Conditional brain-specific knockdown of MAPK using Cre/loxP regulated RNA interference

Abstract

In the last years, RNA interference (RNAi)-mediated gene knockdown has developed into a routine method to assess gene function in cultured mammalian cells in a fast and easy manner. For the use of this technique in developing or adult mice, short hairpin (sh)RNA vectors expressed stably from the genome are a faster alternative to conventional knockout approaches. Here we describe an advanced strategy for conditional gene knockdown in mice, where we used the Cre/loxP system to activate RNAi in a time and tissue dependent manner in the adult mouse brain. By placing conditional RNAi constructs into the defined genomic Rosa26 locus and by using recombinase mediated cassette exchange (RMCE) instead of laborious homologous recombination, we developed a fast, easy and reproducible approach to assess gene function in adult mice. We applied this technique to three genes of the MAPK signaling pathway-Braf, Mek1 and Mek2-and demonstrate here the potential of this new tool in mouse mutagenesis.

Figure 1.

RNAi activity of H1 or U6 promoter based shRNA expression vectors modified by loxP sites. (A) Overview of the tested shRNA vectors. The position of the loxP sites (red triangle) in relation to the start of transcription (black arrow) and the shRNA region (blue arrows) are indicated. In the resulting shRNAs loxP derived sequences are shown in red color. All constructs contain identical sense and antisense shRNA sequences with specificity to β-Galactosidase (lacZ). (B) Sequences of the H1 and U6 promoter constructs. The different positions of the loxP and lox2272 sites (red), respectively, within or outside of the promoter are shown. The nucleotide of the transcriptional start is highlighted in gray. In capital letters: TATA box of the promoter; bold: original loop region and in blue and underlined: shRNA sense and antisense regions. (C) Silencing activity of β-Galactosidase specific shRNA vectors upon transient cotransfection with β-Galactosidase and F-Luciferase expression plasmids into ES cells. Values are expressed as ratio of β-Galactosidase to Luciferase activity in RLU in comparison to the positive control. The results obtained with the parental shRNA vectors H1-shLacZ and U6-shLacZ are shown as black columns. Numbers above the columns indicate the efficiency of silencing of β-Galactosidase activity in percent in comparison to the positive control. All results from duplicate samples are expressed as mean values with standard deviation.

Figure 2.

Cre/loxP controlled conditional shRNA vectors. (A) Scheme for the modification of an shRNA vector into a Cre/loxP controlled version by insertion of a loxP flanked stop cassette into the shRNA loop region. The conditional shRNA vector can be activated by excision of the stop cassette through Cre mediated deletion. A single loxP site remains within the shRNA loop. (B) Transient transfection of β-Galactosidase (lacZ, blue columns) or F-Luciferase (Luc, orange columns) specific conditional shRNA vectors before (U6-lox-lox-shLuc; U6-lox-lox-shLacZ) and after excision of the loxP flanked stop cassette (U6-loxP5-shLuc; U6-loxP5-shLacZ) into ES cells. For F-Luciferase shRNA vectors values are expressed as ratio of Luciferase to β-Galactosidase activity (left Y axis) and for β-Galactosidase shRNA vectors as ratio of β-Galactosidase to Luciferase activity (right Y axis), in comparison to the appropriate positive control. Results are expressed as mean values from duplicate samples with SD. U6-shLuc, U6-shLacZ: parental shRNA vectors; RLU: relative light units.

Figure 3.

Cre mediated activation of a single copy conditional shRNA vector within the Rosa26 locus of murine ES cells. (A) One Rosa26 allele of ES cells was modified by a gene-targeting vector (R26.5) that introduced a splice acceptor-lacZ cassette and a hygromycin resistance gene such that β-Galactosidase (lacZ) is expressed from the endogenous Rosa26 promoter. R26.5 ES cells were further modified with a gene targeting vector (R26.9) that introduced a neomycin resistance gene and the conditional shRNA vector U6-lox-lox-shLacZ. The R26.5/R26.9 ES cells were transiently transfected with a Cre expression vector and subclones that recombined the R26.9 allele (R26.9Δ) were isolated. (B) X-Gal staining of fixed R26.5/R26.9 ES cells in comparison to a R26.5/R26.9Δ clone shows highly reduced β-Galactosidase activity in the latter cells (magnification 20×). (C) Comparison of β-Galactosidase activity in lysates of R26.5/R26.9 ES cells (mean of three non-deleted subclones, blue column) in comparison to two deleted subclones (R26.9Δ-1, R26.9Δ-2, red columns) and wild type ES cells (WT). Values are shown as β-Galactosidase activity in RLU per micrograms protein of the lysates in comparison to the non-deleted clones and are expressed as mean values with SD.

Figure 4.

Vector construction and RMCE for the generation of shBraf-flox and shMek-flox mice. (A) The conditional shRNA expression cassette from the vector pbs-U6-shRNA-flox contains the U6 promoter in front of the sense (s) sequence of the shRNA, the loxP (lox) flanked stop cassette in the loop region and the antisense (as) shRNA sequence. The shRNA expression cassette is cloned into the RMCE donor vector pRMCE behind the promoterless neomycin selection marker (neo-bpA) so that the two attB sites from pRMCE flank the selection marker as well as the shRNA expression cassette. (B) Acceptor ES cells for RMCE harbor a Rosa26 allele where in intron 1 a splice acceptor (SA) is inserted followed by a pgk promoter (pgk) driving a hygromycin selection marker (hygro-bpA) which is flanked by attP sites. (C) Upon RMCE with C31Int, the attP flanked cassette in the acceptor ES cells from B is replaced by the attB flanked cassette from the donor vector in A. FRT (f) sites allow to excise the pgk promoter and the neomycin marker in recombined ES cells or mice. E: EcoRV; B: BamHI; probe: 5′-Rosa26 probe and E1: exon 1.

Figure 5.

Analysis of RNA interference in shBraf^+/flox/CamKII-cre mice. (A) Tissue-specific activation of shRNA in the brain with CamKII-cre. Southern blot analysis of BamHI digested genomic DNA from different brain regions of adult shBraf^+/flox/CamKII-cre mice. Lane 1: DNA from whole brains of shBraf^+/flox control mice; lane 2–10: DNA from indicated regions of shBraf^+/flox/CamKII-cre mice. The wild type Rosa26 allele (wt) gives a 5.8 kb band and the band from the shRNA allele is shifted from 5.4 kb with the stop cassette (flox) to 8.6 kb after Cre recombination (del). (B) Expression of shRNA against Braf in adult forebrain. On a Northern blot with small RNAs against the sequence of shBraf the 21 nt band of the processed siRNA against Braf is only detectable with the control oligonucleotide and in mutant mice but not in control mice. (C) BRAF protein reduction in forebrain regions of shBraf^+/flox/CamKII-cre mice. On a Western blot with protein from the indicated brain regions from adult mouse brain, knockdown of BRAF protein in forebrain regions of mutant mice is shown in comparison to the protein level from control mice. β-ACTIN was used as a loading control. OB: olfactory bulb, HC: hippocampus, St: Striatum, fCx: frontal cortex, pCx: posterior cortex, Th: Thalamus, MB: midbrain, CB: cerebellum, BS: brainstem, +/flox: shBraf^+/flox and +/Δ: shBraf^+/flox/CamKII-cre.

Figure 6.

Analysis of RNA interference in shMek^+/flox/Nestin-cre mice. (A) Tissue specific activation of shRNA in the brain with Nestin-cre. Southern blot analysis of BamHI digested genomic DNA from different brain regions of adult shMek^+/flox/Nestin-cre mice. Lane 1: DNA from whole brains of shMek^+/flox control mice; lane 2–10: DNA from indicated regions of shMek^+/flox/Nestin-cre mice. The wild type Rosa26 allele (wt) gives a 5.8 kb band and the band from the shRNA allele is shifted from 5.4 kb with the stop cassette (flox) to 8.6 kb after Cre recombination (del). (B) Expression of shRNA against Mek1 and Mek2 in whole adult brain. On a Northern blot with small RNAs against the sequence of shMek the 23 nt band of the processed siRNA against Mek1 and Mek2 is highly detectable with the control oligonucleotide and in mutant mice but not in control mice. (C–E) Knockdown of Mek1 mRNA. On a Northern blot of total RNA from whole adult brain (C) Mek1 mRNA from mutant mice is decreased ∼65% compared to control mice. β-Actin was used as a loading control. In situ hybridization against Mek1 mRNA on coronal brain sections shows the knockdown effect throughout the brain of mutant mice (E) in contrast to the high wildtype expression of Mek1 mRNA in control mice (D). (F) MEK1 and MEK2 protein reduction in the brain of shMek^+/flox/Nestin-cre mice. On a Western blot with protein from the indicated regions from adult brain, knockdown of MEK1 and MEK2 in mutant mice is shown in comparison to control mice. β-ACTIN was used as a loading control. OB: olfactory bulb, HC: hippocampus, St: Striatum, fCx: frontal cortex, pCx: posterior cortex, Th: Thalamus, MB: midbrain, CB: cerebellum, BS: brainstem, Cx: cortex, Amy: amygdala, Hy: hypothalamus, +/flox: shMek^+/flox, +/Δ: shMek^+/flox/Nestin-cre; scale bar in D and E: 1 mm.