Biology Reference
In-Depth Information
Although the prevalence of OTEs was underestimated
in early RNAi screens, a number of approaches have now
been developed to minimize their effects
[47,144]
. These
include the development of computational tools to design
RNAi reagents with limited or no homology to genes other
than the intended target; the use of multiple independent
RNAi reagents targeting the same gene; and the rescue of
the RNAi-induced phenotype by an RNAi-resistant version
of the gene. Further, with the availability of the catalogue of
expressed genes in a wide array of commonly used cell
lines by RNA sequencing
[145]
, false positives associated
with a screen performed in Drosophila cells can be iden-
tified and filtered based on whether the targeted gene is
expressed in the cell line being screened. One can also filter
out potential false positives by removing genes that score in
a large majority of RNAi screens.
Large-scale epistasis or synthetic lethality studies (see
Box 5.1
for definition) using sensitized genetic backgrounds
can also uncover new components of signaling pathways
[146]
because they tend to reveal genes that are involved in
redundant or parallel pathways/complexes. Such screens are
similar in concept to the synthetic genetic array (SGA)
analysis in yeast, where the viability of a set of gene dele-
tions has been tested in backgrounds where other genes have
been similarly deleted (synthetic lethal) or overexpressed
[147
genes involved in the early secretory pathway (ESP) in the
budding yeast was constructed which robustly identified
known pathways and relationships, such as the effect of the
unfolded protein response (UPR) pathway on secretory
functions, and the hierarchical relationships of the different
stages of vesicular trafficking. This study also identified
a strong link between endoplasmic reticulum-associated
degradation (ERAD) pathway and lipid biosynthesis,
a connection that had been previously poorly characterized.
The E-MAPs strategy has been successfully extended to
study the networks of genes involved in creating, main-
taining and remodeling the chromatin in response to
various cues
[156]
, and also to identify novel components
of the RNAi machinery in the fission yeast Schizo-
saccharomyces pombe
[157]
. The success of these studies
highlights the power of E-MAPs to provide a systems-level
view of the functional topology of networks that cannot be
obtained by other methods. Recently, the concept of E-
MAPs has been successfully implemented in Drosophila
cells using combinatorial RNAi screens
[158]
. In this study,
pairwise interactions between 93 genes involved in
signaling were evaluated using two independent RNAi
reagents for each per target. This set of 93 genes included
components of the three MAPK pathways (Ras-MAPK,
JNK and p38 pathway) and all expressed protein and lipid
phosphatases. The pairwise knockdowns were analyzed for
their effects on cell number, mean nuclear area and nuclear
fluorescence intensity and resulted in 73 728 measure-
ments, from which interaction scores were estimated. The
success of the strategy was reflected in the high frequency
of interactions observed between known components of the
Ras-MAPK signaling pathway and a clear separation from
regulators of the JNK signaling pathway. In addition, the
authors identified connector of kinase to AP-1 (Cka),
a scaffold protein in the JNK signaling pathway
[159]
,asa
positive regulator of Ras-MAPK signaling, and thus
a putative point of cross-talk between the two pathways was
identified.
Functional genomic approaches at the level of whole
systems are powerful because they can identify most genes
that affect a given signaling network, and have revealed
that, contrary to previous views, hundreds of genes may be
a part of a signaling network. However, genetic studies do
not distinguish between direct and indirect effects, and
therefore it is not clear where in the network the different
genes identified act. Understanding how they contribute to
the overall structure of the cellular signaling network
requires the integration of genetic data with other datasets
such as protein
149]
. The results from these studies showed that
RNAi of many individual genes does not affect growth, but
that many genes do have a synthetic genetic growth
phenotype in combination with other genes. These genetic
interactions include both negative (aggravating) interactions
as well as positive (alleviating) ones, where the phenotype
of eliminating one gene is attenuated by the loss of a second
one. Combinatorial RNAi experiments where dsRNAs are
screened for their ability to suppress or enhance the effect
caused by another dsRNA (or by small molecules) are also
becoming increasingly common
[150,151]
. Examples of
HTS for multiple genes by RNAi include 17 724 combi-
nations that identified regulators of Drosophila JNK
signaling
[150]
, and combinatorial RNAi of disease relevant
genes in C. elegans, which identified ~1750 novel functions
for genes in signaling
[152]
. RNAi microarrays facilitate the
miniaturization of combinatorial RNAi screens and provide
an effective and economical way to conduct large-scale
screens in tissue culture cells
[89,153,154]
.
In addition to identifying new genes involved in
a particular biological process, comprehensive and quan-
titative genetic interaction data can be used to shed light on
the organizing principles of signaling networks and
the ways in which distinct signaling modules are inter-
connected. Schuldiner and colleagues
[155]
developed
a strategy for building large-scale genetic interaction maps
called 'epistatic miniarray profiles' (E-MAPs) that allows
one to group sets of genes based on their signature/patterns
of genetic interactions. Using this strategy, an E-MAP of
e
protein interaction networks.
e
Protein
Protein Interactions
Large-scale protein
e
protein interaction (PPI) mapping
e
complements
genetic
studies
by
revealing
physical
