RNA interference is ineffective as a routine method for gene silencing in chick embryos as monitored by fgf8 silencing

Abstract

The in vivo accessibility of the chick embryo makes it a favoured model system for experimental developmental biology. Although the range of available techniques now extends to miss-expression of genes through in ovo electroporation, it remains difficult to knock out individual gene expression. Recently, the possibility of silencing gene expression by RNAi in chick embryos has been reported. However, published studies show only discrete quantitative differences in the expression of the endogenous targeted genes and unclear morphological alterations. To elucidate whether the tools currently available are adequate to silence gene expression sufficiently to produce a clear and specific null-like mutant phenotype, we have performed several experiments with different molecules that trigger RNAi: dsRNA, siRNA, and shRNA produced from a plasmid coexpressing green fluorescent protein as an internal marker. Focussing on fgf8 expression in the developing isthmus, we show that no morphological defects are observed, and that fgf8 expression is neither silenced in embryos microinjected with dsRNA nor in embryos microinjected and electroporated with a pool of siRNAs. Moreover, fgf8 expression was not significantly silenced in most isthmic cells transformed with a plasmid producing engineered shRNAs to fgf8. We also show that siRNA molecules do not spread significantly from cell to cell as reported for invertebrates, suggesting the existence of molecular differences between different model systems that may explain the different responses to RNAi. Although our results are basically in agreement with previously reported studies, we suggest, in contrast to them, that with currently available tools and techniques the number of cells in which fgf8 gene expression is decreased, if any, is not sufficient to generate a detectable mutant phenotype, thus making RNAi useless as a routine method for functional gene analysis in chick embryos.

Figure 1

siRNA and DIG-siRNA obtained by in vitro Dicer-digestion. A 3% agarose gel electrophoresis of siRNA and DIG-labelled siRNA to fgf8 is shown. Note that Dicer is able to generate a pool of siRNAs from both dsRNA and DIG-labelled dsRNA.

Figure 2

Plasmids and shRNA constructs. A) Wild-type psiRNA-hH1GFPzeo G2 vector map showing Acc65I/HindIII restriction sites (*), the α-peptide coding sequence located downstream of the H1 promoter, and the GFP coding sequence located downstream of hEF1-HTLV promoter. B) Engineered DNA oligonucleotides for the shRNAs: 21-mers are shown in green; the spacer is shown in blue; HindIII and Acc65I restriction sites are shown in orange; and the poly-A sequence is shown in black. The presumptive secondary structure of shfgf8b-1 is also shown.

Figure 3

Localisation of DIG-siRNA 24 h after electroporation. Embryos co-electroporated with a pool of DIG-labelled siRNAs and the pEGFP-C2 plasmid. A) GFP fluorescence into the electroporated side of the neural tube before DIG immunodetection (arrows). Yellow dashed line is delimiting the GFP expression area. B) GFP fluorescence into the electroporated side of the neural tube of the embryo shown in A) after DIG immunodetection (arrows). Note a decrease in GFP fluorescence due to the presence of the NBT/BCIP precipitate. C) Immunodetection of DIG-labelled siRNA in the neural tube (in blue; arrows) in the same embryo than A) and B). D) Merging of B) and C). Yellow dashed line has been extrapolated from A) and is delimiting the area of GFP expression before NBT/BCIP precipitation. Note that most cells exhibiting GFP fluorescence also show the presence of the DIG-siRNAs, although they do not colocalize in all electroporated cells. Also note the absence of siRNA spreading from the electroporated area. Abbreviations: Ist, isthmus; Oc, optic cup; Tel, telencephalon.

Figure 4

α-peptide expression in embryos transfected with the psiRNA-hH1GFPzeo G2 vector. A-B) Two different embryos 12 h after electroporation hybridised with the α-peptide antisense riboprobe. Note the expression of the α-peptide sequence in the electroporated half-side of the neural tube (arrows). C-D) Two different embryos 24 h after electroporation hybridised with the α-peptide antisense riboprobe. Note that the expression of the α-peptide is properly localized into the isthmus (arrows). E) Facial view of the embryo shown in D. Note that the α-peptide is expressed only in electroporated side (arrowhead).

Figure 5

Fgf8 and isthmic morphology in flattened isthmus of embryos transfected with shfgf8b-1 and shfgf8b-2. Dorsal view of a flattened isthmus hybridised with the fgf8 riboprobe 24 h after plasmids electroporation. The black line follows the dorsal midline. Note that no significant differences either in fgf8 expression or neuroectoderm morphology are observed when comparing both sides of the isthmus.

Figure 6

Fgf8 expression and isthmic morphology in whole mount embryos transfected with shfgf8b-1 and shfgf8b-2. Rostral is to the left. A) Embryo 24 h after electroporation, prior to fgf8 in situ hybridisation. Note that the morphology of the isthmic derivatives is completely normal (compare with control embryo shown in C). A') Magnification of A) showing the isthmic area. B) GFP fluorescence (arrowheads) in the electroporated side of the same embryo shown in A). B') Magnification of B) showing the area of GFP expression (arrowheads). C) Fgf8 in situ hybridisation of a control embryo. D) Fgf8 in situ hybridisation in the isthmus of the embryo shown in A) (arrow). Note that fgf8 expression in the isthmus is very similar to that shown in C). Abbreviations: AERfl, forelimb apical ectodermal ridge; AERhl, hindlimb apical ectodermal ridge; BA, branchial arches; Ist, isthmus; Met, metencephalon; Mes, mesencephalon; Tel, telencephalon.

Figure 7

Fgf8 expression and isthmic morphology in sections of embryos transfected with shfgf8b-1 and shfgf8b-2. Transversal sections of the neural tube at the isthmus level. Both sides of neural tube are shown. The electroporated side is to the left. A) Fgf8 expression detected by in situ hybridisation. B) GFP expression in the same section shown in A) detected by GFP immunostaining. Note that GFP is detected only in the electroporated side (arrowheads). C) Merging of A) and B). The expression of fgf8 is very similar in electroporated and control neuroepithelium. D to F) Magnification of sections shown in A to C respectively. Note the presence of cells showing fgf8 and GFP expression simultaneously (arrows and arrowhead). G to I) Magnification of sections shown in A to C respectively in a different focal plane. Note the presence of cells showing GFP expression but not fgf8 expression (double arrowhead).

Figure 8

Apoptosis in embryos transfected with shfgf8b-1 and shfgf8b-2. A) Dorsal view of a transfected embryo in which apoptotic cells have been detected by TUNEL assay (brown nuclei). Note the absence of morphological defects (the transfected side is to the bottom). B) GFP fluorescence in the same embryo. The transfected area includes the isthmus. C) Merging of B) and C). D) Magnification of the isthmic area of A). Note that both sides of the embryo exhibit a similar amount of apoptotic cells (arrows), which affect mostly the ectoderm. E) Magnification of the isthmic area of B). The arrow points an apoptotic cell not expressing GFP. F) Magnification of the isthmic area of C). G) Dissected cephalic cavity showing the internal side of the mesencephalon, the metencephalon and the isthmus. Apoptotic cells have been detected by TUNEL assay (brown nuclei). The morphology of the transfected side (left hand side) is completely normal, and that apoptotic cells are equally distributed at both sides of the embryo. H) GFP fluorescence in the same embryo. The transfected area includes the isthmus. I) Merging of G) and H). J) Apoptosis control in the eye of the electroporated side. Note that apoptotic cells are located in the dorsal portion of the optic cup as described in the literature for normal embryos (Trousse et al., 2001). K) Apoptosis control in the ear of the electroporated side. Note that apoptotic cells are located in the dorsal portion of the otic vesicle as described in the literature for normal embryos (Lang et al., 2000). Abbreviations: Ant, anterior; Mes, mesencephalon; Met, metencephalon; Pos, posterior.