse monoclonal antibody (Sigma), 1:500 dilution; anti-ZO-1 rabbit polyclonal antibody (Zymed), 1:100 dilution. After washing three times in TNT, the specimens were reacted for 1 h at RT with one of the following secondary antibodies 1:500 diluted in 1% blocking reagent/TNT: Alexa 568 goat anti-mouse IgG, Alexa 568 goat anti-rabbit IgG, Alexa 647 goat anti-mouse IgG (Molecular Probes). The reaction was terminated by washing three times in TNT, and the sections were sealed by FluorSave reagent (Calbiochem) containing 4'6'-diamidino-2-phenylindole dihydrochloride (DAPI) to visualize nuclei. Fluorescent images of the sections were obtained using an Axioplan 2 microscope with Apotome system (Carl Zeiss).

Primary cultured WD cells were fixed with 4% PFA in PBS for 10 min at RT, and treated with PBST (PBS containing 0.5% Triton X-100) for 15 min at RT. To detect filamentous actin, the specimens were subjected to a reaction with Alexa Flour 568 phalloidin (Molecular Probes) diluted 1:50 for 30 min at RT after preblocking with 1% blocking reagent (Roche)/TNT. The reaction was terminated by washing three times in TNT, and the specimens were sealed by FluorSave reagent containing DAPI. For immunostaining of active RhoA, the specimens were incubated with anti-active RhoA mouse monoclonal antibody (NewEast Biosciences) diluted 1:200 for 24 h at 4°C. After washing with TNT, the specimens were reacted for 1 h at RT with Alexa 488 goat anti-mouse IgG (Molecular Probes) diluted 1:500 in 1% blocking reagent/TNT. Fluorescent images were obtained using A1R confocal laser scanning microscope (Nikon) and processed with NIS-Elements software (Nikon).

Whole embryo culture and time-lapse imaging

Whole mount embryo culture was performed as described previously (Chapman et al. 2001) with a minor modification: an electroporated embryo was taken out from an egg and mounted on a filter paper ring. The embryo was placed ventral-side up on a small volume of 1% bacto-agar (DIFCO) prepared in a 35-mm glass-bottomed dish (Greiner). Subsequently, the embryos were overlaid with thin albumen, and cultured in an incubation chamber connected to the LSM5 Pascal confocal laser scanning microscope (Carl Zeiss). Images were obtained and processed with LSM Image Browser software (Carl Zeiss).


WD-specific transgenesis by in ovo electroporation

In embryos of Hamburger and Hamilton stage 10 (HH10; 10 somites), WD emerges in the presumptive pronephric region corresponding to the 6th to 12th somites along the A–P axis (Obara-Ishihara et al. 1999; Hiruma & Nakamura 2003; James & Schultheiss 2003). Subsequently, WD extends posteriorly as a straight cord. We confirmed that fluorescent (PKH26)-labeling of the pronephric region results in a WD-specific labeling as previously reported (Fig. 1A,B) (Obara-Ishihara et al. 1999). Exploiting such particular developmental processes of WD, we performed the in ovo electroporation by targeting the pronephric region with exogenous genes to achieve transgenesis of WD. We injected a DNA solution in between the surface ectoderm and pronephric region (Fig. 1C). Twelve hours after electroporation with pCAGGS-EGFP, an extending WD was specifically positive for EGFP in a way similar to the case with PKH26-labeling (Fig. 1D).

Figure 1.

Transgenesis of Wolffian duct (WD) by in ovo electroporation. (A, B) Dorsal views of an E2 embryo 12 h after the PKH26 labeling of the presumptive pronephric region. A pair of elongating WDs was visualized (red). (C) Diagrams showing the procedures of in ovo DNA electroporation into WD. A DNA solution (pCAGGS-EGFP) shown in orange was injected in between the surface ectoderm and pronephric region corresponding to the 9th to presumptive (*) 12th somites. The photo shows an embryo injected with a plasmid solution containing fast green FCF (arrowheads, also see Material and methods). Electrodes were positioned above and below the embryo. Anode (+), cathode (−). (D) An E2 embryo electroporated at 10-somite stage shows enhanced green fluorescent protein (EGFP) signals in WD (arrows). (E) Electroporation at 14-somite stage yielded lower efficiency.

To optimize the efficiency of transgenesis, 10-sm embryos were subjected to electric pulses with different conditions as described in Table 1. We found that the following condition yielded highest efficiency of EGFP-transgenesis with lowest frequency of malformation of embryos: one pre-pulse 50 μs, 50 V, followed by five pulses of 25 ms, 7 V, with a 250-ms interval between pulses. The electrodes were placed above and below a 10-somite stage embryo, parallel to the A–P axis of the embryo (Table 1; Fig. 1C). When younger (6-somites) or older (14-somites) embryos were electroporated with the same condition of electric pulses, higher frequency of malformation (Table 1) and lower efficiency of transgenesis were observed, respectively (Fig. 1E, Table 1). Thus, in the following experiments, 10-somite embryos were used.

EGFP-positive cells constituted a tubular epithelial component of WD

To scrutinize EGFP-positive cells in WD, transverse histological sections of electroporated embryos of HH14 (E2) were prepared (Fig. 2A). EGFP-positive cells, distributed in a mosaic manner, exhibited epithelial morphology characteristic of WD cells as revealed by epithelial markers, E-cadherin (basolateral) and ZO1 (apical) (Fig. 2B–E; = 5). In addition, EGFP-positive cells were enclosed by the basal extracellular matrices, fibronectin and laminin-1 (Fig. 2F–I; = 4). Thus, the electroporation technique described above was successful to introduce exogenous genes into developing WD.

Figure 2.

Enhanced green fluorescent protein (EGFP)-positive cells participated in the epithelial component of forming Wolffian duct (WD). (A) Drawings of a whole-mount and transverse section of EGFP-electroporated E2 embryo. (B–I) Transverse sections, corresponding to the square in (A), were stained with antibodies for E-cadherin (B, C), ZO-1 (D, E), Fibronectin (F, G) and Laminin-1 (H, I). White arrowheads in (E) show apically localized ZO-1 in the electroporated WD cells. Nuclei were stained with 4'6'-diamidino-2-phenylindole dihydrochloride (DAPI). Scale bars: 20 μm.

Long-term tracing of EGFP-positive WD and its derivatives by the Tol2 transposon-mediated gene transfer

In E3.5 embryos, following the completion of WD elongation, the ureteric bud (UB) sprouts from the posterior end of WD near the presumptive cloaca (Fukui et al. 2009). Subsequently, UB extends anteriorly in parallel with WD. We reasoned that the technique we developed above could be extended to EGFP-labeling of UB cells. To ensure a long-term transgenesis, we used the Tol2 transposon-mediated gene transfer (Sato et al. 2007). Briefly, a gene cassette flanked by Tol2-sequences is excised from an electroporated plasmid, and becomes integrated into the chicken chromosome in the presence of transposase. Thus, the expression of an electroporated gene would be stably retained if driven by an appropriate promoter. We used three plasmids, pT2K-CAGGS-EGFP (pT2K-; a plasmid carrying sequences required for transposition (Tol2-cassette)), pCAGGS-T2TP (T2TP; transposase), and pCAGGS-DsRed as a control for genome non-integrated plasmid (Fig. 3A). The three plasmids were co-electroporated into WD cells, and specimens were harvested at E2 (early elongation of WD) or E6.5 (formation of UB). At E2, signals for both EGFP and DsRed were detected in WD-cells (Fig. 3B–D, = 9). In contrast, at E6.5 when UB formed, EGFP, but not DsRed, was detected in both WD- and UB cells (Fig. 3E–J, = 8). The retained EGFP signals were monitored in whole-mount specimens (Fig. 3H–J), and in histological sections (Fig. 3K–O). EGFP-positive cells were found in mature WD and forming UB, both of which were identified by the kidney tubule marker E-cadherin.

Figure 3.

Long-term transgenesis of Wolffian duct (WD) and its derivatives using the Tol2-mediated gene transfer method. (A) Three plasmids were used for co-electroporation. A gene cassette carried by the pT2K-vector is genomically integrated in the presence of transposase (T2TP). (B–D) The three plasmids were co-electroporated into E1.5 (10-sm) embryos and examined at E2. Both EGFP (integrated) and DsRed (non-integrated) were detected in WD-cells. (E–O) At E6.5, EGFP but not DsRed signals were observed in formed WD and its derived ureteric bud (UB). (E–J) Whole mount views of urogenital tissues. (J) Magnification of the square in (H). (K–O) A transverse view at the white line in (J). Both WD cells (L, M) and UB cells (N, O) were marked by immunostaining for E-cadherin. Enhanced green fluorescent protein (EGFP)-positive cells are indicated by white arrowheads. MeTu, mesonephric tubules. Scale bars: 20 μm for (B) and 50 μm for (L, N).

Different populations of WD cells are separately manipulated with EGFP and mCherry by successive electroporations

We previously noticed that co-electroporation with two different DNA plasmids (i.e. EGFP and mCherry) often resulted in co-expression in a cell (Nakaya et al. 2004; Sato et al. 2008; Saito et al. 2012). In contrast, successive electroporations with a single gene for each yields a mixture of EGFP single positive- and mCherry single positive cells with little co-expression. This notion is in agreement with the previous report, where presomitic mesoderm was manipulated by successive electroporations (Iimura & Pourquié 2006).

We therefore tested if successive electroporations would also be applicable to the WD transgenesis, using pCAGGS-H2B-EGFP and pCAGGS-H2B-mCherry. A fusion with the H2B histone-encoding gene leads to a nuclear localization, facilitating the unambiguous distinguishing of signals between single-positive and double-positive. When these plasmids were co-electroporated, 74.8% of the electroporated cells yielded co-expression of EGFP and mCherry signals in WD cells (Fig. 4A–E,K). In contrast, when these two plasmids were successively electroporated, most of the electroporated cells were either EGFP- or mCherry-positive (Fig. 4F–J,K). Thus, the technique of successive electroporations is applicable to WD transgenesis.

Figure 4.

Successive electroporations of Wolffian duct (WD)-cells. (A–E) Single electroporation, where H2B-EGFP and H2B-mCherry were co-electroporated. (C–E) Double-positive cells for enhanced green fluorescent protein (EGFP)and mCherry are indicated by arrows. (F–J) Successive electroporations, where the first shot with H2B-EGFP was immediately followed by the second one with H2B-mCherry. (H–J) Most electroporated cells were single-positive for either EGFP or mCherry. (K) Quantitative representation of the number of double-positive cells per total number of electroporated cells. Scale bars: 50 μm for (B) and (G), 20 μm for (C) and (H).

Live-imaging analyses to visualize the behavior of WD cells

Enhanced green fluorescent protein-labeling of WD further enabled us to directly visualize the behavior of individual cells during WD elongation in vivo. Using a modified New culture method (Chapman et al. 2001), an EGFP-electroporated E2 embryo was placed in a glass-bottomed dish, followed by time lapse confocal microscopy (Fig. 5A). We observed that WD cells migrated caudally in a straight line, although a few cells sporadically deviated from the main stream (Fig. 5B, = 7, Movie S1). To visualize cellular processes with a higher resolution, cDNA of GAP43-EGFP, the membrane bound form of EGFP, was electroporated. Notably, front cells dynamically changed their shapes by extending lamellipodia and filopodia (Fig. 5C, = 3, Movie S2). This is the first demonstration of live-imaged behavior of individual cells during tubular elongation in living amniotes.

Figure 5.

In vivo time-lapse imaging of migrating Wolffian duct (WD)-cells. (A) A 20-sm embryo electroporated with enhanced green fluorescent protein (EGFP) or GAP-EGFP was placed in a glass-bottomed dish, and subjected to time-lapse imaging analyses by confocal microscopy. (B, C) Selected frames from the time-lapse movies (Movie S1, Movie S2). (B) Low magnification. EGFP-positive cells migrated caudally in register with somite segmentation. Yellow arrowhead indicates the level of a newly formed somite. (C) Selected frames of the time-lapse movie (Movie S2), showing magnified WD cells electroporated with GAP-EGFP. Lamellipodia and filopodia are observed (white arrows). Scale bars: 100 μm for (B), 20 μm for (C).

Rho activity is essential for tail retraction of migrating WD

We also noticed a peculiar behavior of cells located slightly behind the front cells. As the cell body moved forward, its tail elongated (~15 min) and subsequently retracted (~36 min). Thus, the tail repeated elongation and retraction independently of the position of cell body (Fig. 6A, E; = 20 cells, see also Movie S3). The tail retraction of migrating cells has previously been reported in vitro for monocytes (Worthylake et al. 2001) and endothelial cells (Lamalice et al. 2007). In these in vitro studies, the RhoA-ROCK pathway was shown to be critical by mediating the generation of traction force via actomyosin contractility (Bito et al. 2000; Yan-Ting et al. 2004; Beningo et al. 2006).

Figure 6.

Retraction of cellular tail and its regulation by RhoA. (A) Selected frames from the time-lapse movie (Movie S3). White and yellow bars indicate the longitudinal length of migrating Wolffian duct (WD) cells. (B) The plasmid used to inhibit RhoA activity. Tetracycline response element (TRE) bidirectionally activates C3 transferase and GAP-enhanced green fluorescent protein (EGFP) genes. The graph shows a quantitative representation of the longitudinal length of electroporated WD cells, measured by Image J software. *< 0.01. (C) Successive electroporations with TRE-C3-GAP-EGFP and TRE-GAP-Tomato (control). (D) In control, the length of EGFP-positive and Tomato-positive cells were comparable. (E) A C3-electroporated cell (EGFP-positive) displayed an aberrantly elongated tail compared to the neighboring Tomato-positive control cells. (F) Selected frames from the time-lapse movie (Movie S4). The aberrantly elongated tail of C3-electroproated cell became fragmented. Scale bars: 50 μm for (A, D–F).

We therefore tested if Rho/ROCK-mediated tail retraction would also occur in migrating WD cells, by electroporating either C3 transferase (Rho inhibitor) or constitutively active form of RhoA (CA-RhoA). For this purpose, pBI-TRE vector was used. TRE (tetracycline response element), a bidirectional promoter activated by tTA (the tetracycline-dependent transcriptional activator), allows simultaneous expression of two different genes (Fig. 6B) (Watanabe et al. 2007). When electroporated with pBI-TRE-GAP43-EGFP as a control, WD cells were 43 μm long along the antero-posterior axis (Fig. 6B). In contrast, cells expressing C3 transferase were 88 μm long, with an aberrantly elongated tail (Fig. 6B). Overexpression of CA-RhoA yielded little effects on cell shape changes (Fig. 6B).

The effects elicited by Rho-inhibition on cell shape changes were unambiguously detected by directly comparing between EGFP (control)- and C3-transfected cells that were adjacent to each other in a forming WD. This analysis was achieved by the successive electroporations as shown in Figure 4. Thus, the first electroporation with the Rho-inhibition plasmid (pBI-TRE-GAP43-EGFP-C3) was immediately followed by the second one with the control (pBI-TRE-GAP43-tdTomato) (Fig. 6C). Whereas the neighboring EGFP+ or Tomato+ cells showed similar size in length in control experiments (Fig. 6D), a RhoA-inhibited EGFP+ cell was much longer compared to the adjacent control Tomato+ cell (Fig. 6E). These results suggest that RhoA activity is required for the tail retraction of migrating cells during WD elongation. Time-lapse imaging analyses further showed that the aberrantly elongated tail of RhoA-inhibited cells often became fragmented, resulting in a degradation of the anucleated piece (Fig. 6F, Movie S4). We observed no detectable effects by the C3 overexpression on the behavior of front cells (data not shown).

Stress fibers in migrating WD cells in vivo

To get insight into the cytoskeleton dynamics during the Rho-mediated tail retraction, we analyzed actin dynamics in WD cells. A piece of WD was dissected from 20-sm embryos, and subjected to a primary culture in vitro (Fig. 7A). In these cells, F-actin, visualized by phalloidin, constituted prominent stress fibers (Fig. 7B). When treated by C3 transferase or Y27632 (ROCK inhibitor), the stress fibers were extinguished (Fig. 7C,D). We also observed that activated (GTP-bound) form of RhoA, detected by immunostaining, was broadly distributed in a normal WD cell with slightly higher signals around the nuclei (Fig. 7E).

Figure 7.

Stress fibers in migrating Wolffian duct (WD)cells and their regulation by Rho/ROCK activities. (A) A piece of normal WD was dissected from 20-sm embryos, and subjected to 2-dimensional in vitro culture. (B–D) Cells were treated by indicated drugs, and stained with Alexa568-labeled phalloidin. (E) Immunostaining with anti GTP-bound RhoA antibody. (F–H) Lifeact-mCherry was co-electroporated with GAP-enhanced green fluorescent protein (EGFP) into embryos. Stress fibers were formed prominently in the rear portion of the cell (white arrowheads). (I–K) Lifeact-mCherry was co-electroporated with C3-GAP-EGFP. The tail portion failed to display stress fibers (black arrowheads). Scale bars: 10 μm.

Last, we explored actin dynamics in the cells behind front cells directly in vivo during WD elongation. F-actin was visualized by electroporating the Lifeact-mCherry gene along with GAP-EGFP (Riedl et al. 2008). Prominent stress fiber structures were observed, and many of them ran in parallel to the elongation axis of the cell (Fig. 7F–H, = 12 cells). These stress fibers failed to form when co-electroproated with C3 transferase (Fig. 7I–K, = 14 cells), suggesting that RhoA activity is required for the stress fiber formation during WD formation in vivo, consistent with in vitro studies. The RhoA-mediated stress fiber formation may generate the traction force in the cellular tail.


The WD formation, particularly in chicken embryos, has long been an excellent model in embryology, where cell lineages and inductive events were extensively studied. To decipher the regulation of tubulogenesis, it is necessary to know how individual cells behave and how such behaviors are governed by the genetic program. Individual cells during tubular morphogenesis have successfully been live-imaged in the trachea of Drosophila embryos (Chihara et al. 2003) and in blood vessels of zebrafish embryos (Siekmann & Lawson 2007). However, such live-imaging analyses have been limited in living amniote embryos.

In this study, using chicken embryos, we have succeeded in live-imaging of WD cells by developing the technique of WD-specific transgenesis. Exploiting the particular mode of WD development, that is, WD emerges in a defined area (6th to 12th somite level), EGFP cDNA is electroporated into this region. This manipulation results in a specific transgenesis of WD, where EGFP-positive cells constitute a tubular structure with the correct apico-basal polarity.

High resolution analyses have revealed that the front cells of WD actively extend and retract filopodia and lamellipodia, the phenomena consistent with the behaviors known for the tip cells in forming blood vessels in vertebrates and the trachea in Drosophila (Chihara et al. 2003; Siekmann & Lawson 2007). In addition, migrating WD cells frequently display an elongated rear process/tail, which subsequently retracts. This event is regulated by Rho activity, since Rho-inhibited cells fail to retract the tail, leaving an abnormally elongated process behind. Thus, the tail dynamics need to be precisely regulated, and Rho is a critical regulator. The tail retraction of individual cells appears to be important for the WD morphogenesis, since in ovo administration of a C3 protein solution resulted in an aberrant WD formation (not shown). This study is the first demonstration of cellular tail dynamics observed directly in living embryos, although such phenomena have been known with cultured cells in vitro (Worthylake et al. 2001; Sastry et al. 2006).

Electroporation with the Lifeact-mCherry gene has revealed stress fibers running in parallel to the elongation axis of the cell. Although stress fibers have extensively been studied using culture cells, a direct visualization of these structures in living bodies has just begun. One recent report showed stress fibers in tip cells of forming blood vessels in mouse retina (Weed et al. 2012). Together with our findings, it is conceivable that in living embryos, stress fibers form where a traction force is generated to assume a particular cell shape such as elongation along the migratory path.

Last, we have described two more useful techniques. One is a combination of the in ovo electroporation with the Tol2-transposon-mediated gene transfer technique (Sato et al. 2007), which enables long-term tracing of EGFP-positive WD cells until later stages when ureteric bud forms. Another useful technique is the successive electroporation method allowing differential transfection with multiple genes into neighboring cells.

The methods described in this study open a way to study the molecular mechanisms underlying tubulogenesis in amniotes. The methods must be particularly useful to understand how individual cells contribute to morphogenesis of three-dimensional tubular structures, an essential but unsolved question in developmental biology. Knowledge provided by WD studies could be applicable to understanding more complex tubulogenesis, that is, metanephric tubules, trachea, blood vessels, and mammary glands.


This work was supported by a Grant-in-Aid for Scientific Research on Innovative Areas, and Grant-in-Aid for Scientific Research (A) from the Ministry of Education, Culture, Sports, Science and Technology of Japan, CREST (JST), The Mitsubishi Foundation, and Takeda Science Foundation. Y.A. is a fellow of the Japan Society for the Promotion of Science. We thank Drs Kaibuchi, Nakagawa, Nakaya and Fukuhara for providing useful materials.