66 128
stable through day 18 in culture (Figure S3B). Importantly, the
exogenous BAM factors and gRNAs were significantly depleted
by day 18 in culture after transient transfection (Figure S3C),
though levels of activation from the endogenous genes remained
high in cells treated with CR-BAM (Figure S3A).
Direct Activation via
VP64
dCas9
VP64
Rapidly Remodels
Chromatin at Target Loci
The kinetics of gene activation led us to speculate whether
the rapid and sustained elevated levels of endogenous gene
expression achieved with CR-BAM corresponded to an altered
epigenetic program at the target loci. We used chromatin
immunoprecipitation followed by next-generation sequencing
(ChIP-seq) data generated as part of the Encyclopedia of DNA
Elements (ENCODE) Project (Mouse ENCODE Consortium,
2012) to identify histone modifications enriched at the transcrip-
tionally active BAM factor loci in mouse embryonic brain tissue,
including H3K27ac and H3K4me3 (Figures 3A, 3C, and S4A). We
hypothesized that targeting the endogenous BAM factors for
activation with
VP64
dCas9
VP64
in PMEFs could recapitulate the
chromatin signatures found at these loci in developing brain
tissue.
To investigate the effects of BAM-factor induction on the
epigenetic programming at the target loci, we performed
chromatin immunoprecipitation (ChIP) qPCR in PMEFs trans-
duced with
VP64
dCas9
VP64
and transfected with pLuc, pBAM,
or CR-BAM plasmids (Figures 3 and S4). We used qPCR primers
tiled along intragenic and regulatory regions of the
Brn2, Ascl1,
and
Myt1l
loci. We detected a significant enrichment in
H3K27ac and H3K4me3 at the
Brn2
and
Ascl1
loci on day 3
post-transfection of CR-BAM (Figures 3B and 3D). H3K4me3
was enriched along the gene bodies of
Brn2
and
Ascl1
.
H3K27ac was enriched along the gene bodies and regions sur-
rounding the putative promoter sequences of both genes. In
contrast, targeted activation of
Myt1l
only induced modest
detectable enrichment in H3K27ac at the gRNA target sites
directly upstream of the first coding exon (Figure S4B). No signif-
icant change in H3K27ac or H3K4me3 was measured within the
putative
Myt1l
promoter. Though overexpression of the BAM
factors induced modest levels of expression of the endogenous
genes by day 3 post-transfection (Figures 1C and S3A), we
did not detect corresponding enrichment in H3K27ac and
H3K4me3 at the endogenous loci (Figures 3B, 3D, and S4B).
Generation of Induced Neuronal Cells with Multiplex
gRNA Lentiviral Vectors
To explore a strategy for stable expression of the CRISPR/Cas9
transcription factors, and to see if the same outcomes observed
with transient expression held true with constitutive expression,
we used a single lentiviral vector capable of expressing four
gRNAs from four independent RNA polymerase III promoters
(Kabadi et al., 2014) (Figure 4A). Co-transduction of lentiviruses
encoding
VP64
dCas9
VP64
and a set of four gRNAs targeting each
of the three BAM factors (lentiCR-BAM) permitted concurrent
activation of the endogenous BAM factors in PMEFs by day 6
post-transduction (Figure 4B). For comparison, we used lentiviral
vectors directly encoding theBAMfactors (lentiBAM), anddemon-
strated activation of the corresponding endogenous genes by day
6 post-transduction (Figure 4B). Similar to the results we obtained
with transient transfection of expression plasmids, targeted
activation of the endogenous genes via lentiviral delivery gener-
ated significantly more endogenous transcript from the
Brn2
and
Ascl1
loci than that induced through ectopic expression of
the BAM factors. However, unlike the transfection experiments,
endogenous
Myt1l
expression was significantly higher with trans-
duction of lentiBAM compared to lentiCR-BAM (Figure 4B).
Following extended culture for 2 weeks in neurogenic medium,
we readily identified Tuj1
+
Map2
+
cells with complex neuronal
morphologies (Figure 4C). All Tuj1
+
cells identified also co-ex-
pressed Map2. To promote further neuronal maturation and for
electrophysiological assessments, PMEFs were replated onto
a previously established monolayer of primary rat astrocytes
following transduction of
VP64
dCas9
VP64
and gRNAs (Vierbuchen
et al., 2010). Synapsin-RFP expression and cell morphology
were used to select the most mature neuronal cells for patch-
clamp analysis after 21 days in culture. In current-clamp mode,
single or multiple action potentials were readily elicited in
response to depolarizing current injections (six out of seven cells
analyzed; Figure 4D). The same cells displayed voltage-depen-
dent inward and outward currents. The transient inward currents
were abolished in the presence of the voltage-gated Na
+
channel
blocker tetrodotoxin (TTX; Figure 4E). The average resting mem-
brane potential, action potential (AP) threshold and AP amplitude
were 41 ± 3.8 mV, 33 ± 2.6 mV, and 49 ± 9.7 mV, respectively
(mean ± SEMs, n = 7 cells).
In contrast to what we observed by transient transfection of
the reprogramming factors, constitutive expression of the BAM
factor transgenes via lentiviral vectors generated significantly
more Tuj1
+
Map2
+
cells than that detected with
VP64
dCas9
VP64
(Figure 4F). We hypothesized that the prolonged and high levels
of expression of the BAM factor transgenes enabled by lentiviral
delivery permitted further epigenetic and transcriptional reprog-
ramming that improved the efficiency of iN generation when
compared to transient transfection methods. Consequently, we
revisited the analysis of chromatin remodeling at the endogenous
BAM factor loci in the context of lentiviral delivery of the
reprograming factors. We found that, as shown with transient
transfection, targeted activation of the endogenous genes via len-
tiCR-BAM transduction led to the rapid deposition of H3K27ac at
the
Brn2
and
Ascl1
loci as early as day 3 post-transduction that
persisted at day 6 (Figure 4G). Also, as seen with transient trans-
fection, we did not detect enrichment of H3K27ac at the
Myt1l
locus with lentiCR-BAM transduction, although we did measure
an increase in
Myt1l
mRNA (Figures 4B and 4G). In contrast to
what we observed with transient transfection of the BAM factors,
we detected enrichment of H3K27ac along regions of all three
endogenous genes with lentiBAM transduction (Figure 4G).
Furthermore, we only detected minor enrichment in H3K27ac at
all three genes at day 3 post-transduction of lentiBAM; however,
both
Ascl1
and
Myt1l
showed a substantial increase in H3K27ac
deposition by day 6 post-transduction (Figure 4G).
DISCUSSION
In this study, we demonstrate direct cellular reprogramming
to induced neuronal cells through targeted activation of
endogenous genes. We utilized the CRISPR/Cas9 system as a
programmable, locus-specific transcriptional regulator for the
410
Cell Stem Cell
19
, 406–414, September 1, 2016