ARTICLES
In vitro reprogramming of fibroblasts into a
pluripotent ES-cell-like state
Marius Wernig1*, Alexander Meissner1*, Ruth Foreman1,2*, Tobias Brambrink1*, Manching Ku3*,
Konrad Hochedlinger1{, Bradley E. Bernstein3,4,5 & Rudolf Jaenisch1,2
Nuclear transplantation can reprogramme a somatic genome back into an embryonic epigenetic state, and the
reprogrammed nucleus can create a cloned animal or produce pluripotent embryonic stem cells. One potential use of the
nuclear cloning approach is the derivation of ‘customized’ embryonic stem (ES) cells for patient-specific cell treatment, but
technical and ethical considerations impede the therapeutic application of this technology. Reprogramming of fibroblasts to
a pluripotent state can be induced in vitro through ectopic expression of the four transcription factorsOct4 (also calledOct3/
4 or Pou5f1), Sox2, c-Myc and Klf4. Here we show that DNA methylation, gene expression and chromatin state of such
induced reprogrammed stem cells are similar to those of ES cells. Notably, the cells—derived from mouse fibroblasts—can
form viable chimaeras, can contribute to the germ line and can generate live late-term embryos when injected into tetraploid
blastocysts. Our results show that the biological potency and epigenetic state of in-vitro-reprogrammed induced pluripotent
stem cells are indistinguishable from those of ES cells.
Epigenetic reprogramming of somatic cells into ES cells has attracted
much attention because of the potential for customized transplanta-
tion therapy, as cellular derivatives of reprogrammed cells will not be
rejected by the donor1,2. Thus far, somatic cell nuclear transfer and
fusion of fibroblasts with ES cells have been shown to promote the
epigenetic reprogramming of the donor genome to an embryonic
state3–5. However, the therapeutic application of either approach has
been hindered by technical complications as well as ethical objec-
tions6. Recently, a major breakthrough was reported whereby expres-
sion of the transcription factors Oct4, Sox2, c-Myc and Klf4 was
shown to induce fibroblasts to become pluripotent stem cells (desig-
nated as induced pluripotent stem (iPS) cells), although with a low
efficiency7. The iPS cells were isolated by selection for activation of
Fbx15 (also called Fbxo15), which is a downstream gene of Oct4. This
important study left a number of questions unresolved: (1) although
iPS cells were pluripotent they were not identical to ES cells (for
example, iPS cells injected into blastocysts generated abnormal chi-
maeric embryos that did not survive to term); (2) gene expression
profiling revealedmajor differences between iPS cells and ES cells; (3)
because the four transcription factors were transduced by constitu-
tively expressed retroviral vectors it was unclear why the cells could be
induced to differentiate and whether continuous vector expression
was required for the maintenance of the pluripotent state; and (4)
the epigenetic state of the endogenous pluripotency genes Oct4 and
Nanog was incompletely reprogrammed, raising questions about the
stability of the pluripotent state.
Here we used activation of the endogenousOct4 orNanog genes as
a more stringent selection strategy for the isolation of reprogrammed
cells. We infected fibroblasts with retroviral vectors transducing the
four factors, and selected for the activation of the endogenousOct4 or
Nanog genes. Positive colonies resembled ES cells and assumed an
epigenetic state characteristic of ES cells. When injected into blasto-
cysts the reprogrammed cells generated viable chimaeras and
contributed to the germ line. Our results establish that somatic cells
can be reprogrammed to a pluripotent state that is similar, if not
identical, to that of normal ES cells.
Selection of fibroblasts for Oct4 or Nanog activation
Using homologous recombination in ES cells we generated mouse
embryonic fibroblasts (MEFs) and tail-tip fibroblasts (TTFs) that
carried a neomycin-resistance marker inserted into either the endo-
genousOct4 (Oct4-neo) orNanog locus (Nanog-neo) (Fig. 1a). These
cultures were sensitive to G418, indicating that the Oct4 and Nanog
loci were, as expected, silenced in somatic cells. These MEFs or TTFs
were infected with Oct4-, Sox2-, c-Myc- and Klf4-expressing retro-
viral vectors and G418 was added to the cultures 3, 6 or 9 days later.
The number of drug-resistant colonies increased substantially when
analysed at day 20 (Fig. 1i). Most colonies had a flat morphology
(Fig. 1h, right) and between 11% and 25% of the colonies were ‘ES-
like’ (Fig. 1h, left) when selection was applied early (Fig. 1k), a per-
centage that increased at later time points. At day 20, ES-like colonies
were picked, dissociated and propagated in G418-containing media.
They gave rise to ES-like cell lines (designated as Oct4 iPS or Nanog
iPS cells, respectively) that could be propagated without drug selec-
tion, displayed homogenous Nanog, SSEA1 and alkaline phosphatase
expression (Fig. 1b–g and Supplementary Figs 1 and 5), and formed
undifferentiated colonies when seeded at clonal density on gelatin-
coated dishes (see inset in Fig. 1b). Four out of five analysed lines had a
normal karyotype (Supplementary Table 1).
Although the timing and appearance of colonies were similar
between the Oct4 and Nanog selection, we noticed pronounced
quantitative differences between the two selection strategies: whereas
Oct4-selected MEF cultures had 3- to 10-fold fewer colonies, the
fraction of ES-like colonies was 2- to 3-fold higher than in Nanog-
selected cultures. Accordingly, approximately four times more Oct4-
selected ES-like colonies gave rise to stable and homogenous iPS cell
*These authors contributed equally to this work.
1Whitehead Institute for Biomedical Research and 2Department of Biology,Massachusetts Institute of Technology, Cambridge,Massachusetts 02142,USA. 3Molecular PathologyUnit
and Center for Cancer Research, Massachusetts General Hospital, Charlestown,Massachusetts 02129, USA. 4Broad Institute of Harvard andMIT, Cambridge, Massachusetts 02142,
USA. 5Department of Pathology, HarvardMedical School, Boston,Massachusetts 02115, USA. {Present address: Center for RegenerativeMedicine and Cancer Center,Massachusetts
General Hospital, Harvard Medical School and Harvard Stem Cell Institute, Boston, Massachusetts 02414, USA.
Vol 448 | 19 July 2007 |doi:10.1038/nature05944
318
Nature ©2007 Publishing Group

lines compared with Nanog-selected ES-like colonies (Fig. 1k). This
suggests that although theNanog locus was easier to activate, a higher
fraction of the drug-resistant colonies inOct4-neo cultures was repro-
grammed to a pluripotent state. Therefore, the overall estimated effi-
ciency of 0.05–0.1% to establish iPS cell lines from MEFs was similar
between Oct4 selection and Nanog selection, despite the larger num-
ber of total Nanog-neo resistant colonies (Fig. 1k). Next we investi-
gated the time course of reprogramming by studying the fraction of
alkaline-phosphatase-, SSEA1- and Nanog-positive cells in Oct4-
selected MEF cultures. Fourteen days after infection some cells had
already initiated alkaline phosphatase activity and SSEA1 expression,
but lacked detectable amounts of Nanog protein (Fig. 1j), whereas by
day 20, alkaline phosphatase and SSEA1 expression had increased and
,8% of the cells were Nanog-positive. Thus, the reprogramming
induced by the four transcription factors (Oct4, Sox2, c-Myc and
Klf4) is a gradual and slow process.
Expression and DNA methylation
To characterize the reprogrammed cells on a molecular level we used
quantitative polymerase chain reaction with reverse transcription
(qRT–PCR) to measure the expression of ES-cell- and fibroblast-
specific genes. Figure 2a shows that in Oct4 iPS cells the total level
of Nanog and Oct4 was similar to that in ES cells but decreased on
differentiation to embryoid bodies.MEFs did not express either gene.
Using specific primers for endogenous or total Sox2 transcripts
showed that most Sox2 transcripts originated from the endogenous
locus rather than the viral vector (Fig. 2b). In contrast, Hoxa9 and
Zfpm2 were highly expressed in MEFs but were expressed at very low
levels in iPS or ES cells (Fig. 2c). Western blot analysis showed that
multiple iPS clones expressed Nanog and Oct4 proteins at similar
levels compared to ES cells (Fig. 2d). Finally, we used microarray
technology to compare gene expression patterns on a global level.
Figure 2f shows that the iPS cells clustered with ES cells in contrast to
wild-type or donor MEFs.
To investigate the DNA methylation level of the Oct4 and Nanog
promoters we performed bisulphite sequencing and combined bisul-
phite restriction analysis (COBRA) with DNA isolated from ES cells,
iPS cells and MEFs. As shown in Fig. 2g, both loci were demethylated
in ES and iPS cells and fully methylated in MEFs. To assess whether
the maintenance of genomic imprinting was compromised we
NanogDAPI
d e
f g
DAPI SSEA1
a
j
0
20
40
60
80
Day 0
Total col./
100,000
cells
ES-like col./
total col.
(%)
iPS line/
ES-like col.
(%)
Estimated
efficiency
(%)
Oct4
Nanog
156 ± 31 24.0 ± 7 22
5 0.050
0.080
11.5 ± 4947 ± 187
Day 14 Day 20
AP
SSEA1
Nanog
i
h
k
ES-like colony Flat colony
0
100
200
300
400
500
Day 3 Day 6 Day 9
Oct4-neo
Nanog-neo
Bcll Ncol
IRES GFP neo
1 kb
Ncol
Ncol NcolNcol
5′probe
To
ta
l c
ol
on
y
nu
m
be
r
P
er
ce
nt
ag
e
of
to
ta
l
se
le
ct
ed
c
el
ls
b c
Figure 1 | Generation of Oct4- and Nanog-selected iPS cells. a, Targeting
strategy to generate an Oct4-IRES-GFPneo allele. The resulting GFPneo
fusion protein has sufficient neomycin-resistance activity in ES cells; GFP
fluorescence, however, is not visible. b, Phase-contrast micrograph of Oct4
iPS cells (clone 18) grown on irradiated MEFs. Inset: an ES-cell-like colony
5 days after seeding in clonal density without feeder cells. iPS clone 18 cells
exhibited strong alkaline phosphatase activity (c) and were homogenously
labelled with antibodies against SSEA1 (d, e) and Nanog (f, g). h, One
example of an ES-like colony 16 days after infection (left). Most G418-
resistant colonies, however, consisted of flat non-ES-like cells (right): b, 103;
c–g, 203; h, 43. i, Gradual activation of the Nanog and Oct4-neo alleles.
Shown are the total colony numbers of one experiment at day 20 after
infection starting neomycin selection at day 3, 6 and 9. j, Fraction of total
selected cells expressing alkaline phosphatase, SSEA1 and Nanog 0, 14, and
20 days after infection (counted were more than ten visual fields containing
n. 1,000 total cells for every time point; error bars indicate s.d.).
k, Estimated reprogramming efficiency of Oct4 selection and Nanog
selection (n5 3 different experiments; s.e.m. is shown). Indicated are the
total number of drug-resistant colonies per 100,000 plated MEFs 20 days
after infection; the fraction of ES-like colonies per total number of colonies;
the fraction of iPS cell lines that could be established from picked ES-like
colonies as defined by homogenous alkaline phosphatase, SSEA1 andNanog
expression. After determining the fraction of Sox2- (83.4%), Oct4- (53.2%)
and c-myc- (46.3%) infected MEFs 2 days after infection by
immunofluorescence and assuming 50% were infected by Klf4 viruses, we
estimated the overall reprogramming efficiency as the ratio of quadruple-
infected cells and the extrapolated total number of iPS cell lines that could be
established with G418 selection starting at day 6 after infection.
NATURE |Vol 448 | 19 July 2007 ARTICLES
319
Nature ©2007 Publishing Group

genes reported to be bivalent in pluripotent ES cells13. Figure 3a
shows that the fibroblast-specific genes Zfpm2 and Hoxa9 carried
stronger H3K4 methylation than H3K27 methylation in the donor
MEFs, whereas the silent genes Nkx2.2, Sox1, Lbx1 and Pax5 prim-
arily carried H3K27methylation. In contrast, in the Oct4 iPS cells, all
of these genes showed comparable enrichment for both histonemod-
ifications, similar to normal ES cells (Fig. 3a). Identical results were
obtained in Nanog iPS clones selected from Nanog-neo MEFs
(Supplementary Fig. 2). These data suggest that the chromatin con-
figuration of somatic cells is re-set to one that is characteristic of ES
cells.
iPS cells tolerate genomic demethylation
Tolerance of genomic demethylation is a unique property of ES cells
in contrast to somatic cells, which undergo rapid apoptosis on loss of
the DNA methyltransferase Dnmt1 (refs 15–17). We investigated
whether iPS cells would be resistant to global demethylation after
Dnmt1 inhibition and would be able to re-establish global methyla-
tion patterns after restoration of Dnmt1 activity. To this end, we used
a conditional lentiviral vector harbouring a Dnmt1-targeting short
hairpin (sh)RNA and a green fluorescence protein (GFP) reporter
gene (Fig. 3b and ref. 18). Infected iPS cells were plated at low density
and GFP-positive colonies were picked and expanded. Southern blot
analysis using HpaII-digested genomic DNA showed that global
demethylation of infected iPS cells (Fig. 3c, lanes 6, 7) was similar
to Dnmt12/2 ES cells (lane 2). In contrast, uninfected iPS cells or
MEFs (lanes 4, 5) displayed normal methylation levels. Morph-
ologically, the GFP-positive cells were indistinguishable from the par-
ental line or from uninfected sister subclones, indicating that iPS cells
tolerate global DNA demethylation. In a second step, the Dnmt1
shRNA was excised through Cre-mediated recombination and GFP-
negative cloneswere picked (Fig. 3d). The cells had excised the shRNA
vector (Fig. 3e) and normal DNA methylation levels were restored
(Fig. 3c, lane 8) and were able to generate chimaeras (see below,
Table 1), as has been reported previously for ES cells19. These observa-
tions imply that the de novo methyltransferases Dnmt3a and Dnmt3b
were reactivated in iPS cells20, leading to restoration of global methy-
lation levels. As expected19, the imprinted genes Snrpn and Peg3 were
unmethylated and resistant to remethylation (Fig. 3f).
Maintenance of the pluripotent state
Southern blot analysis indicated that Oct4-neo iPS clone 18 carried
four to six copies of the Oct4, c-myc and Klf4 retroviral vectors and
only one copy of the Sox2 retroviral vector (Fig. 4a). Because these
four factors were under the control of the constitutively expressed
retroviral long terminal repeat, it was unclear in a previous study why
iPS cells could be induced to differentiate7. To address this question,
we designed primers specific for the four viral-encoded transcription
factor transcripts and compared expression levels by qRT–PCR in
MEFs 2 days after infection in iPS cells, in embryoid bodies derived
from iPS cells, and in demethylated and remethylated iPS cells
(Fig. 4b). Although the MEFs represented a heterogeneous popu-
lation composed of uninfected and infected cells, virally encoded
RNA levels of Oct4, Sox2 and Klf4 RNA were 5-fold higher and of
c-myc more than 10-fold higher than in iPS cells. This suggests silen-
cing of the viral long terminal repeat by de novo methylation during
the reprogramming process. Accordingly, the total Sox2 and Oct4
RNA levels in iPS cells were similar to those in wild-type ES cells,
and the Sox2 transcripts in iPS cells were mostly, if not exclusively,
transcribed from the endogenous gene (compare Fig. 2b). On differ-
entiation to embryoid bodies, both viral and endogenous transcripts
were downregulated. All viral Sox2, Oct4 and Klf4 transcripts were
upregulated by approximately twofold in Dnmt1 knockdown iPS
cells, and again downregulated on restoration of Dnmt1 activity.
This is consistent with previous data that Moloney virus is efficiently
de novo methylated and silenced in embryonic but not in somatic
cells21,22. Transcript levels of c-myc were about 20-fold lower in iPS
cells than in infected MEFs, and did not change on differentiation or
demethylation.
To follow the kinetics of vector inactivation during the reprogram-
ming process, we isolated RNA from drug-resistant cell populations
at different times after infection. Figure 4c shows that the viral-vector-
encoded transcripts were gradually silenced during the transition
from MEFs to iPS cells with a time course that corresponded to
the gradual appearance of pluripotency markers (compare Fig. 1j).
Finally, to visualize directly Oct4 and Nanog expression during dif-
ferentiation, we injected Oct4 iPS cells into SCID mice to induce
teratoma formation (Fig. 4d). Immunostaining revealed that Oct4
and Nanog were expressed in the centrally located undifferentiated
cells but were silenced in the differentiated parts of the teratoma
(Fig. 4e, f). Our results suggest that the retroviral vectors are subject
to gradual silencing by de novo methylation during the reprogram-
ming process. Themaintenance of the pluripotent state and induction
of differentiation strictly depends on the expression and normal regu-
lation of the endogenous Oct4 and Nanog genes.
Developmental potency
We determined the developmental potential of iPS cells by teratoma
and chimaera formation. Histological and immunohistochemical
analysis of Oct4- or Nanog-iPS-cell-induced teratomas revealed that
the cells had differentiated into cell types representing all three
embryonic germ layers (Supplementary Figs 3 and 4). To assess more
stringently their developmental potential, various iPS cell lines were
injected into diploid (2N) or tetraploid (4N) blastocysts. After injec-
tion into 2N blastocysts both Nanog iPS and Oct4 iPS clones derived
from MEFs (Fig. 5a) or from TTFs (Fig. 5b,c), as well as iPS cells that
had been subjected to a consecutive cycle of demethylation and
remethylation (compare Fig. 3b, c), efficiently generated viable
Table 1 | Summary of blastocyst infections
2N injections 4N injections
Cell line Injected
blastocysts
Live chimaeras Chimaerism
(%)
Germ line Injected
blastocysts
Dead
embryos (arrested)
Live
embryos (analysed)
O6 ND ND ND ND 13 0 2 (E12.5)
O9 30 5 30–70 Yes 90 3 (E11–13.5) 12 (E10–12.5)
O16 15 3 10–30 Yes ND ND ND
O18 95 8 5–50 No 134 7 (E9–11.5) 4* (E10–12.5)
O3-2 ND ND ND ND 25 2 (E8,11.5) 0
O4-16 ND ND ND ND 35 4 (E11–13.5) 3 (E14.5)
N7 30 1 30 ND ND ND ND
N8 90 14 5–50 No 118 9 (E9–11.5) 1* (E12.5)
N14 30 5 5–20 ND 46 2 (E8,11.5) 1 (E12.5)
TT-O25 50 2 30{ ND 39 3 (E9.5) 0
O18 rem/3.1 25 1 30 ND ND ND ND
The extent of chimaerism was estimated on the basis of coat colour or EGFP expression. ND, not determined. 4N injected blastocysts were analysed between embryonic day E10.5 and E14.5.
‘Analysed’ indicates the day of embryonic development analysed; ‘arrested’ indicates the estimated stage of development of dead embryos.
* Developmentally retarded or abnormal. O18 rem/3.1 is a de- and remethylated iPS clone (Fig. 3c).
{On the basis of GFP fluorescence.
NATURE |Vol 448 | 19 July 2007 ARTICLES
321
Nature ©2007 Publishing Group

high-contribution chimaeras (summarized in Table 1). To test for
germline transmission, chimaeras derived from two different iPS
lines (Oct4 iPS O9 and O16) were mated with normal females, and
blastocysts were isolated and genotyped by three different PCR reac-
tions for the presence of the multiple viral Oct4 and c-myc genes and
for the single-copy GFPneo sequences inserted into the Oct4 locus of
the donor cell (Fig. 1a). Figure 5f shows that 9 out of 16 embryos from
two chimaeras were positive for the viral copies. As expected, only
half of the viral-positive blastocysts contained the GFPneo sequences
(5 out of 9 embryos, Fig. 5f, left panel). When embryonic day (E)10
embryos derived from an Oct4 iPS line O16 chimaera were geno-
typed, three out of eight tested embryos were transgenic (Fig. 5f, right
panel). Finally, we injected iPS cells into 4N blastocysts as this repre-
sents the most rigorous test for developmental potency, because the
resulting embryos are composed only of the injected donor cells (‘all
ES embryo’). Figure 5d, e shows that both Oct4 and Nanog iPS cells
could generate mid- and late-gestation ‘all iPS embryos’ (summar-
ized in Table 1). These findings indicate that iPS cells can establish all
lineages of the embryo and thus have a similar developmental poten-
tial as ES cells.
Discussion
The results presented here demonstrate that the four transcription
factors Oct4, Sox2, c-myc and Klf4 can induce epigenetic reprogram-
ming of a somatic genome to an embryonic pluripotent state. In
contrast to selection for Fbx15 activation7, fibroblasts that had reac-
tivated the endogenous Oct4 (Oct4-neo) or Nanog (Nanog-neo) loci
grew independently of feeder cells, expressed normal Oct4, Nanog
and Sox2 RNA and protein levels, were epigenetically identical to ES
cells by a number of criteria, and were able to generate viable chi-
maeras, contribute to the germ line and generate viable late-gestation
embryos after injection into tetraploid blastocysts. Transduction of
the four factors generated significantlymore drug-resistant cells from
Nanog-neo than from Oct4-neo fibroblasts but a higher fraction of
Oct4-selected cells had all the characteristics of pluripotent ES cells,
suggesting that Nanog activation is a less stringent criterion for plur-
ipotency than Oct4 activation.
Our data suggest that the pluripotent state of Oct4 iPS and Nanog
iPS cells is induced by the virally transduced factors but is largely
maintained by the activity of the endogenous pluripotency factors
including Oct4, Nanog and Sox2, because the viral-controlled
transcripts, although expressed highly in MEFs, become mostly
silenced in iPS cells. The total levels of Oct4, Nanog and Sox2 were
similar in iPS and wild-type ES cells. Consistent with the conclusion
that the pluripotent state is maintained by the endogenous pluripo-
tency genes is the finding that the Oct4 and the Nanog genes became
hypomethylated in iPS cells as in ES cells, and that the bivalent his-
tone modifications of developmental regulators were re-established.
Furthermore, iPS cells were resistant to global demethylation
induced by inactivation of Dnmt1, similar to ES cells but in contrast
to somatic cells. Re-expression of Dnmt1 in the hypomethylated ES
cells resulted in global remethylation, indicating that the iPS cells had
also reactivated the de novomethyltransferases Dnmt3a andDnmt3b.
All these observations are consistent with the conclusion that the iPS
cells have gained an epigenetic state that is similar to that of normal
ES cells. This conclusion is further supported by the recent obser-
vation that female iPS cells, similar to ES cells, reactivate the soma-
tically silenced X chromosome23.
Expression of the four transcription factors proved to be a robust
method to induce reprogramming of somatic cells to a pluripotent
state. However, the use of retrovirus-transduced oncogenes repre-
sents a serious barrier to the eventual use of reprogrammed cells for
therapeutic application. Much work is needed to understand the
molecular pathways of reprogramming and to eventually find small
0
20
40
60
80
100
120
140
a
Oct4 c-Myc Klf4 Sox2
*
*
*
* *
* O18
de
m/
1
O1
8 d
em
/3EB
s
2 d
ay
infe
cte
d-M
EF
s
ME
Fs
O1
8.1O1
8
R
el
at
iv
e
ex
pr
es
si
on 1.0
0.8
0.6
0.4
0.2
b
c
Ex
pr
es
si
on
p
er
ce
nt
ag
e
of
G
ap
dh
d4 d9 d18 d28 d32 O1.3 O9
Oct4
Sox2
Klf4
c-myc
d
H&E Nanog Oct4
e f
O1
8 r
em
/3.
1
Figure 4 | Efficient silencing of retroviral transcripts in induced pluripotent
cells. a, Southern blot analysis of proviral integrations in iPS cloneO18 (left
lanes) for the four retroviral vectors. Uninfected ES cells (right lanes) show
only one or two bands corresponding to the endogenous gene (marked by an
asterisk). b, Quantitative RT–PCR using primers specifically detecting the
four viral transcripts. Shown are Oct4-neo iPS clone 18 and a GFP-labelled
subclone, Oct4-neo MEFs, 2-week-old embryoid bodies generated from
clone 18, two demethylated clones (18 dem/1 and 18 dem/3), a remethylated
clone (18 rem/3.1), and Oct4-neo MEFs 2 days after infection with all four
viruses but not selected with G418 (n5 3 independent experiments; error
bars indicate s.d.). c, Viral transcript levels at various time points in cell
populations after infection and Oct4 selection and in the two Oct4 iPS cell
lines O1.3 and O9 (n5 3 independent experiments; error bars indicate s.d.).
d–f, Paraffin sections of a teratoma 26 days after subcutaneous injection of
Oct4 iPS clone 18 cells into SCID mice. H&E, haematoxylin and eosin.
Nanog (e) and Oct4 (f) expression was confined to undifferentiated cell
types as indicated an immunohistochemical analysis.
ARTICLES NATURE |Vol 448 | 19 July 2007
322
Nature ©2007 Publishing Group

molecules that could achieve reprogramming without gene transfer
of potentially harmful genes.
METHODS SUMMARY
Cell culture, gene targeting and viral infections.ES and iPS cells were cultivated
on irradiated MEFs. Using homologous recombination we generated ES cells
carrying an IRES-GFPneo fusion cassette downstream of Oct4 exon 5 (Fig. 1a).
The Nanog gene was targeted as described24. Transgenic MEFs were isolated and
selected from E13.5 chimaeric embryos after blastocyst injection of Oct4-IRES-
GFPneo- or Nanog-neo-targeted ES cells.MEFs were infected overnight with the
Moloney-based retroviral vector pLIB (Clontech) containing the murine com-
plementary DNAs of Oct4, Sox2, Klf4 and c-myc.
Southern blot, methylation and chromatin analyses. To assess the levels of
DNA methylation, genomic DNA was digested with HpaII and hybridized to a
probe for the minor satellite repeats25 or with an IAP probe26. Bisulphite treat-
ment was performed with the Qiagen EpiTect Kit. For the methylation status of
Oct4 and Nanog promoters, bisulphite sequencing analysis was performed as
described previously27. For imprinted genes, a COBRA assay was performed.
PCR primers and conditions were as described previously28. The status of biva-
lent domains was determined by chromatin immunoprecipitation followed by
quantitative PCR analysis, as described previously12
Expression analysis. Total RNA was reverse-transcribed and quantified using
the QuantTtect SYBR green RT–PCR Kit (Qiagen) on a 7000 ABI detection
system. Western blot and immunofluorescence analysis was performed as
described29,30. Microarray targets from 2mg total RNA were synthesized and
labelled using the Low RNA Input Linear Amp Kit (Agilent), hybridized to
Agilent whole-mouse genome oligonucleotide arrays (G4122F) and analysed
as previously described31.
Full Methods and any associated references are available in the online version of
the paper at www.nature.com/nature.
Received 27 February; accepted 22 May 2007.
Published online 6 June 2007.
1. Hochedlinger, K. & Jaenisch, R. Nuclear transplantation, embryonic stem cells,
and the potential for cell therapy. N. Engl. J. Med. 349, 275–286 (2003).
2. Yang, X. et al. Nuclear reprogramming of cloned embryos and its implications for
therapeutic cloning. Nature Genet. 39, 295–302 (2007).
3. Hochedlinger, K. & Jaenisch, R. Nuclear reprogramming and pluripotency. Nature
441, 1061–1067 (2006).
4. Tada, M., Takahama, Y., Abe, K., Nakatsuji, N. & Tada, T. Nuclear reprogramming
of somatic cells by in vitro hybridization with ES cells. Curr. Biol. 11, 1553–1558
(2001).
5. Cowan, C. A., Atienza, J., Melton, D. A. & Eggan, K. Nuclear reprogramming of
somatic cells after fusion with human embryonic stem cells. Science 309,
1369–1373 (2005).
6. Jaenisch, R. Human cloning—the science and ethics of nuclear transplantation. N.
Engl. J. Med. 351, 2787–2791 (2004).
7. Takahashi, K. & Yamanaka, S. Induction of pluripotent stem cells from mouse
embryonic and adult fibroblast cultures by defined factors. Cell 126, 663–676
(2006).
8. Labosky, P. A., Barlow, D. P. & Hogan, B. L. Mouse embryonic germ (EG) cell lines:
transmission through the germline and differences in the methylation imprint of
insulin-like growth factor 2 receptor (Igf2r) gene compared with embryonic stem
(ES) cell lines. Development 120, 3197–3204 (1994).
9. Boyer, L. A. et al. Core transcriptional regulatory circuitry in human embryonic
stem cells. Cell 122, 947–956 (2005).
10. Loh, Y. H. et al. The Oct4 andNanog transcription network regulates pluripotency
in mouse embryonic stem cells. Nature Genet. 38, 431–440 (2006).
11. Lee, T. I. et al. Control of developmental regulators by Polycomb in human
embryonic stem cells. Cell 125, 301–313 (2006).
12. Boyer, L. A. et al. Polycomb complexes repress developmental regulators in
murine embryonic stem cells. Nature 441, 349–353 (2006).
13. Bernstein, B. E. et al. A bivalent chromatin structure marks key developmental
genes in embryonic stem cells. Cell 125, 315–326 (2006).
14. Azuara, V. et al. Chromatin signatures of pluripotent cell lines. Nature Cell Biol. 8,
532–538 (2006).
15. Jackson-Grusby, L. et al. Loss of genomic methylation causes p53-dependent
apoptosis and epigenetic deregulation. Nature Genet. 27, 31–39 (2001).
16. Li, E., Bestor, T. H. & Jaenisch, R. Targetedmutation of theDNAmethyltransferase
gene results in embryonic lethality. Cell 69, 915–926 (1992).
17. Meissner, A. et al. Reduced representation bisulfite sequencing for comparative
high-resolution DNA methylation analysis. Nucleic Acids Res. 33, 5868–5877
(2005).
18. Ventura, A. et al. Cre-lox-regulated conditional RNA interference from
transgenes. Proc. Natl Acad. Sci. USA 101, 10380–10385 (2004).
19. Holm, T. M. et al. Global loss of imprinting leads to widespread tumorigenesis in
adult mice. Cancer Cell 8, 275–285 (2005).
20. Okano, M., Bell, D. W., Haber, D. A. & Li, E. DNA methyltransferases Dnmt3a and
Dnmt3b are essential for de novo methylation and mammalian development. Cell
99, 247–257 (1999).
21. Stewart, C. L., Stuhlmann, H., Ja¨hner, D. & Jaenisch, R. De novo methylation,
expression, and infectivity of retroviral genomes introduced into embryonal
carcinoma cells. Proc. Natl Acad. Sci. USA 79, 4098–4102 (1982).
22. Ja¨hner, D. et al. De novo methylation and expression of retroviral genomes during
mouse embryogenesis. Nature 298, 623–628 (1982).
23. Maherali, N. et al. Global epigenetic remodeling in directly reprogrammed
fibroblasts. Cell Stem Cells (in the press)..
a
b c
d e
O16 O9-1 O9-2 + –
Oct4 virus
c-myc virus
Oct4-GFPneo
allele
f
Blastocysts E10 embryos
O16
Figure 5 | Developmental pluripotency of reprogrammed fibroblasts. a, A
6-week-old chimaeric mouse. Agouti-coloured hairs originated from Oct4
iPS cell line O18.1. b, c, Two live pups after 2N blastocyst injection, one of
which shows a high contribution (c) of the TTF-derived Oct4 iPS cell line
TT-O25, which had been GFP-labelled with a lentiviral ubiquitin-EGFP
vector. d, ‘All iPS cell embryos’ were generated by injection of iPS cells into
4N blastocysts34. Live E12.5 embryos generated fromOct4 iPS line O6 (left),
fromNanog iPS line N14 (middle) and fromV.6.5 ES cells (right) are shown.
e, A normally developed E14.5 embryo was derived from Oct4 iPS cell line
O4-16 after tetraploid complementation andwas isolated by screeningMEFs
for activation of GFP inserted into the Oct4 locus. f, Germline contribution
of Oct4 iPS clones O9 and O16. Genotyping of blastocysts from females
mated with three chimaeric males demonstrated the presence of Oct4 and
c-myc virus integrations and the Oct4-IRES-GFPneo allele (left panel).
Because of the multiple integrations (Fig. 4a) all embryos with iPS cell
contribution are expected to be positive for proviral sequences in this assay.
In contrast, the single-copy Oct4-IRES-GFPneo allele segregated into only 5
of the 9 virus-positive embryos. All six blastocysts from O9 chimaera 1 were
iPS-cell-derived, suggesting that this chimaera was a pseudo-male.
Additional genotyping identified 13 out of 72 tested blastocysts derived from
iPS line O9 and 4 out of 13 blastocysts derived from iPS line O16 chimaeras
carrying the viral transgenes. The right panel shows that 3 out of 8 tested E.10
mid-gestation embryos were sired by a chimaera derived from the donor iPS
line O16. 1, positive control; 2, negative control.
NATURE |Vol 448 | 19 July 2007 ARTICLES
323
Nature ©2007 Publishing Group

24. Mitsui, K. et al. The homeoprotein Nanog is required for maintenance of
pluripotency in mouse epiblast and ES cells. Cell 113, 631–642 (2003).
25. Chapman, V., Forrester, L., Sanford, J., Hastie, N. & Rossant, J. Cell lineage specific
undermethylation of mouse repetitive DNA. Nature 307, 284–286 (1984).
26. Walsh, C. P., Chaillet, J. R. & Bestor, T. H. Transcription of IAP endogenous
retroviruses is constrainedby cytosinemethylation.NatureGenet.20, 116–117 (1998).
27. Blelloch, R. et al. Reprogramming efficiency following somatic cell nuclear transfer
is influenced by the differentiation and methylation state of the donor nucleus.
Stem Cells 24, 2007–2013 (2006).
28. Lucifero, D., Mertineit, C., Clarke, H. J., Bestor, T. H. & Trasler, J. M. Methylation
dynamics of imprinted genes inmouse germ cells.Genomics 79, 530–538 (2002).
29. Hochedlinger, K., Yamada, Y., Beard, C. & Jaenisch, R. Ectopic expression of Oct-4
blocks progenitor-cell differentiation and causes dysplasia in epithelial tissues.
Cell 121, 465–477 (2005).
30. Wernig,M. et al. Functional integration of embryonic stem cell-derived neurons in
vivo. J. Neurosci. 24, 5258–5268 (2004).
31. Brambrink, T., Hochedlinger, K., Bell, G. & Jaenisch, R. ES cells derived from cloned
and fertilized blastocysts are transcriptionally and functionally indistinguishable.
Proc. Natl Acad. Sci. USA 103, 933–938 (2006).
32. Eads, C. A. & Laird, P. W. Combined bisulfite restriction analysis (COBRA).
Methods Mol. Biol. 200, 71–85 (2002).
33. Peitz, M., Pfannkuche, K., Rajewsky, K. & Edenhofer, F. Ability of the hydrophobic
FGF and basic TAT peptides to promote cellular uptake of recombinant Cre
recombinase: a tool for efficient genetic engineering of mammalian genomes.
Proc. Natl Acad. Sci. USA 99, 4489–4494 (2002).
34. Eggan, K. et al. Hybrid vigor, fetal overgrowth, and viability of mice derived by
nuclear cloning and tetraploid embryo complementation. Proc. Natl Acad. Sci. USA
98, 6209–6214 (2001).
Supplementary Information is linked to the online version of the paper at
www.nature.com/nature.
Acknowledgements We thank H. Suh, D. Fu and J. Dausman for technical
assistance; J. Love for help with themicroarray analysis; S.Markoulaki for help with
blastocyst injections; F. Edenhofer for a gift of Tat-Cre; and S. Yamanaka for the
Nanog-neo construct. We acknowledge L. Zagachin in the MGH Nucleic Acid
Quantitation core for assistance with real-time PCR. We also thank C. Lengner,
C. Beard and M. Creyghton for constructive criticism. M.W. was supported in part
by fellowships from the Human Frontiers Science Organization Program and the
Ellison Foundation; B.B. by grants from the BurroughsWellcome Fund, the Harvard
Stem Cell Institute and the NIH; and R.J. by grants from the NIH.
Author Contributions M.W., A.M. and R.J. conceived and designed the
experiments and wrote the manuscript; M.W. derived all iPS lines; M.W. and A.M.
performed the in vitro and in vivo characterization of the iPS lines (teratoma, 2N and
4N injections and IHC) and the conditional Dnmt1 experiment; A.M. investigated
the promoter and imprinting methylation; M.K. and B.B. performed and analysed
the real-time PCRs and ChIP experiments; R.F. and K.H. generated the selectable
MEFs and TTFs; R.F. performedwestern blot and PCR analyses; and T.B. performed
the microarray analysis and the proviral integration Southern blots.
Author Information All microarray data from this study are available from Array
Express at the EBI (http://www.ebi.ac.uk/arrayexpress) under the accession
number E-MEXP-1037. Reprints and permissions information is available at
www.nature.com/reprints. The authors declare no competing financial interests.
Correspondence and requests for materials should be addressed to R.J.
(jaenisch@wi.mit.edu).
ARTICLES NATURE |Vol 448 | 19 July 2007
324
Nature ©2007 Publishing Group

METHODS
Cell culture, MEF isolation, gene targeting and viral infections. ES and iPS
cells were cultivated on irradiatedMEFs inDMEcontaining 15% fetal calf serum,
leukaemia inhibiting factor (LIF), penicillin/streptomycin, L-glutamine, and
non-essential amino acids. All cells were depleted of feeder cells for two passages
on 0.2% gelatin before RNA, DNA or protein isolation. Transgenic MEFs were
isolated and selected in 2mgml21 puromycin (Sigma) from E13.5 chimaeric
embryos after blastocyst injection ofOct4-inducible KH2ES cells29 that had been
previously targeted with either Oct4-IRES-GFPneo or Nanog-neo constructs
(Fig. 1a and ref. 24). Using homologous recombination in ES cells, an IRES-
GFPneo fusion cassette was inserted into theBclI site downstreamofOct4 exon 5.
Correctly targeted ES cell clones were screened by Southern analysis of NcoI-
digested DNA using a 59 external probe. The murine cDNAs for Oct4, Sox2, Klf4
and c-mycwere PCR amplified from ES cell cDNA, sequence-verified and cloned
into the Moloney-based retroviral vector pLIB (Clontech). 23 105 MEFs or
TTFs at passage 2–4 were infected overnight with pooled viral supernatant
generated by transfection of 43 106 HEK293T cells (Fugene, Roche) with
10 mg of viral vectors and the packaging plasmid pCL-Eco in a 10-cm dish35.
Blastocyst injection.Diploid or tetraploid blastocysts (94–98 h after HCG injec-
tion) were placed in a drop of DMEMwith 15% FCS undermineral oil. A flat-tip
microinjection pipette with an internal diameter of 12–15mm was used for iPS
cell injection (using a Piezo micromanipulator34). A controlled number of cells
was injected into the blastocyst cavity. After injection, blastocysts were returned
to KSOM media and placed at 37 uC until transferred to recipient females.
Recipient females and caesarean sections. Ten to fifteen injected blastocysts
were transferred to each uterine horn of 2.5 days post coitum pseudo-pregnant
B6D2F1 females. To recover full-termpups, recipientmothers were killed at 19.5
days post coitum. Surviving pups were fostered to lactating BALB/c mothers.
Southern blot, methylation and chromatin analyses. To assess the levels of
DNA methylation, genomic DNA was digested with HpaII, and hybridized to
pMR150 as a probe for the minor satellite repeats25, or with an IAP-probe26.
Bisulphite treatment was performedwith theQiagen EpiTect Kit. For themethy-
lation status of Oct4 and Nanog promoters, bisulphite sequencing analysis was
performed as described previously27. A total of 10–20 clones of each sample was
sequenced in both directions. For imprinted genes, a COBRA assay was per-
formed. PCR primers and conditions were as described previously28. PCR pro-
ducts after bisulphite treatment and gel purification were digested with BstUI
(CGCG; H19, Peg3 and Snrpn) or HpyCH4 IV (ACGT; Peg1) and resolved on a
2% agarose gel. Unmethylated CpGs in the recognition sequence will be con-
verted to T and not cut. The status of bivalent domains was determined by
chromatin immunoprecipitation followed by quantitative PCR analysis, as
described previously12.
Expression analysis. Fifty nanograms of total RNA isolated using TRIzol reagent
(Invitrogen) was reverse-transcribed and quantified using the QuantTtect SYBR
green RT–PCR Kit (Qiagen) on a 7000 ABI detection system. Western blot and
immunofluorescence analysis was performed as described29,30. Primary antibod-
ies included Oct4 (monoclonal mouse, Santa Cruz), Nanog (polyclonal rabbit,
Bethyl), actin (monoclonal mouse, Abcam) and SSEA1 (monoclonal mouse,
Developmental Studies Hybridoma Bank). Fluorophor-labelled secondary anti-
bodies were purchased from Jackson Immunoresearch. Microarray targets from
2 mg total RNA were synthesized and labelled using the Low RNA Input Linear
Amp Kit (Agilent) and hybridized to Agilent whole-mouse genome oligo arrays
(G4122F). Arrays were scanned on an Agilent G2565B scanner and signal intens-
ities were calculated in Agilent FE software. Data sets were normalized using anR
script (available at http://www.ebi.ac.uk/arrayexpress) and clustered as prev-
iously described31.
Viral integrations.Genomic DNAwas digested with SpeI (Oct4, c-myc, klf4) or
HindIII (Sox2) overnight, followed by electrophoresis and transfer, and the blots
were hybridized to the respective radioactively labelled cDNAs of the four tran-
scription factors.
Genotyping. Blastocysts were lysed for 4 h in 10ml 50mMTris, pH 8.8, contain-
ing 1mM EDTA, 0.5% Tween20 and 200mgml21 proteinase K. After heat inac-
tivation for 15min, PCRwas performedwith the following conditions: 95 uC30 s
(1 cycle); 95 uC 10 s, 60 uC 15 s, 72 uC 15 s (40 cycles); 72 uC 5 min.
Primer sequences for genotyping. GFP-F, TCCATGGCCAACACTAGTCA;
GFP-R, TCCCAGAATGTTGCCATCTT; pLIB-FW1, CCCCCTTGAAC-
CTCCTCGTTCGAC; Oct4R, GAGGTTCCCTCTGAGTTGCTTT; MycR,
CGAATTTCTTCCAGATATCCTCAC.
Primer sequences for viral-specific qRT–PCR. rtKlf4_virusF1, TCTCTA-
GGCGCCGGAATTC; rtKlf4_virusR1, CCATGTCAGACTCGCCAGGT;
rtMyc_virusF1, CTTCTCTAGGCGCCGGAATT; rtMyc_virusR1, TGGT-
GAAGTTCACGTTGAGGG; rtOct4_virusF1, TACACCCTAAGCCTCCGCCT;
rtOct4_virusR1, ATTCCGGCGCCTAGAGAAG; rtSox2_virusF1, TACACCC-
TAAGCCTCCGCCT; rtSox2_virusR1, ATTCCGGCGCCTAGAGAAG.
Dnmt1 hairpin target sequence DZ. GGAAAGAGATGGCTTAACA.
35. Naviaux, R. K., Costanzi, E., Haas, M. & Verma, I. M. The pCL vector system: rapid
production of helper-free, high-titer, recombinant retroviruses. J. Virol. 70,
5701–5705 (1996).
doi:10.1038/nature05944
Nature ©2007 Publishing Group