Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2016 Apr 7;18(4):541-53.
doi: 10.1016/j.stem.2016.01.022. Epub 2016 Mar 10.

CRISPR Interference Efficiently Induces Specific and Reversible Gene Silencing in Human iPSCs

Free PMC article

CRISPR Interference Efficiently Induces Specific and Reversible Gene Silencing in Human iPSCs

Mohammad A Mandegar et al. Cell Stem Cell. .
Free PMC article


Developing technologies for efficient and scalable disruption of gene expression will provide powerful tools for studying gene function, developmental pathways, and disease mechanisms. Here, we develop clustered regularly interspaced short palindromic repeat interference (CRISPRi) to repress gene expression in human induced pluripotent stem cells (iPSCs). CRISPRi, in which a doxycycline-inducible deactivated Cas9 is fused to a KRAB repression domain, can specifically and reversibly inhibit gene expression in iPSCs and iPSC-derived cardiac progenitors, cardiomyocytes, and T lymphocytes. This gene repression system is tunable and has the potential to silence single alleles. Compared with CRISPR nuclease (CRISPRn), CRISPRi gene repression is more efficient and homogenous across cell populations. The CRISPRi system in iPSCs provides a powerful platform to perform genome-scale screens in a wide range of iPSC-derived cell types, dissect developmental pathways, and model disease.


Figure 1
Figure 1. Generation and characterization of inducible CRISPRi and CRISPRn iPSCs
(A, B) Schematic overview of the strategy for TALEN-mediated targeting to the AAVS1 locus to generate the CRISPRi and CRISPRn iPSC lines. The doxycycline-controlled transcriptional activator (rtTA) is driven by a strong constitutive promoter (CAG). The third-generation doxycycline-response element (TRE3G) drives transcription of either Cas9 (CRISPRn) or dCas9-KRAB-P2A-mCherry (CRISPRi) and is oriented in the opposite direction of the transactivator to ensure no leaky expression without doxycycline treatment. (C, D) Immunostaining of CRISPRi and CRISPRn colonies before and after 48 h of doxycycline treatment with an antibody against Cas9 (green). Nuclei are stained with DAPI (blue). All nuclei showed expression of dCas9-KRAB or Cas9 after adding doxycycline. (E, G) Flow cytometry analysis of CRISPRi and CRISPRn iPSC lines before and after 48 h of doxycycline treatment. Doxycycline treatment of CRISPRi and CRISPRn produced expression of mCherry and FLAG in all cells, respectively. The doxycycline-untreated sample is plotted in grey. (F, H) CRISPRi and CRISPRn iPSC lines were treated with doxycycline (2 µM) for 24 h, which was then removed to measure the protein half-life of dCas9-KRAB and Cas9. Total protein was extracted from samples and analyzed by western blot with antibodies against Cas9 and GAPDH as a loading control. Both the CRISPRi and CRISPRn clones express dCas9-KRAB and Cas9 at similar levels after doxycycline treatment and the half-life of both proteins was approximately 12 h in iPSCs. Scale bars = 100 µm.
Figure 2
Figure 2. Comparison of the efficiency of CRISPRi knockdown and CRISPRn knockout
Immunostaining of representative (A) CRISPRi and (B) CRISPRn stable clones, each containing the same gRNA targeting the first exon of NANOG (NANOG g+358). After 7 days of doxycycline treatment, NANOG expression (green) was completely lost in all CRISPRi clones, but showed a variegated pattern of knockout in multiple independent CRISPRn clones. The mKate2 signal indicates the presence of the gRNA-expression vector in all cells within the clone. Nuclei are counterstained with DAPI. Western blot and flow cytometry analyses of (C, E) CRISPRi and (D, G) CRISPRn stable clones that contain the same gRNA against the first exon of NANOG. With CRISPRi, NANOG expression was uniformly decreased during doxycycline treatment and did not increase thereafter; however, with CRISPRn, the percentage of NANOG-positive cells fluctuates over the period of doxycycline treatment. Even after 12 days of continuous doxycycline treatment, ~30% of the population stained positive for NANOG. Genomic DNA was extracted from (F) CRISPRi and (H) CRISPRn stable lines containing a gRNA against NANOG before and after continuous doxycycline treatment for up to 17 days and subjected to sequencing. Red, out-of-frame INDELs; orange, in-frame INDELs; green, non-mutated alleles. Even after 12–17 days of continuous doxycycline treatment, 50–70% of sequenced alleles from CRISPRn contained no mutation, and 30–50% of mutated alleles were in-frame INDELs. No mutations were observed in either CRISPRi or CRISPRn without doxycycline, and the CRISPRi clones did not contain any mutations after doxycycline treatment. The total number of sequenced colonies is listed below each pie graph. Scale bars = 500 µm.
Figure 3
Figure 3. CRISPRi knockdown is efficient in iPSCs
(A) Efficiency of gRNA knockdown based on proximity to the transcription start site (TSS). The binding location of each gRNA is indicated relative to the TSS of the OCT4 locus and whether it targets the template (T) or non-template (NT) strand. Only gRNAs targeting near the TSS (approximately ± 150 bp) effectively knocked down OCT4. (B) TaqMan qPCR analysis of stable iPSCs containing gRNA against the gene of interest showed greater than 90% knockdown efficiency after 7 days of doxycycline induction in different endogenous genetic loci. (C) Immunostaining of stable clones containing a single gRNA against the gene of interest (OCT4, SOX2, NANOG, and BAG3). After 7 days of doxycycline treatment, there was a complete knockdown of the protein of interest (green). As expected, DAPI staining revealed that knocking down OCT4, NANOG and SOX2 resulted in loss of pluripotency and clear morphological changes. Also, knocking down BAG3 did not cause a loss of pluripotent morphology, as indicated by the distinct and round colony edges.
Figure 4
Figure 4. CRISPRi knockdown is reversible and tunable
A CRISPRi clone containing gRNA against the GCaMP transgene (GCaMP g+56) and endogenous BAG3 locus were used to test the knockdown efficiency and reversibility of the CRISPRi system in iPSCs. (A) Flow cytometry analysis of GCaMP expression showed that after 7 days of doxycycline induction, GCaMP was knocked down by ~99%, and was completely restored after doxycycline withdrawal for 14 days. (B) Using TaqMan qPCR, BAG3 transcript levels were knocked down to nearly undetectable levels, and expression was restored after doxycycline withdrawal. (C) Schematic diagram of the GCaMP-expression vector in which the GCaMP open reading frame (ORF) is driven off the CAG promoter. The locations of three gRNAs (g+24, g+56, and g+91) are schematically highlighted on the GCaMP ORF. The coordinates of GCaMP gRNA are based on the translation start site. pA, poly A signal. (D) Three stable CRISPRi colonies, each containing a different gRNA against GCaMP, were selected using blasticidin and cultured with doxycycline for 10 days. The percentage of GCaMP-positive cells for each gRNA-containing clone was plotted as a function of time based on flow cytometry analysis. Variable levels of GCaMP knockdown (~30%, ~50%, and ~99%) were achieved with different gRNA sequences. Error bars = standard deviation. N=1–3 technical replicates for each time point. (E) Flow cytometry plots of GCaMP fluorescence of stable CRISPRi clones on day 10 of doxycycline treatment. Using different gRNAs that target near the same region, variable levels of knockdown can be achieved. A scramble gRNA-containing CRISPRi and a GCaMP-negative iPSC population are displayed as controls. (F) Partial schematic diagram of the OCT4 locus marked with the location of the TSS and two gRNA-binding locations. Asterisk, a single-nucleotide polymorphism (SNP); green box, exon 1; grey box, 5’ untranslated region (5’ UTR). (G) Three stable CRISPRi colonies, two with different gRNAs against OCT4 and one with a scrambled control, were selected with blasticidin. Stable clones that contain either a scramble gRNA, a gRNA that targets a PAM sequence containing a SNP (OCT4 g−4), or a gRNA that does not target a SNP (OCT4 g+22) were treated with doxycycline. The percentage of the maximal median intensity of OCT4 staining for each gRNA-containing clone is plotted as a function of time by flow cytometry analysis. Complete loss of OCT4 expression (>98% knockdown) was observed after 7 days of doxycycline induction only when both alleles were targeted using OCT4 g+22. While using OCT4 g−4, which targets only one OCT4 allele (due to SNP in the PAM sequence), a gradual loss of OCT4 staining intensity is observed over time (down by ~40% by day 7). Error bars = standard deviation. N=1–3 technical replicates for each time point. (H) Flow cytometry plots of OCT4 staining on day 7 of doxycycline treatment. Dashed lines highlight the loss of OCT4-staining intensity (~40%) when using OCT4 g−4 compared to the scramble control. By targeting only one allele of OCT4, the OCT4-staining intensity homogeneously shifts (while remaining OCT4-positive), indicating that each cell experiences approximately the same level of knockdown. Note that the x-axis is a log-scale of OCT4 intensity. Differentiated iPSC-derived fibroblasts (OCT4 Cntrl) and a non-doxycycline-treated (–Dox) sample are displayed as controls.
Figure 5
Figure 5. RNA-Seq and TaqMan qPCR analysis
(A) RNA-sequencing RPM (reads per million) are plotted for CRISPRi cells stably expressing a gRNA targeting the GCaMP transgene (GCaMP g+56) cultured in the absence or presence of doxycycline. CRISPRi knockdown is specific to the GCaMP transcript, and few off-target transcriptional changes are observed. Data represent two independent biological replicates. (B) Heat map of TaqMan qPCR of stable clones containing a single gRNA against the gene of interest (OCT4, NANOG, and SOX2) as a function of days after doxycycline treatment. Analysis shows that by day 3, over 80% of the target transcript is depleted. Three housekeeping genes (18S, GAPDH, and UBC) were used to measure relative transcript levels. Each data point is an average of 2–4 technical replicates. TaqMan probes are listed in Table S5.
Figure 6
Figure 6. CRISPRi knockdown in differentiated cell types and cardiac-disease modeling
(A) Using CRISPRi, MESP1 was knocked down by ~90% in polyclonal cardiac progenitors, and MYBPC3 and HERG were knocked down by ~90% and 60% in polyclonal iPS-CMs, respectively. (B) Immunostaining of day-35 lactate-purified iPS-CMs stained with antibodies against MYBPC3 (green) and ACTN2 (red). Using CRISPRi knockdown, loss of MYBPC3 was observed in over 85% of analyzed cells in a polyclonal population. Nuclei were counterstained with DAPI. Scale bar = 100 µm. (C) Western blot of day-35 lactate-purified iPS-CMs with antibodies against MYBPC3, ACTN2, and GAPDH. Using CRISPRi, MYBPC3 protein was knocked down by ~90%. (D) GCaMP fluorescence in iPS-CMs containing gRNA against HERG and cultured in doxycycline (red). Recordings show a prolonged beat duration compared to untreated controls (green). (E) Quantified ratio of the downstroke-to-upstroke duration of doxycycline-treated iPS-CMs shows a significant difference in untreated iPS-CMs containing a gRNA against HERG, but not in iPS-CMs containing gRNA against OCT4 (negative control). (F) Patch-clamp recordings from single iPS-CMs show prolonged action potential durations in doxycycline-treated samples containing HERG gRNA.

Similar articles

See all similar articles

Cited by 92 articles

See all "Cited by" articles

Publication types

MeSH terms