Fixation/Permeabilization: New Alternative Procedure for Immunofluorescence and mRNA In Situ Hybridization of Vertebrate and Invertebrate Embryos


Correspondence to: Juan Fernández, Laboratory of Developmental Cell Biology, Department of Biology, Faculty of Sciences, University of Chile, Casilla 653, Santiago, Chile. E-mail:


A new procedure is described to visualize the spatial pattern of expression of proteins and mRNAs in cryosections or whole-mounted leech, Drosophila, zebrafish, and chick embryos. Our principal contribution is in the use of a nonconventional fixation/permeabilization procedure based on the use of formaldehyde or paraformaldehyde combined with a short C-chain carboxylic acid. Detergents, methanol, and proteinases were omitted. Hybridization procedures were modified from those of routinely used protocols developed for the same embryos. Results showed that cytoskeletal and other cytoplasmic proteins, as well as different mRNAs, were clearly visualized in the expected regions of the embryos. Our procedure has several advantages over currently used protocols: is simpler, produces better general preservation of cells, yields reliable results, and can be used for embryos of different taxa at different developmental stages. It is hypothesized that short C-chain aliphatic carboxylic acids modulate the cross-linking effect of aldehyde fixatives on cell proteins. Developmental Dynamics 242:493–507, 2013. © 2013 Wiley Periodicals, Inc.


Immunofluorescence and RNA in situ hybridization are microscopy techniques that are extensively used for the localization of antigens and mRNAs with single cell resolution. Both techniques include a preparatory fixation and permeabilization steps. Fixation is a procedure intended to stabilize the cell structure, preserving it as closely as possible to that of the living cell. It may be achieved by precipitating or additive fixatives. Precipitating fixatives, such as cold organic solvents (methanol, acetone, ethanol, or mixtures of them), denature proteins and hence disrupt their three-dimensional (3D) structure. In addition, they provoke dehydration and cell shrinkage. Under these conditions, cell structure is poorly preserved. However, antigenic sites and mRNAs are not covalently modified and, thus, may be available to bind antibodies and mRNA probes, respectively. Additive fixatives, such as formaldehyde and glutaraldehyde, cross-link proteins preserving their native 3D structure (Hayat, 2002). Formaldehyde is a gas that dissolves in water to 37–40%, and reacts with it to form methanediol (Kiernan, 2000). As polymerization of this compound increases, its fixation properties decrease. For this reason, monomeric formaldehyde obtained from paraformaldehyde is commonly used to prepare cells for immunocytochemical procedures. Aldehyde-driven cross-linking is achieved by formation of methylene bridges between amino acids (Puchtler and Meloan, 1985; Kiernan, 2000; Sutherland et al., 2008) or between amino acids and nucleic acids (Dedon et al., 1991; Orlando et al., 1997; Spencer and Davie, 2002; Schmiedeberg et al., 2009). Formation of methylene bridges by formaldehyde appears to occur in two steps. The first step involves the formation of highly reactive methylol groups by the reaction of methylene hydrate with: (a) reactive groups, mainly NH2 groups of N-terminal amino acid residues and the side chains of arginine, cysteine, histidine, and lysine, and (b) cytosine amino groups of nucleic acids. During the second step, the reaction with an another amino group leads to a condensation reaction with the formation of methylene bridges between adjacent polypeptide chains. Additive fixatives covalently modify proteins and mRNAs that become less able to bind antibodies and mRNA probes. To solve this problem, cell and tissues must be permeabilized before, during or after fixation. Permeabilization is a process by which membranes are solubilized by detergents and/or enzymes allowing antibodies and RNA probes to reach the antigenic sites and native mRNAs, respectively (Puchtler and Meloan, 1985; Orlando et al., 1997; Kiernan, 2000; Hayat, 2002; Spencer and Davie, 2002; Renshaw, 2007). In the case of immunofluorescence, permeabilization and fixation are followed by incubation in the primary antibody and then its recognition by a secondary antibody linked to a fluorophore. For mRNA in situ hybridization, fixed and permeabilized samples must first be hybridized with a labeled RNA probe after which the hybridization product is visualized by immunofluorescence or immunocytochemistry. This last process is called staining.

Preparation of cells or tissues for immunofluorescence and mRNA in situ hybridization thus involve three main steps: fixation, permeabilization and staining. Because the first two steps may be performed in a different sequence, separately or together, one may consider three modalities for applying these techniques: (a) Permeabilization, fixation, and staining. This is the less common modality but one that has successfully been used for the immunofluorescence staining of the cytoskeleton of leech embryos (Fernández et al., 1990, 1994; Fernández and Olea, 1995). In this case, permeabilization is performed first using an extraction buffer including a nonionic detergent, protein stabilizers and antiproteases (Fernández et al., 1998; Cantillana et al., 2000). Fixation is performed afterward using paraformaldehyde-Triton-X-100. This detergent is also added to the antibody staining and washing solutions. (b) Fixation, permeabilization, and staining. This is the most common procedure for immunofluorescence and mRNA in situ hybridization in which samples are first fixed with paraformaldehyde (Schulte-Merker et al., 1992; Strähle and Jesuthasan, 1993; Strähle et al., 1993; Jesuthasan, 1998; Thisse and Thisse, 2008) or formaldehyde (Henrique et al., 1995; Streit and Stern, 2001; Kang et al., 2002). Small amounts of detergent, glutaraldehyde, calcium chelators, Mg salts or cytoskeleton stabilizers may be added to either of the fixatives (Hammati-Brivanlou and Harland, 1989; Harland, 1991; Stachel et al., 1993; Henrique et al., 1995; Leung et al., 2000; Streit and Stern, 2001; Dekens et al., 2003). In the majority of cases, permeabilization is achieved by treating the fixed embryos with detergents, cold organic solvents or a mixture of both (Hammati-Brivanlou and Harland, 1989; Harland, 1991; Schulte-Merker et al., 1992; Stachel et al., 1993; Leung et al., 2000; Thisse and Thisse, 2008). For mRNA in situ hybridization, proteinase K or pronase E are used for further permeabilization (Allende et al., 1996; Kang et al., 2002). After permeabilization, cells or tissues are subjected to immunostaining or mRNA in situ hybridization. Antigen retrieval after fixation may be also achieved by heat (Shi et al., 2001; Hayat, 2002; Renshaw, 2007) and this procedure is compatible with mRNA in situ hybridization and detection of fluorescent proteins in transgenic fish lines (Inoue and Wittbrodt, 2011). (c) Fixation/permeabilization and staining. In this case, fixation/permeabilization is carried out simultaneously and can be achieved by using cold organic solvents such as methanol, that fix proteins while keeping the antigenic sites of proteins and the nucleic acids available to bind antibodies and hybridization probes, respectively. To further improve accessibility to antigenic sites, detergents, calcium chelators, or proteolytic enzymes may be also added to the fixative or during the washing and staining of the samples (Wühr et al., 2008). In other cases, a detergent is added to the aldehyde fixative that was prepared with stabilizers or in an assembly buffer (Gard, 1991; Schroeder and Gard, 1992; Pelegri et al., 1999).

In this study, we show that combined fixation/permeabilization, using a mixture of formaldehyde or paraformaldehyde with a short C-chain aliphatic carboxylic acid (particularly glacial acetic acid), that acts as a permeabilizer, allows improved preparation of embryonic cells and tissues for immunofluorescence and mRNA in situ hybridization. Detergents, organic solvents, proteinases, and refixation steps were omitted. With this tool, we developed protocols for immunostaining and mRNA in situ hybridization of different vertebrate and invertebrate embryos. The advantage of using these protocols is discussed.


Immunofluorescence Staining of the Leech Embryo

Good immunofluorescence staining of cytoskeletal structures was achieved by standard fixation/permeabilization, a situation that improved when microtubule and/or microfilament stabilizers were used. A low percentage of useful preparations was obtained with propionic or isobutyric acids. Swelling and disaggregation of the embryonic cells by the acetic acid was diminished by combining acetic and isobutyric acids: 3–4 and 1–2 droplets, respectively. Under these conditions, the immunofluorescence signals were less intense than when using acetic acid only, but the integrity of the embryos was much better preserved. Time of incubation in the antibodies could be drastically reduced: 3–6 hr for the primary antibody, 2 hr for the secondary antibody. DAPI (4′,6-diamidino-2-phenylidole-dihydrochloride) staining was successful after 10–15 min incubation. Our new procedure led to a very good preservation of the cytoplasm, improved staining of deep cytoskeletal structures and minor distortion of the embryonic cells.


The fixation/permeabilization procedure nicely revealed both peripheral and deep structures such as the sperm nucleus and aster, the meiotic spindles, the perinuclear plasm, the polar rings, and the meridional bands (Fig. 1A–F). The latter two structures condense to form the polar cytoplasmic domains, or teloplasms (Fernández et al., 1998). For comparison, control zygotes were prepared by a permeabilization fixation procedure according to Fernández and Olea (1995). In both experimental and controls zygotes, cytoskeleton stabilizers were used. Comparison of Figures 1E and 1F reveal a much better preservation of the zygote structure, for example of the microtubules, using our new immunofluorescence procedure.

Figure 1.

Immunofluorescence staining of whole-mounted leech eggs (A,B) and zygotes (C–F). A: Microtubule-stained first meiotic spindle (sp) and sperm aster (as) [Form-Acet]. B: Double-stained (microtubules green, pronucleus red) male pronucleus (pr) with the aster (as) [Form-Acet]. C,D: Co-distribution of microtubules (C) and actin filaments (D) in the animal (ar) and vegetal (vr) rings and the perinuclear plasm (pp) [Form-Acet]. E: Accumulation of microtubules in the animal pole region (arrowhead), animal ring (ar) and meridional bands (mb) [Form-Acet-Tx-Ph]. F: Control zygote stained for microtubules according to Fernández and Olea (1995). Notice that the zygote periphery is not well preserved and microtubules appear shorter and less numerous. The bundle of microtubules at the animal pole was not disclosed. ar, animal ring; mb, meridional bands. Scale bar = 90 μm in A, 12.5 μm in B, 145 μm in C,D, 65 μm in E,F.


During the first asymmetric cleavage division, the huge mitotic spindle appeared partly sandwiched between the cytoskeleton-rich teloplasms accumulated in the CD blastomere. The teloplasms included microtubules and actin filaments that formed two domains. Blastulae showed that the A, B, C blastomeres inherited the perinuclear plasm domain enclosing the nucleus and centrosome. Meanwhile, the D cell-derived blastomeres, or proteloblasts, inherited perinuclear plasm domain embedded in the teloplasms (Fernández, 1980). The teloplasms were finally sequestered by stem cells or teloblasts (M, N, O, P, and Q cells) that used the teloplasm to generate rows of blast cells (bandlets) that originate the germinal bands (Fig. 2A–D) (Fernández and Olea, 1982; Weisblat et al., 1984).

Figure 2.

Immunofluorescence staining of whole mounted leech embryos. A: Double-stained (microtubules green, actin red) first cleavage division embryo showing the nascent AB and CD blastomeres and the mitotic spindle (sp). Astral microtubules (as) are seen leaving the AB cell centrosome (ce). The animal (top) and vegetal (bottom) teloplasms (te), consist of an inner actin-rich and an outer microtubule-rich sectors. [Form-Acet-Tx-Ph]. B: The O and P blastomeres have been just formed and microtubules are seen accumulated in the subdivided teloplasm (te) and division furrow (arrowhead) [Form-Prop]. C: Double-stained (microtubules green, chromosomes red) prophase nucleus in the syncytial B blastomere. ce, centrosomes with astral fibers [Form-Acet]. D: Double- stained (microtubules green, nuclei red) gastrula showing the O and P teloblasts and the bandlets of blast cells (arrows). ge, germinal band; mt, bundles of microtubules; te, teloplasm [Form-Acet-Isobut]. Scale bar = 35 μm in A,B, 9 μm in C, 15 μm in D.

Immunofluorescence Staining of the Drosophila Embryo

Good immunostaining of the Drosophila cytoskeleton was obtained after a short fixation/permeabilization with glacial acetic or propionic acid. Isobutyric acid gave poorer results. In some cases, cytoskeleton stabilizers did not produce detectable differences in the immunostaining of the embryo. One-hour incubation in the primary and secondary antibodies and 15 min in DAPI or TO-PRO-3 may be sufficient for optimum staining. The quality of our immunofluorescence preparations was equivalent to that reached by conventional methods (Karr and Alberts, 1986; Sharp et al., 1999). The resolution of the images was remarkably improved after deconvolution of either fluorescent or confocal images.


Our immunostaining protocol nicely revealed the organization of the cytoskeleton in the syncytial blastula (Fig. 3A–D). The structure of the mitotic spindle, and of its surrounding cytoplasm (energid), during different phases of the mitosis was similar to that described by Karr and Alberts (1986). Immunofluorescent controls prepared according to the latter authors showed similarly preserved mitotic spindles.

Figure 3.

Immunofluorescence staining of whole-mounted Drosophila syncytial blastulae. A: Low magnification image showing numerous microtubule-stained (green) cleavage spindles (sp) [Form-Acet]. B: Double stained (microtubules green, chromosomes red) metaphase cleavage spindles. en, energids cytoplasm [Form-Acet]. C: Higher magnification deconvolved image showing microtubule-stained mitotic spindles. The spindle poles and asters (as) are readily visualized [Form-Acet]. D: Control similarly stained cleavage spindles prepared according to Karr and Alberts (1986). Scale bar = 25 μm in A, 12 μm in B, 10 μm in C, 5 μm in D.


In the gastrula stage embryos microtubules were seen accumulated at the periphery of the ectodermal and mesodermal cells. In the germ band embryos, our procedure succeeded in showing the organization of microtubules in the body segments and in particular in the nerve cord (Fig. 4A–D).

Figure 4.

Immunofluorescence staining of whole-mounted Drosophila embryos [nForm-Acet]. A,B: Confocal projection (A) and optical section (7/13) (B) of an embryo stained for microtubules (green) and nuclei (red) showing the boundaries (arrowheads) between the segments of the retracted germinal band. Notice accumulation of microtubules in cells of the nerve cord (nc), walls of the foregut (fg) and hindgut (hg). C: Confocal image showing a dorsal view of a similar embryo also stained for microtubules (green) and nuclei (red). Notice the ladder-like appearance of the ventral nerve cord (nc). D: Higher magnification of the nerve cord to show the distribution of neuroblast nuclei and the organization of neural processes in the connectives (cn), commissures (cm), and nerves (ne). Scale bar = 40 μm in A–C, 9 μm in D.

Immunofluorescence Staining of the Zebrafish Embryo and Larva

The best results were obtained after standard fixation/permeabilization using neutralized formaldehyde. The zygote blastodisc is hard to permeabilize and its fixation/permeabilization requires a greater amount of glacial acetic acid (up to 16%). Under these conditions, good staining of the pronuclei and first cleavage spindle was possible. Polyethylenglycol, MgSO4, dimethyl sulfoxide (DMSO) and Taxol, added to the fixative separately or in combination, improved the preservation of the embryos and the quality of the immunofluorescence. This was particularly noticed in the staining of the mitotic spindles. Propionic and isobutyric acids can also be used but yield less consistent results. Visualization of mitotic spindles was also improved when tyrosinated antitubulin primary antibody was used. It was found that 6 hr in the primary antibody, 2 hr in the secondary antibody and 10 min in DAPI or TO-PRO-3 may be sufficient to suitably stain microtubules, actin filaments, and nuclei.


Immunofluorescence staining of the cytoskeleton provided very satisfactory visualization of the first mitotic spindle and of microtubules and actin filaments across the blastodisc and endoplasmic lacunae. Blastulae were examined after staining for microtubules, actin filaments, γ-tubulin, β-catenin, and DNA. Mitotic cells appeared as good as those prepared using other immunofluorescence procedures (Fig. 5A–H) (Dekens et al., 2003; Kishimoto et al., 2004; Pfaff et al., 2007; Yesbe et al., 2007). The same happened with microtubules associated with the cleavage furrows (Jesuthasan, 1998; Lee et al., 2004). γ-Tubulin was detected in the centrosomes and spindle poles, whereas β-catenin was seen at the periphery of the cleaving blastomeres (Knaut et al., 2000; Pelegri, 2003; Yesbe et al., 2007). For comparison, blastulae were also prepared by a fixation/permeabilization procedure according to Solnica-Krezel and Driever (1994). Comparison of Figures 5C and D shows that our fixation/permeabilization procedure produced results similar to those obtained by the latter authors. Addition of microtubule stabilizers, and in some cases DMSO, to our acid fixative allowed preservation and staining of more and larger astral fibers (compare Figs. 5E,F with 5G,H). Taken into consideration the fact that the extent of the aster in zebrafish mitotic spindles depends on the mitotic stage, being larger in anaphase/telophase stages (Wühr et al, 2010), the anaphase of Figure 5F (fixed/permeabilized without Taxol) has a much smaller aster than those of the mitotic spindles of Figure 5G,H (fixed/permeabilized with Taxol).

Figure 5.

Immunofluorescence staining of zebrafish whole-mounted embryos. A: Double stained (microtubules green, chromosomes red) metaphase first cleavage spindle (sp). The arrow points to the DNA left by the extruded second pole cell. bd, blastodisc [Form-Acet-Peg]. B: Similarly stained blastula with numerous methaphase mitotic spindles (sp) [Form-Acet]. C,D: Confocal images of advanced blastulae showing asynchronously dividing blastodermal cells (microtubules, green; nuclei, blue) prepared with our method [Form-Acet] and stained for β-tubulin (C) or according to Solnica-Krezel and Driever (1994) (D). E,F: High magnification double-stained (microtubules, green; chromosomes, blue) metaphase (E) and anaphase (F) mitotic spindles stained for tyrosinated tubulin [Form-Acet-Peg]. G,H: Confocal images (microtubules, green; chromosomes, red) of anaphase mitotic spindles stained for tyrosinated (G) and α-tubulin (H). The latter image was deconvolved. Notice the successful staining of peripheral and astral microtubules [nForm-Acet-DMSO-Tx]. Scale bar = 30 μm in A, 55 μm in B, 25 μm in C,D, 3 μm in E,F, 10 μm in G,H.


Double or triple immunostained frozen sections revealed clearly the organization of the cytoskeleton and the distribution of other proteins, such as sarcomeric and nonmuscle myosin and β-catenin, in the neural tube, somites, and eye (Fig. 6A–J). With the latter organ, we obtained excellent staining of the retina layers and lens. In general, results showed the high degree of structural preservation of the embryonic tissue achieved by the application of our method. Examination of 5- to 6-day postfertilization whole-mounted larvae immunostained for microtubules allowed visualization of the peripheral innervation of the head, fins, somites, olfactory placodes, spinal nerves, ganglia and developing neurons.

Figure 6.

Immunofluorescent staining of freeze-sectioned 3-day-old (A–H) and whole-mounted 6-day-old (I,J) zebrafish larvae. A–C: Microtubule-stained (A), actin-stained (B), and merged image (C) of a triple stained (DAPI) larva. Microtubules and actin filaments co-distribute in the outer sector of the neural rod (nr) and developing musculature (mu). ep, epidermis; no, empty notochord [Form-Acet]. D: Double-stained (microtubules, green; nuclei, red) merge image of the neural rod showing its inner nuclear (in) and outer fibrillar (of) sectors [Form-Acet]. E: Confocal deconvolved double-stained image (microtubules green, nuclei blue) of a longitudinally sectioned neural rod [Form-Acet]. F: Triple-stained (microtubules, green; β-catenin, red; nuclei, blue) merged image of a coronal section along the head of a larva showing the eye (ey), optic nerve (ov), optic chiasm (oc), and tectal fiber tracts (tf). The central lens (le) cells express β-catenin and microtubules [Form-Acet]. G: Confocal deconvolved merge image of a double-stained (microtubules, green; nuclei, blue) eye showing the retina layers: pt, photoreceptors; on/op, outer nuclear and plexiform layers; in/ip, inner nuclear and plexiform layers; gc, ganglion cell layer; le, lens [Form-Acet]. H: Sector of the same retina showing the inner nuclear (in) and plexiform (ip) layers. I: Spinal cord (sc), nerve roots (nv), and sensory ganglia (sg) stained for microtubules. J: Optical section of a similarly stained spinal cord with developing neurons (dn) [Form-Acet]. Scale bar = 25 μm in A–C, 10 μm in D,E, 35 μm in F, 15 μm in G, 6 μm in H, 25 μm in I,J.

Immunofluorescence Staining of the Chick Embryo

The best results were obtained after fixation/permeabilization with glacial acetic acid of sliced embryos. Isobutyric acid yielded good results but in a smaller number of preparations (approximately 20%). There was no need for protein stabilizers Improved visualization of labeled structures was achieved in frozen sections of embryonic slices incubated overnight in the primary and then secondary antibodies. Shorter incubations were not performed. Although reasonable immunostaining was possible with whole-mounted embryos, results were much better when performed on freeze-sectioned embryos.


Transverse sections of immunostained embryos clearly show the distribution of microtubules, actin filaments, β-catenin, myosin and nuclei in the developing neural tube, notochord, somites, myotomes, blood vessels, pronephros, and the retina (Fig. 7A–D). As shown in Figure 7A, our procedure combines good preservation of large pieces of embryonic tissue with neat immunofluorescence signals.

Figure 7.

Immunofluorescence staining of 3-day-old freeze-sectioned chick embryos. A: Triple-stained (microtubules, green; β-catenin, red; nuclei, blue) merged image of a low magnification transversally sectioned embryo. The walls of the aorta (ao), anterior cardinal veins (cv), notochord (no), and pharynx (ph) have abundant microtubules. The ectoderm (ec), neural tube (nt), ventral roots (vt), somites (so), and pronephritic ducts (pd) reveal co-distribution of microtubules and β-catenin. me, mesenchyme [Form-Acet]. B: Higher magnification of the neural tube in panel A [Form-Acet]. C: Higher magnification of the somite in panel A. ec, ectoderm; de, dermatome; my, myotome [Form-Acet]. D: Merged image of a triple-stained (microtubules, green; β-catenin; red, nuclei, blue) developing eye showing co-distribution of microtubules and β-catenin in the ectoderm (ec), optic cup (ou), and walls of the diencephalon (di) [Form-Acet]. Scale bar = 100 μm in A, 45 μm in B, 25 μm in C, 40 μm in D.


Although the distribution of cytoskeletal components and other proteins was examined in different developing organs, especial emphasis was placed on the development of the neural tube, a delicate structure difficult to preserve (Fig. 8A–G). Microtubules as well as actin filaments are expressed in the three layers of the neural tube. A particular high concentration of microtubules was observed in the apical border of the matrix cells, particularly in those of the roof and floor plates. As expected, the Shh protein was accumulated in cells of the notochord and floor plate.

Figure 8.

Immunofluorescence staining of 3 (A), 5.5 (B–E), and 9 (F,G) day-old freeze-sectioned chick embryos. A: Double-stained (microtubules, green; nuclei, blue) confocal deconvolved merged image of the neural tube showing the matrix (ml) and mantle (mn) layers. vt, ventral root [Form-Acet]. B: Low magnification triple-stained (microtubules, green; actin, red; nuclei, blue) confocal deconvolved image showing the three neural tube layers and sensory ganglia (sg). Notice co-distribution of microtubules and actin filaments in several regions [Form-Acet]. C: Double-stained (microtubules, green; nuclei, blue) confocal merged image showing microtubule-rich fibers in the three layers of the neural tube: matrix (ml), mantle (mn), and marginal (mg) [Form-Acet]. D: Higher magnification of the mantle layer of fig. C [Form-Acet]. E: Double-stained (shh protein, red; nuclei, blue) image showing the distribution of the shh protein in the notochord (no) and floor plate (fp) cells. F: Double-stained (microtubules, green; nuclei, blue) image showing the layers of a more developed neural tube. ml, matrix; mn, mantle; mg, marginal layers. vh, ventral motor horn [Parf-Acet]. G: Similarly stained embryo with microtubule-rich floor plate cells (fp) [Parf-Acet]. Scale bar = 45 μm in A, 65 μm in B, 3 μm in C, 1 μm in D, 95 μm in E, 65 μm in F, 15 μm in G.

mRNA In Situ Hybridization of the Drosophila Embryo

Transcripts of the polarity gene bcd were seen concentrated at the anterior tip of the egg (Kaufman et al., 1990), whereas transcripts of the gap gene kni accumulated in the cytoplasm of cells located at the anterior blastoderm and in a band located close to the middle of the blastula (Nüsslein-Volhard and Wieschaus, 1980. Reviewed by St Johnston and Nüsslein-Volhard, 1992). Transcripts of the pair rule gene eve accumulated in the cytoplasm of blastodermal cells forming seven bands along the anterior posterior axis of the embryo (Macdonald et al., 1986; Frasch et al., 1987) (Fig. 9A–F). Preservation of the cells was very good as judged by observations under high resolution light microscopy. Transcripts of the pair rule gene ftz, that we revealed by a conventional method of mRNA in situ hybridization (Lehmann and Tautz, 1994), gave similar results.

Figure 9.

A–H: mRNA in situ hybridization and combined immunofluoresce/mRNA in situ hybridization of whole-mounted Drosophila eggs (A,B) and blastulae (C–H) [Form-Acet]. A: bcd mRNA expression (arrowhead). B: Distribution of bcd mRNA (arrowheads) among the yolk platelets (yp). C: Distribution of kni mRNA (arrowheads). D: Accumulation of kni mRNA (arrowheads) in the cytoplasm. nu, nuclei. E: Control showing the bands of expression of ftz mRNA (arrowheads) using the Lehmann and Tautz (1994) protocol. F: Accumulation of ftz mRNA (arrowheads) among nuclei. G,H: Combined bcd mRNA in situ hybridization (G) and microtubule immunostaining (H) of the same blastula showing remnants of bcd mRNA expression (arrowhead) and mitotic spindles (sp). Scale bar = 25 μm in A,C,E, 7 μm in B, 8.5 μm in D, 6 μm in F, 20 μm in G,H.

Combined mRNA In Situ Hybridization/Immunofluorescence Staining of the Drosophila Embryo

We studied the distribution of bcd mRNA and microtubules in the same embryo. Results showed fluorescent mitotic spindles scattered across the blastula and bcd mRNA at its anterior pole (See Lu et al., 2009) (Fig. 9G,H). The structural preservation of the embryo and the strength of the in situ and immunofluorescence signals were very satisfactory.

mRNA In Situ Hybridization of the Zebrafish Zygote, Embryo, and 1-Day-Old Larvae

The fixation/permeabilization procedure, with acetic or isobutyric acid gave the best results (Fig. 10A–G). Embryo and larvae did not need to be sliced during the fixation/permeabilization procedure because the hybridization mixtures penetrated the tissues without difficulty. In addition, the specimens remained sufficiently transparent to precisely visualize the stained cells. As compared with mRNA in situ hybridizations performed by us with another method (Thisse and Thisse, 2008), our procedure produced similar hybridization signals, in the predicted regions of the embryo or larva and often within better preserved cells.

Figure 10.

A–G: mRNA in situ hybridization of whole-mounted zebrafish zygotes (A,B), gastrulae (C,D), and 1-day-old (E–G) larvae [Form-Acet]. A: Distribution of snai1a mRNA in the blastodisc (bd) and endoplasmic lacunae (el). B: Distribution of vas mRNA in meridional streamers (ms) and base of the blastodisc (bd). C: Expression of gsc mRNA in cells of the embryonic shield (es). D: Control showing the expression of gsc mRNA in cells of the embryonic shield (es) using the Thisse and Thisse (2008) protocol. Differences between the quality of the in situ hybridizations shown in C and D are confirmed when we compare our results with those obtained by others with a similar method in the same embryonic material (see Shao et al., 2012). E: Expression of brul mRNA in the lens (arrowhead). F: Expression of atoh1a mRNA in the hindbrain (arrowheads). G: Dorsal view of the hindbrain showing expression of atoh1a in the developing cerebellum (cr) and rhombencephalon (ro). Scale bar = 90 μm in A,C,D, 50 μm in B, 200 μm in E, 140 μm in F, 35 μm in G.

Zygote and embryo

The mRNA of snai1a, gsc, and sqt were distributed across very well preserved endoplasmic lacunae, axial streamers and the blastodisc. The same occurred with vas, which was detected along the ectoplasm and meridional streamers (Fuentes and Fernández, 2010) and also at the distal end of the cleavage furrows in the four-cell embryo (Yoon et al., 1997; Knaut et al., 2000; Hashimoto et al., 2004). Expression of gsc was also visualized in the embryonic shield of 50% epiboly gastrulae (Schulte-Merker et al., 1992; Stachel et al., 1993). Similar gastrulae, that we stained for gsc mRNA using a conventional procedure (Thisse and Thisse, 2008), revealed poorer cell preservation and less precise localization of the hybridization signal (compare Figs. 10C and D) (see also Shao et al., 2012).


Transcripts of brul were detected in the lens (Suzuki et al., 2000; Thisse et al., 2001) whereas those of atoh1a were seen accumulated in the walls of the hindbrain (Köster and Fraser, 2001). The preservation of the larvae and the strength of the hybridization signal were at least as good as those of the referred authors.

mRNA In Situ Hybridization of the Chick Embryo

Good results were obtained using fixation/permeabilization with acetic or isobutyric acid (Fig. 11A–C). The distribution of shh mRNA was successfully studied in whole mounts of 3-day-old or sectioned 3- to 5-day-old embryos. Shh transcripts were seen in notochordal and floor plate cells. Distribution of shh in the floor plate formed a gradient along the dorso-ventral axis of the neural tube (Roelink et al., 1995; Ruiz i Altaba et al., 1995).

Figure 11.

mRNA in situ hybridization of whole-mounted (A) and freeze-sectioned (B,C) 3-day-old chick embryos [Form-Acet]. A: Dorsal view showing the expression of shh along the head and trunk regions of the embryo. The signal probably comes from the notochord and neural tube floor plate. ey, eye. B: Transverse section across the trunk showing the expression of shh in the floor plate (fp) and notochord (no) cells. C: Accumulation of shh in the cytoplasm of the floor plate cells (arrowheads). nu, nuclei. Scale bar = 280 μm in A, 40 μm in B, 8 μm in C.


The formaldehyde/glacial acetic acid mixture has been successfully used in the past for the visualization of ooplasmic domains in leech (Fernández, 1980: Fernández and Olea, 1982) and zebrafish (Fernández et al., 2006; Fuentes and Fernández, 2010) eggs and embryos. In contact with the fixation mixture (acid fixation), the ooplasm turns into opaque masses that can be digitally converted into bright structures. To our surprise, the acid fixation turned out to be a suitable fixative/permeabilizer for immunofluorescence and mRNA in situ hybridization, not only of leech and zebrafish embryos, but also of fly and chick embryos. Results indicate that the fixation/permeabilization procedure affords a good general structural preservation of the embryonic cells and tissues and a convenient accessibility of antibodies to antigenic sites and of mRNA probes to native mRNA. When we compare our results with those obtained by the application of conventional methods, we find that our method is simpler, leads to a comparable or better preservation of the cytoplasm, is often more reliable and can be successfully completed in a shorter period of time. An additional advantage of our procedure is its versatility in that the same fixation/permeabilization protocol yields equivalent good results in embryos of different taxa at different developmental stages. In this study, we have obtained satisfactory results using conventional fluorescent microscopy. However, results were further improved using confocal laser scanning microscopy. In both cases, image deconvolution added further quality to the images. Although formaldehyde and paraformaldehyde were successfully used in this study, formaldehyde generally produced better results. This is surprising because paraformaldehyde is considered to be a better fixative for the preservation of immunogenic sites. Formaldehyde prepared from paraformaldehyde is monomeric whereas formaldehyde prepared from the commercial 37–40% solution contains small polymers (Kiernan, 2000). According to this author, to be useful as a fixative formaldehyde must contain the monomer in a hydrated state. We have neither analyzed the ratio of monomer to polymer in our stock formaldehyde nor in the freshly prepared aqueous working solution of it. However, the fact that the formaldehyde/carboxylic acid-fixed embryonic samples hardened in a rather short time, appeared well preserved, and could be successfully processed for immunofluorescence and mRNA in situ hybridization indicates that our acid formaldehyde produced adequate cross-linking of the proteins and nucleic acids. It is possible, however, that a lower monomer concentration in the fixative may have produced less cross-linking of proteins in the tissue, a condition that may contribute to better penetration of the antibodies and mRNA probes afterward. Although some authors claim that protein cross-linking by formaldehyde is slow (Kiernan, 2000), our experience indicates that fixation of the embryonic material with formaldehyde or paraformaldehyde seems to be quick because swelling of the tissue is greatly diminished or imperceptible when a carboxylic acid is added a few minutes after beginning the fixation. Under these fixation/permeabilization conditions immunofluorescence and mRNA in situ hybridization failed. This is considered to indicate that, in a short period, cross-linking of the proteins had advanced beyond the point where carboxylic acids can do their job. We propose that, when carboxylic acids are applied shortly after initiation of the aldehyde fixation, they modulate the rate of further methylene bridge formation and may perhaps trigger a moderate destruction of these bridges. The final outcome of this process is a less drastic distortion of the cytoplasm than that caused by detergents, organic solvents and heat, and a successful entry into the cells of antibodies and mRNA probes that bind to their respective reactive sites. Our method of preserving antigenic sites is thus different to those applied to retrieve them. In the latter case, heat or proteolytic enzymes are considered to loose or break the cross-linkages already formed by the aldehyde, presumably by hydrolysis of methylene bridges. In this manner, the structure of the fixed proteins appear to be re-natured (Shi et al., 2001).

In routine light microscopy techniques, glacial acetic acid is widely used as an agent that induces swelling of the cells, counteracting shrinkage caused by aldehyde fixatives. Swelling caused by the carboxylic acids is a consequence of water entry and retention inside the cells by hydrogen bonding, in part to released lyophilic radicals (Presnell and Schreibman, 1997). It is not surprising, then, that many of the most commonly used fixative mixtures such as the Allen's, Bouin's, Carnoy's, Gendre's, Heidenhain's, and Hollandes's fluids include acetic acid in their formulation (Romeis, 1928; Baker, 1950; Lillie, 1955; Hayat, 2002; Hoppert, 2003). In these cases, acetic acid probably contribute to the permeabilization of the tissues and entry of the dyes. Propionic and isobutyric acids produce less swelling of the cells and in general weaker penetration of the antibodies and mRNA probes. Longer C-chain carboxylic acids probably penetrate more slowly than acetic acid and hence cause less alteration of the formaldehyde-driven protein cross-linking, probably resulting in the formation of tighter protein networks. We are not aware of the use of short C-chain carboxylic acids as permeabilizing agents for immunofluorescence or mRNA in situ hybridization. Glacial acetic acid ranks as the best of the three carboxylic acids tested in our study, not only because it worked well with the embryos of all four organisms but also because cell preservation and the strength of the immunofluorescence and mRNA in situ hybrization signals were better. Methylene bonds cross-link proteins at distances spanning 2.3–2.7 Å (Sutherland et al., 2008), thus producing a tight network. It is proposed that short C-chain carboxylic acids would lead to formation of a looser protein network thereby allowing the penetration of antibodies and mRNA probes. Another way in which carboxylic acids may regulate the degree of protein cross-linking is by lowering the tissue pH, a condition that increases the number of positively charged sites in the peptides that are not available for cross-linking (Hayat, 2002). Finally, it cannot be ruled out that, although the tested permeabilizers are short C-chain carboxylic acids (less than 6C-long), which are generally not considered to act as detergents, they are nevertheless amphipathic molecules that may function as weak detergents. Hence, part of their permeabilization effect may be due to a mild detergent effect. It would be interesting to search whether other short C-chain carboxylic acids might have similar permeabilizing effects.

It was beyond our capabilities to test exhaustively whether our fixation/permeabilization procedure worked well for other invertebrate and vertebrate embryos or even for adult tissues. It was encouraging to find, however, that the cytoskeleton of adult tissues, such as that of rat nerves, was satisfactorily immunostained with our fixation/permeabilization method (J. Fernández and R. Fuentes, unpublished observation).


Embryonic material was obtained from four organisms: leech (Theromyzon trizonare), insect (Drosophila melanogaster), fish (Danio rerio), and bird (Gallus gallus).

Collection, Handling, and Preparation of Embryonic Material

Leeches were maintained in 0.03% Instant Ocean Salt at 20°C. Ripe eggs were removed by laparotomy of pregnant mothers and developed at 20°C in the same salt solution. Drosophila adults were maintained at 20–25°C on standard Drosophila medium. Embryos were removed from the medium, bleached with 50% sodium hypochlorite, and cleaned with tap water and 0.4% NaCl. Zebrafish were maintained at 28°C in aerated aquaria subjected to a 14-hr light/10-hr dark photoperiod. Laid eggs were collected and developed at the same temperature. Fertilized chick eggs were obtained from a local dealer and incubated at 37–38°C for 3–9 days. The embryos were removed and rinsed in warm saline (0.9% NaCl). Handling of animals complied with the institutional guidelines (Ethics Committee, Faculty of Sciences, University of Chile).

Fixation/Permeabilization for Immunofluorescence and mRNA In Situ Hybridization

Detergents, methanol, rehydration, proteinase K, and refixation were omitted in all the following protocols.

Leech, zebrafish, and chick embryos

Embryos were fixed/permeabilized for 2–3 hr at room temperature (RT) and continuous agitation (CA) in freshly prepared 5% formaldehyde (Form) (37%, Merck) in distilled water or 4% paraformaldehyde (Parf) (Electron Microscopy Sciences) in 0.1 M phosphate buffer pH 7.4, to which droplets of either of the following carboxylic acids was added 5–10 sec after initiation of the fixation: glacial acetic (Acet) (100%, Merck), propionic (Prop) or isobutyric (Isobut) acids (both 99% and from Sigma). In some cases, we used paper filtered 37% formaldehyde kept neutralized with chalk (nForm). The standard fixation/permeabilization procedure was 2 hr in the fixative containing 5–8% glacial acetic acid (3 ml of formaldehyde + 3–5 droplets of the carboxylic acid, final pH about 3). In some cases, two carboxylic acids were simultaneously used. Carboxylic acids, particularly the glacial acetic, provokes mild swelling of the samples. The later the carboxylic acid is added, the less the embryo swells but its permeabilization diminishes. Protein stabilizers, chelators, or solvents such as 35 × 103 Da polyethylenglycol (Peg, final concentration 4%, Merck), Taxol (Tx) and/or Phalloidin (Ph) (final concentration 1 μg/ml, Sigma), MgSO4 (final concentration 1 mM, Merck), EGTA (final concentration 5 mM, Sigma), and DMSO (final concentration 0.5%, Sigma) were sometimes added to the fixative. Addition of glutaraldehyde (final concentration 0.125%. Electron Microscopy Science) impaired penetration of the antibodies. Before or soon after initiation of the fixation/permeabilization procedure leech and zebrafish zygotes and embryos were mechanically dechorionated, while large chick embryos were transversally sectioned into 2- to 3-mm-thick slices.

Drosophila embryos

Embryos dechorionated in 50% bleach solution were transferred to a scintillation tube with 500 μl of 0.4% NaCl and an equal amount of heptane (Merck). After shaking for 30 sec, 50 μl of formaldehyde (final concentration 5%) and 50 μl of the carboxylic acid (final concentration 5%) were added. After 20–30 min of intermittent gentle agitation, the aqueous phase was removed and replaced with 500 μl of 100% methanol (Merck) and vortexed for 1 min. Protein stabilizers were sometimes added to the fixative mixture.

Immunofluorescence Staining

After fixation/permeabilization samples were washed 4× 15 min in 1× phosphate buffered saline (PBS). For freeze sectioning (15–20 μm thickness) embryos were soaked in 30% sucrose until the samples reached the bottom of the vial. After blocking for 1 hr in 2% bovine serum albumin (BSA; Merck) in 1× PBS, whole or sectioned embryos were incubated for 1–18 hr, at RT and CA, in the primary antibody in 2% BSA in 1× PBS, containing 0.2% sodium azide (Fisher). The following IgG primary antibodies were used: polyclonal (Sigma) or monoclonal (Calbiochem) anti-α-or-β-tubulin (1:100), monoclonal acetylated or tyrosinated anti-tubulin (1:100, Sigma), monoclonal anti-γ-tubulin (1:50, Sigma), polyclonal nonmuscle anti-myosin (1:100, Sigma), polyclonal sarcomeric anti-myosin (1:100, Sigma), polyclonal anti-β-catenin (1:100, Sigma), monoclonal anti-shh (1:100, Santa Cruz). IgM monoclonal anti-actin (1:100, Calbiochem) was also tested. After 1–2 hr washes in several changes of 1× PBS at RT and CA, samples were incubated for 1–18 hr in 1:100 secondary antibody in 2% BSA in 1× PBS with 0.2% sodium azide at RT, darkness and CA.

In some cases, the incubation in the secondary antibody was extended for another 24 hr at 4–8°C. Good immunostaining may also be achieved by 1- to 2-day staining in each of the antibodies at 4–8°C. The secondary antibodies were labeled with Cy2 (Jackson), Alexa Fluor 488 or 594 (Molecular Probes). After 1–2 hr wash in several changes of 1× PBS at RT and CA, whole or sectioned embryos were mounted between 2 cover slips in 9:1 glycerol (Merck) 1× PBS. One cover slip was sealed with nail enamel to a metallic slide having a rectangular window. In the case of whole embryos, plasticine spacers were placed between the cover slips. F-actin was also stained with Phalloidin labeled with Alexa Fluor 488, Alexa Fluor 594 (Molecular Probes) or TRITC (Sigma) at a concentration of 1–10 μg/ml. In this case staining was performed during the fixation/permeabilization step or in combination with the secondary antibody incubation. Nuclei and chromosomes were stained during 10–15 min with either DAPI (Sigma) or TO-PRO-3 (1–10 μg/ml. Molecular Probes) during the washes that followed incubation in the secondary antibody.

Controls were incubated in the secondary antibody only. Samples were examined in a Zeiss Axiovert 135 inverted fluorescence microscope (Oberkochen, Germany) equipped with a Z-motor (Prior Scientific Instruments. Cambridge, UK) and a Hamamatsu chilled CCD camera (model C5985. Hamamatsu City, Japan). Image grabbing and analysis were performed using the 6.1 version of Metamorph. Stack of images were also acquired with an Olympus (1X81-ZDC2) or a Zeiss (LSM 510 Meta) confocal microscope. Some immunofluoresce and confocal images were deconvolved with the Huygens Professional software version 3.5. Projection of confocal images was sometimes prepared with the ImageJ software. To compare results obtained with our procedure with those obtained with other techniques (controls), leech embryos were permeabilized, fixed, and immunostained as in Fernández and Olea (1995), zebrafish embryos according to Solnica-Krezel and Driever (1994) and Drosophila embryos as recommended by Karr and Alberts (1986).

mRNA In Situ Hybridization

Leech embryos were not used because no probes were available for Theromyzon trizonare. For the other embryos, we modified current protocols for mRNA in situ hybridization. The following transcripts were detected using the corresponding anti-sense mRNA. Zebrafish: squint (sqt), goosecoid (gsc), vasa (vas), snail 1a (snai1a), bruno-like (brul), and atonal 1a (atoh1a). Drosophila: bicoid (bcd), knirps (kni), even-skipped (eve), and fushy tarazu (ftz). Chick: sonic hedgehog (shh). Gsc, sqt, snai1a, brul, and vas were obtained as cDNA inserted in plasmids, whereas bcd, kni, eve, ftz, shh, and atoh1a as labeled antisense mRNA (riboprobes). To synthesize labeled antisense mRNA from cDNA, plasmids were linearized with restriction enzymes, the DNA purified by phenol/chloroform, precipitated with ethanol, centrifuged, resuspended, and transcribed in vitro using Dig mix NTPs (Roche) and RNA polymerase (Fermentas) following manufacturer's instructions. After digesting the template DNA with DNase (Fermentas), the antisense mRNA was precipitated with 5 M LiCl, centrifuged, washed, dried, and re-suspended in RNase-free water (for more details see Thisse and Thisse, 2008). In situ hybridizations were examined in an Olympus MVX10 dissecting microscope or a Zeiss 135 inverted microscope.

Drosophila (modified from Lehmann and Tautz, 1994)

After fixation/permeabilization, samples were washed at RT and CA for 4× 5 min in 1× PBS and 10 min in 500 μl of each of the following hybridization mixes (HM): 1:1 HM B/1× PBS and HM B. For prehybridization, embryos were incubated for 1 hr at 65°C in 500 μl of HM A (without salmon sperm DNA) and for hybridization 18 hr or overnight at 65°C in the same amount of HM A containing the riboprobe: 10–20 ng of antisense digoxigenin-labeled mRNA. After 5 min heating at 80°C, the riboprobe mixture was cooled on ice/water and ready to be used. Hybridization was followed by washes: 3× 20 minutes in HM B at 65°C, 20 min in 1:1 HM B/1× PBS at RT and 4× 20 min in 1× PBS at the same temperature. After blocking with 2% BSA in 1× PBS for 1 hr, embryos were incubated for 2 hr in 1:2,000 anti-digoxigenin antibody coupled to alkaline phosphatase (Roche) in 2% BSA in 1× PBS at RT. Embryos were then washed under the same conditions for 4× 20 min in 1× PBS, 4× 5 min in alkaline phosphatase (AP) buffer (5 ml 1M Tris pH 9.5, 2.5 ml 1M MgCl2, 7.5 ml 1M NaCl, 35 ml dH20) and stained 15–30 min in 200 μl of BM purple (Roche) in darkness, RT and discontinuous agitation (DA). Staining was checked under the microscope. After 2× 3 min washes in 1× PBS to stop staining, embryos were dehydrated, cleared in xylene, and mounted with Permount (Fisher Scientific). For comparison Drosophila embryos were also subjected to mRNA in situ hybridization following the Lehmann and Tautz (1994) protocol.

Zebrafish (modified from Thisse and Thisse, 2008)

After fixation/permeabilization samples were washed 4 × 15 min in 1× PBS at RT and CA. When needed, bleaching of the pigment cells was performed using hydrogen peroxide. Prehybridization was carried out by 2–4 hr incubation in 800 μl of the HM at 70°C and hybridization by 18 hr or overnight incubation at the same temperatures in the HM with the riboprobe. This was prepared in 200 μl of HM containing 100–200 ng of antisense digoxigenin-labeled mRNA. After heating for 5 min at 70°C, the HM/riboprobe was ready to be used. Hybridization was followed by 10 min washes at 70°C in each of the following solutions: 75%, 50%, 25% formamide in 2× standard saline citrate (SSC) and 100% 2× SSC. Then, samples were washed in 0.2× SSC in 1× PBS according to the original protocol and blocked for 2 hr in 2% BSA in 1× PBS at RT. Samples were then incubated in 1:2,000 antidigoxigenin antibody labeled with alkaline phosphatase in 2% BSA in 1× PBS for 8 to 18 hr at RT. After 5× 20 min washes in 1× PBS and 3× 5 min in AP buffer, samples were stained in BM purple for 2–6 hr (early embryos) or 16–20 hr (larvae) at RT and DA (check under the microscope). Subsequently, samples were washed 2× 1 hr in 1× PBS. Some embryos and larvae were whole-mounted in 9:1 glycerol/1× PBS while others were soaked in 30% sucrose, frozen and sectioned at 15–20 μm and then similarly mounted. For comparison, zebrafish embryos were subjected to mRNA in situ hybridization according to Thisse and Thisse (2008).

Chick (modified from Henrique et al. (1995)

After fixation/permeabilization whole (early) or sliced (late) embryos were washed 4× 15 min in 1× PBS at RT and CA. Samples were prehybridized for 3 hr by incubation at 65°C in HM and hybridized 18 hr or overnight at the same temperature in HM with the riboprobe. The riboprobe was prepared in HM containing 30–60 ng/ml (for frozen sections) or 300–600 ng/ml (for whole embryos) of antisense digoxigenin-labeled mRNA. The mixture was heated for 5 min at 70°C before use. EDTA (ethylenediaminetetraacetic acid) and CHAPS were not added to the HM. After 2× 30 min washes at 65°C in solution A (12.5 ml 20× SCC pH 4.5, 25 ml 100% formamide, 12.5 ml dH2O), 3× 30 min washes in solution B (6.25 ml 20× SSC pH 4.5, 25 ml 100% formamide, 18.75 ml dH20), and 2× 5 min washes at RT in MAB (100 mM maleic acid/150 mM NaCl pH 7.5), samples were blocked in 2% BSA in 1× PBS for 3 hr at RT. Specimens were then incubated in 1:2,000 antidigoxigenin antibody labeled with alkaline phosphatase in 1× PBS overnight at 4°C. After 6× 40 min washes at RT in MAB and 5 min in AP buffer, samples were stained in NBT/BCIP (Roche) at RT until labeling of the mRNA (visualized under the microscope) was acceptable. After two washes in 1× PBS, samples were whole-mounted or soaked in sucrose, frozen, sectioned, and mounted in 9:1 glycerol 1× PBS.

Combined mRNA In Situ Hybridization/Immunofluorescence Staining

To combine mRNA in situ hybridization with immunofluorescence, Drosophila embryos were used. Embryos were fixed/permeabilized as indicated above. They were first processed for mRNA in situ hybridization and then for immunofluorescence. After BM purple staining, embryos were washed in 1× PBS, incubated in the desired primary and secondary antibodies, washed, dehydrated, cleared in xylene, and mounted in Permount. The concentration of the riboprobes and antibodies were the same as those used for the procedures performed separately.


We thank Drs. John Ewer and Bruce Cassels for critical reading of the manuscript. Drs. J. Ewer, V. Gallardo, W. Gehring, V. Palma, F. Pelegri, A. Reyes, J. Botelho, and L. Valdivia for donation of cDNA-inserted plasmids, riboprobes or antibodies. Drs. M. Allende, A. Couve, A. Glavic, V. Palma, and J. Mpodozis for allowing us to use their laboratory facilities. V. Guzman, F. Saavedra, L. Saragoni, M. Salinas, L. Ossa, and O. Ramírez for technical assistance. Financed by the University of Chile.