The RNaseIII enzyme Dicer is required for morphogenesis but not patterning of the vertebrate limb

Abstract

The RNaseIII-containing enzyme Dicer is believed to be required for the processing of most, if not all, microRNAs (miRNAs) and for processing long dsRNA into small interfering RNAs. Because the complete loss of Dicer in both zebrafish and mice results in early embryonic lethality, it has been impossible to determine what role, if any, Dicer has in patterning later tissues in the developing vertebrate embryo. To bypass the early requirement of Dicer in development, we have created a conditional allele of this gene in mice. Using transgenes to drive Cre expression in discrete regions of the limb mesoderm, we find that removal of Dicer results in the loss of processed miRNAs. Phenotypically, developmental delays, in part due to massive cell death as well as disregulation of specific gene expression, lead to the formation of a much smaller limb. Thus, Dicer is required for the formation of normal mouse limbs. Strikingly, however, we did not detect defects in basic patterning or in tissue-specific differentiation of Dicer-deficient limb buds.

Fig. 1.

Construction of a Dicer conditional mouse allele. (A) Schematic of the Dicer conditional targeting construct. (B) Southern blot of correctly targeted embryonic stem cells. By using a 3′ probe, a 2.9-kb fragment is detected in addition to the wild-type 4.0-kb fragment in targeted embryonic stem cells. (C) Upon Cre-mediated recombination of the Dicer^flox allele in the germ line, homozygous mutant embryos resemble the reported null allele (15). (D) RNA was extracted from primary fibroblasts from wild-type or Dicer^flox/Dicer^flox mice. Parallel experiments with GFP adenovirus indicated near 90–100% infection efficiency. The presence of a small amount of processed let-7 and miR-21 RNA in Dicer-null cells may be due to the inclusion of Dicer-positive cells in our RNA preparations or may indicate that a small amount of miRNA processing is independent of Dicer. (E) A representative example of a cell displaying defects in chromosome segregation upon removal of Dicer. The arrowhead points to a micro nucleus, and the brackets indicate an anaphase spindle. Counting the number of anaphase spindles and satellite nuclei present among four randomly chosen fields of DAPI-stained cells revealed that ≈25% of the Dicer-null cells displayed defects in chromosome segregation and accumulation of satellite nuclei despite a normal morphological appearance under the light microscope. In control adeno-GFP-treated cells, 1.4% of the cells contained satellite nuclei but no anaphase bridges.

Fig. 2.

let-7 is not correctly processed in limbs that lack Dicer. RNA was extracted from wild-type, heterozygous, or Dicer-null E12.5 forelimbs. The small amount of mature let-7 present in Dicer^flox/Dicer^flox;prx1cre forelimbs may result from the correct processing of let-7 in the limb ectoderm [let-7 is known to be expressed in the limb ectoderm (35), and the prx1cre allele removes Dicer activity only from the limb mesoderm].

Fig. 3.

Loss of Dicer in the mouse limb leads to cell death. A, C, and E show wild-type limbs, and B, D, and F show limbs in which Dicer has been removed from the limb mesoderm using the prx1cre allele. An increase in cell death in Dicer-minus limbs compared with age-matched littermates was observed in E10.5 limbs (compare A and B). By E11.5, Dicer-minus limbs were narrower than wild-type limbs and exhibited an increase in cell death throughout the limb mesoderm. Note that loss of a single copy of Dicer does not lead to an increase in cell death (compare C and D). By E12.5, a wild-type pattern of cell death was observed in the apical ectodermal ridge (AER) of wild-type limbs (arrow in E). Dicer-minus limbs exhibited cell death in the AER in addition to ectopic cell death throughout the limb mesoderm (F). Dicer-minus limbs were also smaller and digit condensations were not visible in E12.5 mutant limbs. The ×2 magnification insets in E and F are of the distal, posterior region of the respective limbs. All pictures are of forelimbs from littermates. Images in A–D were taken at the same magnification.

Fig. 4.

Skeletal preparations of E20 Dicer-minus limbs. (A) The upper image shows a wild-type E20 forelimb, and the lower image shows a Dicer^flox/Dicer^flox;prx1cre forelimb from an age-matched littermate. (B) Magnification of the mutant limb shown in A. Notice that despite its small size the mutant limb contains most of the bones associated with normal limb development. (C) The upper image shows a wild-type E20 hindlimb, and the lower image shows a Dicer^flox/Dicer^flox;prx1cre hindlimb from an age-matched littermate. (D) Magnification of the mutant limb shown in C.

Fig. 5.

Loss of Dicer in cells that express Shh does not alter Shh signaling or cell migration. Removal of Dicer in Shh cells of E18.5 embryos was accomplished by mating Dicer^flox/+;Shhgfpcre/+ males with Dicer^flox/Dicer^flox;R26R/+ females. Using the R26R allele, we were able to identify and follow the fate of cells that expressed the Shhgfpcre allele, had undergone a recombination event activating the LacZ gene, and lacked Dicer. Previously, we had demonstrated that cells that express Shh in the ZPA migrate to the distal limb and form digits 5 and 4 and part of digit 3 (19). In limbs in which Dicer was removed in Shh-expressing cells, abnormal cell migrations were not observed. No blue cells (Dicer-minus) were observed in abnormal locations compared with wild-type age-matched littermates.

Fig. 6.

Gene expression in a Dicer-minus limb. (A) RNA in situ hybridizations for MyoD (muscle) staining. The upper image shows a Dicer^flox/Dicer^flox;prx1cre E12.5 forelimb, and the lower image shows a Dicer^flox/+;prx1cre control limb. (B) RNA in situ hybridizations for MyoD staining in an E11.5 wild-type forelimb. Notice that the staining pattern in an E12.5 Dicer^flox/Dicer^flox;prx1cre E12.5 forelimb is similar to MyoD expression in an E11.5 limb. (C) RNA in situ hybridizations for Gdf5, a marker of joint formation. Staining is in an E13.0 wild-type forelimb. (D) Gdf5 staining in a Dicer^flox/Dicer^flox;prx1cre forelimb (×2 magnified compared with C). All control limbs are from age-matched littermates. (E and F) Spry2 whole-mount RNA in situ hybridizations on Dicer^flox/Dicer^flox (E) and Dicer^flox/Dicer^flox;prx1cre (F) E11.75 littermates, respectively. Notice that Spry2 is expressed at much higher levels in mesoderm that lacks Dicer (F). For all probes, more than four mutant and wild-type embryos were examined.

Fig. 7.

FGF signaling is reduced in limb buds lacking Dicer activity. Frontal sections through E10.5 control (A) and Dicer^flox/Dicer^flox;prx1cre (B) limbs were immunostained with an antibody against phosphorylated ERK/mitogen-activated protein kinase (MAPK), an indicator of active FGF and other RTK signal transduction as described in ref. . Phosphorylated ERK/MAPK is detectable at high levels in the mesoderm directly apposed to the apical ectodermal ridge (AER), a potent source of FGF signals, and is present in a gradient extending proximally through the mesoderm from the distal tip (A) (31). In contrast, in limb buds lacking Dicer activity (B), this graded signal is absent, indicating a reduction in FGF/RTK signaling. High levels of phosphorylated ERK/MAPK are still detected beneath the AER, consistent with maintained FGF signaling in the limb ectoderm.