Embryonic stem cells, derived from the inner cell mass of the blastocyst stage mammalian embryos 
, can self-renew nearly indefinitely in culture and give rise to all cell types of the three germ layers, ectoderm, mesoderm and endoderm, during differentiation. ESCs possess distinctive transcriptional regulatory circuits and chromatin signatures that are critical for maintaining pluripotency and self-renewal 
. Recent studies suggest that ESCs exhibit a relatively “open” chromatin state, and during differentiation, heterochromatin formation increases 
. However, whether this “open” chromatin state is necessary for pluripotency and whether the compaction of chromatin is required for ESC differentiation remain to be addressed.
Linker histone H1 is the major chromatin architectural protein in mediating higher order chromatin folding. H1 TKO ESCs have an H1/nucleosome ratio of 0.25, equivalent to 1 H1 per 4 nucleosomes, a nearly 50% reduction in total H1 levels in comparison with WT ESCs 
. The H1 level is especially low in H1 TKO ESCs when compared with an H1/nucleosome ratio of 0.75~0.8 in differentiated cell types from various adult tissues 
. H1 TKO ESCs have globally decondensed chromatin 
, offering an approachable means to examine the effect of chromatin decondensation on ESC pluripotency and differentiation. H1 TKO ESCs maintain ESC colony morphology, express pluripotency factors (), propagate and self-renew normally as wild-type ESCs, suggesting that a more “open” chromatin structure than normal WT ESCs does not interfere with the “basal” state of ESCs, and may even promote the maintenance of this primitive state. This prediction is consistent with the fact that H1 TKO ESCs are easier to maintain and have sustainable OCT4 pluripotency factor expression and robust growth even under conditions normally promoting spontaneous differentiation, such as culturing ESCs in the absence of LIF and feeder cells for a prolonged period. ESCs are found to have hyperdynamic chromatin with loosely bound major chromatin architectural proteins, such as H1 and HP1 
. A more “open” chromatin in H1 TKO ESCs may suggest a more dynamic chromatin structure due to the lack of structural constraints. However, it is not clear at present whether the remaining H1 proteins in H1 TKO ESCs undergo a change in post-translational modifications, such as phosphorylation, which would change the binding affinity of these remaining H1 subtypes to chromatin 
. We also note the considerable amount of H1s remaining in these TKO ESCs, thus further reducing H1 amount by knockout or siRNA could help determine if a minimal level of H1 is required to permit self-renewal of ESCs.
While a significant reduction in H1 levels does not interfere with ESC self-renewal, it appears to clearly impair ESC differentiation. This is manifested in static culture conditions that promote spontaneous ESC differentiation, in a rotary suspension culture system which induces highly reproducible and robust EB formation and differentiation 
, as well as in a well defined neural differentiation regimen. H1 TKO EBs formed in rotary culture have a reduced level of activation of many developmental genes and markers from all three germ layers, suggesting that the effects of H1 depletion on differentiation and cell fate decision broadly impact early developmental gene expression. This may explain why only 50% of H1 TKO embryos are present at E7.5 
. Furthermore, H1 TKO ESCs are defective in forming neuronal cells, glial cells, and lack formation of neural network, which are essential for nervous system development in vivo
. Total levels of H1 increases progressively in EB formation and differentiation, suggesting an increasingly more condensed chromatin state during EB differentiation in WT cultures. H1 TKO EBs have an H1 to nucleosome ratio lower than WT ESCs. The fact that H1 TKO ESCs cells are unable to execute normal differentiation programs suggests that an especially low H1 level (and the resulting more open chromatin structure 
) impairs ESC pluripotency and differentiation. Thus, elevated levels of the total H1 amount as well as a more compact chromatin are not mere consequences of differentiation processes, but a necessity to enable it to proceed normally.
H1c, H1d, H1e and H10
are four H1 subtypes that increase significantly during ESC differentiation. H1x, although whose mRNA expression has been reported to increase during differentiation of human ESCs and embryocarcinoma cells 
, is not detected in HPLC profiles of both WT and TKO ESCs throughout differentiation despite a 2-fold increase in mRNA levels in TKO ESCs compared with WT (
and data not shown). Thus, this more distantly related H1 subtype (H1x) is present at a negligible level compared with the 6 somatic H1 subtypes (H1a-e and H10
) in ESCs and EBs. In contrast, H1a and H1b are abundantly present in ESCs, together accounting for one third of total H1 content in WT ESCs. Although both H1a and H1b increase approximately 50% in TKO ESCs upon depletion of H1c, H1d and H1e, the levels of H1a and H1b do not increase during EB differentiation of WT or TKO cultures. Thus, H1c, H1d, H1e, and H10
, but not H1a and H1b, are likely to be the major contributors for the effects of H1 on ESC differentiation and repression of pluripotency genes during ESC differentiation. In particular, H10
, a subtype highly expressed in differentiated cells and tissues 
, progressively increases in bulk chromatin and at the Oct4
promoter during EB differentiation and largely accounts for the increase in total H1 levels in TKO EBs during differentiation (, Figure S3
, and Figure S7
). Thus it would be very interesting to investigate if further deletion of H10
in the face of H1 TKO will result in a complete inhibition of ESC differentiation. Nevertheless, none of these four H1 subtypes alone appears to be required for mouse ESC differentiation, because knockout mice with deletion of one of these four H1 subtypes develop normally 
, suggesting that the differentiation defects we observed here are more likely caused by a marked reduction of total H1 content in H1 TKO cells. Furthermore, we show that a partial rescue of H1 content by reintroduction of H1d into TKO cells mitigates the impairment of differentiation. Together, we surmise that a potential threshold of H1 levels, but not necessarily a specific H1 subtype, is required for proper ESC differentiation.
The effects of H1 depletion on gene expression in EBs are significant and wide-spread, drastically affecting many genes (, and Figure S2C
), in sharp contrast to the limited number of genes with altered expression in H1 TKO ESCs 
. It is conceivable that H1 depletion in ESCs and a marked decondensation of the chromatin pose little effects on the “basal” state of ESCs, but more so on impairing the capability of ESCs to transit to differentiated cells which exhibit more compact chromatin. Nevertheless, the influence of H1 on many developmental genes in EBs could be a secondary effect resulting from the lack of effecient repression of pluripotency gene expression, such as Oct4
, which associate with repressor complexes to silence developmental genes 
. The effects might also be caused by misregulation of multiple key developmental genes required for normal differentiation to proceed. It is interesting to note that 50% of H1 TKO embryos are able to progress to mid-gestation, suggesting that early differentiation in three germ layers in vivo
is possible for some TKO embryos 
. Consistently, H1 TKO ES cells are capable of forming EBs (), albeit mostly impaired in differentiation, and teratomas that contain a small fraction of cells differentiated into the three germ layers (data not shown). The impairment of ESC differentiation in vitro
yet survival of some knockout embryos to mid-gestation stage is reminiscent of several other knockouts of ubiquitously expressed proteins that bind and modify chromatin 
, which probably reflects more heterogenous cell populations and conditions in vivo
Importantly, we discovered that, compared with WT ESCs, the H1 TKO cells fail to effectively silence the expression of pluripotency genes Oct4
, which are critical for pluripotency 
. We believe that this effect of H1 on repression of Oct4
is direct because 1) Oct4
expression is higher in H1 TKO compared with WT both in vivo
in embryos and in vitro
using three differentiation schemes for ESCs and EBs, although the degree of effects varies according to different differentiation schemes employed; 2) reconstitution of H1d into H1 TKO ESCs restores the effective repression of expression and dynamic changes in histone modifications and DNA methylation levels during differentiation; 3) the level of H1 is cumulatively increased at the Oct4
promoter during differentiation of WT, but not of H1 TKO, cultures. We suggest that the H1 occupancy at Oct4
promoter in ESCs could be the basal/minimal level for detection by qChIP assay, as H1 has been found to be relatively depleted from active promoters compared with other regions 
. Interestingly, qChIP analysis showed that the association of H10
promoters was significantly higher in RES cells than TKO cells (Figure S7
), suggesting that the presence of sufficient H1 proteins may facilitate H10
binding. We surmise that the progressive increase of H1c, H1d and H1e during differentiation and the increased H1 occupancy at Oct4
promoter lead to a transition to a more condensed local chromatin structure necessary for stable silencing of Oct4
during differentiation (). These results together with the observation that OCT4 is present at the promoters of several H1 subtypes in human ESCs 
suggest a potential feedback loop between OCT4 and H1 in stem cell fate determination.
Interestingly, we found that CpG methylation of Oct4
promoter in H1 TKO embryos is significantly reduced compared with wild-type littermates. Although less pronounced in EB differentiation, the effects of H1 depletion on DNA methylation at Oct4
promoter are also apparent in day 10 EBs. This observation reinforces the link between H1 and DNA methylation, which was initially discovered at imprinting control regions (ICRs) of H19
and later at regulatory regions of the immunoglobin heavy chain locus and homeobox Rhox
gene cluster 
. Future studies on how DNA methylation changes at these regions in H1 TKO ESCs during differentiation will provide additional insights on dynamic profiles of DNA methylation upon differentiation in the face of minimal level of H1 and/or open chromatin structure.
H1 TKO EBs do not exhibit the opposite changes in the levels of the active histone mark (H3K4me3) and the repressive histone mark (H3K9me3) at promoters of Oct4
that normally occur in wild-type EBs during differentiation. Interestingly, we did observe significant changes in the levels of histone modifications in wild-type EBs at day 7 in rotary culture, before an increase in DNA methylation levels occurred at Oct4
promoter. This result reinforces the notion that DNA methylation is a slower mark to establish compared with histone marks 
. It is noteworthy that the levels of DNA methylation at the Nanog
promoter do not display a difference in WT and H1 TKO embryos at day 8.5 and are not altered during EB differentiation, suggesting that DNA methylation is unlikely to be responsible for gene expression changes of Nanog
during this period of time.
Our results suggest a role of H1 and chromatin compaction in epigenetic regulation of the pluripotency gene Oct4
, likely mediated through DNA methylation and histone modifications. To our knowledge, this represents a novel mechanistic link by which bulk chromatin compaction is directly linked to pluripotency, by participating in repression of the pluripotency genes. In ESCs, DNMT3b has been shown to interact with H1 
. In vitro
studies demonstrated that H1 interacts with HP1 
which can in turn bind to SUV39H which methylates H3K9. Moreover, H1 has been shown in vitro
to stimulate the activity of PRC2 toward methylation of H3K27me3 when H1 is incorporated into nucleosomes 
, and we have also observed interactions between H1 and PRC2 components in ESCs (Cao, Ho, Lasater, and Fan, unpublished observation). Therefore, we envision that during ESC differentiation, H1 levels increase, which may facilitate the recruitment of DNMTs, SUV39H and PRC2 to Oct4
promoter, promoting the establishment and/or maintenance of repressive epigenetic modifications and silencing the expression of this pluripotency gene ().
In summary, we have demonstrated that loss of linker histone subtypes H1c, H1d, and H1e impairs embryonic stem cell differentiation. Furthermore, our results indicate that H1 contributes to silencing of pluripotency factors and participates in mediating changes in DNA methylation and histone marks necessary for silencing of pluripotency genes during differentiation. Thus, modulating the levels of H1 linker histones and chromatin compaction may potentially serve as a new strategy for regulating stem cell pluripotency.