Role of post- transcriptional regulators in the establishment and maintenance of cell identity

Regulation of gene expression, which is essential for the unfolding of all processes taking
place in multicellular organisms, is very complex. Gene expression is controlled at the level of
transcription, RNA processing and localization, translation and protein modification and
decay. Among the various post-transcriptional regulators of gene expression, microRNA
(miRNA)s contribute to the maintenance of gene expression patterns among various cell
types in an organism. miRNAs are small, evolutionarily conserved non-protein coding RNAs,
whose biogenesis involves multiple steps in the nucleus and cytoplasm of the cell. So far
35,828 miRNAs have been reported from 233 species and in humans they are present as one
of the abundant gene families comprising over 2500 miRNAs. Mature miRNAs are loaded into
Argonaute (AGO) proteins to form RNA induced silencing complexes (RISC), which find their
targets via nucleotide complementarity between sites mostly present in 3’ untranslated
region (3’UTR)s of mRNAs and miRNAs. The outcome is destabilization or translational
repression of the miRNA targets. Although the components of miRNA biogenesis are
relatively well characterized, the mechanisms through which miRNAs execute their functional
activities remain less understood. In the first chapter of this thesis, we have addressed two
important aspects of miRNA mediated gene regulation.
Differential expression analysis based on high-throughput data sets generated upon
modulating the expression of a given miRNA in a given model has helped to identify miRNA
targets. Many computational target prediction models have been proposed. They are
typically trained on high-throughput data sets, and are based on few parameters such as
seed complementarity of targets, evolutionary conservation etc. Validation of predicted
miRNA targets remains non-trivial and we believe that one reason could be lack of methods
that consider the miRNA activity at multiple levels. An aspect that has been largely ignored so
far is the time scale on which miRNAs regulate their targets. In one study we have addressed
the kinetics aspects of miRNA regulation, and proposed a model that takes these aspects into
account. The parameters of this model were inferred from a variety of low and highthroughput
experimental data sets and we found that the model well describes the time
dependent changes in the level of mRNA, proteins and ribosome density levels upon miRNA
transfection and induction. We also found that miRNAs may not generally act as fast
regulators of gene expression due to two bottlenecks, one is the miRNA loading into
Argonaute proteins and the other is the rate of protein decay. These influence the time-scale
and magnitude of miRNA mediated gene regulation.
Several recent studies have indicated that the miRNA binding sites present in the coding
region (CDS) of an mRNA are functional, but their implications remain unclear. Use of highthroughput
approaches such as cross-linking and immunoprecipitation (CLIP) to isolate AGO
bound target sites indicate that there are as many sites located in CDS as in 3’UTRs. The
second study presented in this thesis concerns itself with the function of coding regionlocated
miRNA binding sites. Using published high-throughput data sets of Argonaute CLIP
and ribosome protected fragment profiles upon miRNA transfections, we have shown that
miRNA binding sites that are located in CDS and 3’UTRs have co-evolved and have similar
sequence and structure properties. We also found that the miRNA binding sites located in
CDS are capable of inhibiting translation, while those located in 3’UTR are more efficient in
triggering mRNA degradation. This particular observation was validated experimentally using
an inducible miRNA expression cell line, and with a luciferase reporter system containing CDS
located binding sites of the cognate miRNA. Our study therefore suggests that miRNAs may
co-target CDS and 3’UTR to fine-tune the time scale and magnitude of the posttranscriptional
regulatory effect imposed by them.
Recent studies reported that miRNAs from miR302/367 cluster enhance the somatic cell
reprogramming induced with embryonic stem cell (ESC) specific transcription factors: OCT4,
SOX2, KLF4 and c-MYC. Few other reports also claimed that miR-302/367 cluster alone is
enough to reprogram somatic cells to induced pluripotent stem cell (iPSC)s. However, the
mechanisms underlying the miRNA mediated reprogramming are not clear. We tried to
establish the miR-302/367 mediated reprogramming of fibroblasts in the primary mouse
embryonic fibroblast (MEF)s, but, similar to other labs, we were unable to reproduce the
initial result. However, we have succeeded in enhancing the reprogramming efficiency in the
secondary transgenic mouse embryonic fibroblasts (TNG)-MEFs that contained a
pluripotency marker “Nanog” tagged with a green fluorescent protein, along with miR-
302/367. Apart from miRNAs, other post-transcriptional regulators we have focused in this
thesis are tissue specific splicing factors. Recent evidence indicated that knockdown of
muscle blind like (MBNL) proteins enhance the reprogramming efficiency. Our own analysis
of already published mRNA-seq data sets of iPSCs and their parental cells also showed that
the tissue specific splicing factors are differentially expressed between iPSCs and their
parental cells. Especially ESRP1, ESRP2 and MSI1 showed striking changes in their expression
levels during the course of reprogramming. Based on this analysis we hypothesized that
these factors may enhance the reprogramming efficiency. To investigate this hypothesis we
again used secondary TNG-MEFs as a model and we have transduced them with both
lentiviruses and retroviruses as carriers to deliver our candidate splicing factors. Our
experiment indeed revealed an increase in the reprogramming efficiency of 1.4 fold to 2 fold,
with ESRP2 showing highest enhancement. As a follow up of these experiments, we aim to decipher the cascade of events through which miR-302/367 and splicing factors enhance the
efficiency of reprogramming induced by the ESC set of transcription factors. Since the major
changes involved in reprogramming occur during the early and late phases, we plan to
perform an early and late time series of mRNA profiling upon the induction of
reprogramming with miR-302/367 and splicing factors in secondary TNG-MEFs.
In conclusion this thesis presents two main contributions to the field of miRNA-based
regulation of gene expression. The first is a mathematical model that describes the kinetics of
miRNA dependent gene regulation and the second shows that the miRNA binding sites
located in CDS sites are functional and are more effective in inhibiting translation than the
sites present in 3’UTR. Besides these two studies, I have obtained evidence that tissue
specific splicing factors, in particular ESRP1, ESRP2 and MSI1 are able to enhance
reprogramming efficiency up to 2 fold. Experiments are under way to uncover the
mechanisms involved in the enhancement of reprogramming efficiency by tissue specific
factors.
Since relatively little is known about the function of alternative splice forms in iPSC
generation, these preliminary studies could set the ground for future research in iPSC and
also towards clinical research. Being able to obtain iPSCs with more efficient and safer
methods will enable studies of various diseases at the clinical level. On the other side, as
miRNAs are currently being considered for various therapeutic approaches, a deeper
understanding of the underlying mechanisms by which miRNAs regulate gene expression
would help in the better design of therapeutic compounds. The work presented in this thesis
may thus be beneficial for both the miRNA and the iPSC fields.