Biallelic Expression of Nanog Protein in Mouse Embryonic Stem Cells

Transcription   factors   (TFs)   and    their networks are central effectors controlling pluripotency  (Young,   2011).   Numerous involved  TFs  have  been  identified,  but  a subset of core  pluripotency TFs regulates the  majority  of others. One  such  factor, Nanog,  is expressed in  pluripotent  cells, is required for self-renewal of mouse em- bryonic stem cells (ESCs) in vitro, is able to  force  ESC  self-renewal upon  overex-  pression in the  absence  of  LIF, and   is necessary for the normal development of early mouse embryos (reviewed in Young, 2011).  Several  studies have  shown  that Nanog   expression  is  heterogeneous  in populations  of  pluripotent ESCs,   which can  express high  or  low  Nanog  levels (reviewed     in    Young,    2011),    making Nanog  regulation an interesting model  for analyzing the dynamic regulation of fluctu- ating   but   stable  TF  expression  states. Recently,  allele-specific   expression  of Nanog—as  assessed  by  a  combination of fluorescent in situ hybridization (FISH) to detect Nanog mRNA and protein-based assays  involving   fusion   of  destabilized fluorescent proteins connected to Nanog via  a  self-cleavable peptide—has been described as  a  potential  mechanism for regulation   of   

Nanog expression and, consequently, pluripotency (Miyanari and Torres-Padilla,    2012).    These   studies suggested  that Nanog is predominantly expressed  in  a   monoallelic   manner   in serum/LIF-cultured  ESCs but  biallelically in 2i ‘‘ground state’’  conditions, and  they led  to  the  conclusion  that  switching   to higher biallelic Nanog  expression is asso- ciated with a more stable pluripotent state. However, the underlying mechanisms and functional relevance remained unclear.
To  examine the  allelic  distribution of Nanog   expression at  the  protein  level, we  created knockin  ESC  lines  in  which the  two  endogenous  Nanog  alleles  are targeted with a  yellow (VENUS) and  red (KATUSHKA) fluorescent   protein   (FP), respectively   (Figures    S1A    and    S1B available   online).   The   FPs   are   fused  to the  C terminus  of the  Nanog  protein, so    they    reflect    all   of    the    regula- tory mechanisms influencing  the amount of Nanog  protein  in ESCs  and  measure functionally   relevant   levels   of   Nanog protein,    not    separate    markers   that could   have  different  stability  or  regula-  tion.  To confirm  the  functionality  of  the Nanog-FP   fusions,  the   pluripotency  of

the   NanogVENUS/KATUSHKA       ESC  reporter

lines was  tested in vitro and  in vivo. Loss of  Nanog   leads  to   differentiation  and loss  of  ESC  maintenance, and   Nanog- deficient embryos  do  not  develop  past the   implantation   stage  (Mitsui   et   al.,

2003).  In contrast,  NanogVENUS/KATUSHKA

ESCs survived and proliferated  normally over  at  least   250  population doublings in  vitro,   exhibited  normal   morphology of  undifferentiated  ESCs   (Figure   S1C), and  expressed  other  ESC-pluripotency- specific   TFs   like   Oct3/4,  Sox2    (Fig- ure  S1D),  and   Rex1  (data   not  shown). Both  Nanog-FP reporters  also   showed normal   downregulation  during   induced ESC  differentiation upon  LIF  withdrawal (Figure    S1H).    We    also   verified    the functionality    of   the    NanogVENUS      and NanogKATUSHKA   fusion  proteins through a tetraploid aggregation  assay, the  most  stringent   test    for    ESC    pluripotency: normal  day  9.5  embryos can  be  gener-

ated  from   NanogVENUS/KATUSHKA     ESCs

without   contribution  of  tetraploid   cells (Figure  S1E).  In addition, the  stability  of NanogVENUS   and  NanogKATUSHKA     fusion proteins   is   identical    to   that   of   wildtype   Nanog   protein   (Figure  S1F).  Thus, the    normal    function    and   stability    of NanogVENUS   and  NanogKATUSHKA     fusion proteins indicates that  they  can  be  used as   faithful  reporters  of  Nanog   protein expression.

We  used the  labeled cells  to  examine Nanog  expression.   As   previously   des- cribed  (Chambers et al., 2007), we saw  a range  of  Nanog  expression levels  when the   ESCs   were   cultured  in   serum/LIF conditions, although  the  dynamic range  was  not  as  broad as  in some previous reports. We found  that  the  extent of this variability of Nanog expression depended on  culture  conditions  and   strain   back- ground  and   could   also   vary   between genetically identical   ESC  clones. How- ever,    we    unexpectedly   did    not    see evidence for widespread monoallelic exp- ression of  Nanog  protein   (Figure  S1G). Instead,   Nanog  expression was  highly correlated  between  the   two   alleles   in terms  of   the   expression   level   within individual  cells.  This  situation remained unchanged in ESCs  cultured over  many weeks   (data    not    shown).    Consistent with   prior   reports,   Nanog    expression changed to a more  uniform high distribu- tion  in  ESC populations cultured in 3i ground  state   conditions  (Ying   et   al.,

2008) (Figure S1G).  We  cannot  exclude potential    monoallelic    Nanog     protein  expression in a  very  small  subset (less than  2%) of ESCs  due  to potential noise levels  of  FACS  analysis  (individual dots in FACS  plots  of  Figure  S1G).  We  can, however,    conclude   that    we    do    not see  evidence for  significant  monoallelic Nanog  expression in ESCs  at the protein level.  Although  we  did  not  analyze the potential  for monoallelic  Nanog protein expression in other  ESC lines, the normal self-renewal and  pluripotency  properties of our cells suggest that monoallelic  regu-  lation  of  expression is  not  required  for wild-type  Nanog  function.

It is unclear at this point what the basis is for the  difference  between  our  results and  those of Miyanari and  Torres-Padilla (2012).  One  possible  explanation could lie   with   transcriptional bursts,   which seem to  occur  at  a  low  frequency even for   actively    expressed   genes   (Suter et  al.,  2011).  Thus,  FISH data from  one point  in time  might  detect  transcription of   only   one   allele   because  of   burst behavior rather  than  overall  monoallelic Nanog  expression.  Differences in terms of stability between the separate reporter proteins  and   Nanog   itself   could   also influence  the  results seen at  the  protein  level.

It is important to  note  that  we  did  not analyze  the   potential  for  allele-specific bias   of  Nanog   transcription.  However, even  if it occurs, our data suggest that  it would  not  lead  to  prevalence  of  Nanog protein  from one allele in ESCs,  and  thus it is not  likely to  be functionally  relevant  as   a   central  mechanism of regulating pluripotency or heterogeneity in  pluripo- tency  TF expression.  Instead, we  would suggest  that   other   regulatory   mecha- nisms, including   Nanog  autorepression (Fidalgo   et   al.,   2012,   Navarro   et   al.,

2012)  and   the  topology of  the  pluripo- tency  TF and  signaling networks (MacAr- thur  et  al.,  2012),  underlie  the  heteroge- neous molecular states seen in individual pluripotent cells.  A  related paper in this issue   from  Faddah et  al.  (2013)  draws similar conclusions to ours regarding bial- lelic expression of Nanog, and in addition  looks more broadly  at variability in Nanog expression at the transcriptional level and the  activity  of  a  range  of reporter con- structs. Together, these studies will help inform  future  analysis of  the  regulation of  Nanog   expression and  pluripotency networks.

SUPPLEMENTAL INFORMATION

Supplemental Information  for this article  includes Supplemental  Experimental  Procedures and  one figure and  can  be found  with this article  online at http://dx.doi.org/10.1016/j.stem.2013.04.025.