Biology Reference
In-Depth Information
associations and helps to define the physical signaling
network. In PPI networks, nodes represent proteins and the
edges represent a physical association between them. The
methods most widely used to map PPI networks include the
yeast two-hybrid (Y2H) systemand its derivatives [160, 161] ,
and affinity- or immunoprecipitation followed by mass
spectrometry (AP/MS) [162
PPI approaches, as implemented thus far, have been
incomplete for investigations of signal transduction
because they (1) do not provide functional information,
and (2) often take place outside the context of endogenous
signaling. These issues can be addressed by combining
proteomics with RNAi. For instance, Y2H was used to
identify interactors of the DAF-7/TGF- b pathway in
C. elegans, resulting in a network of 59 proteins, and
RNAi was used to show that nine novel interactors func-
tionally interact with the TGF- b pathway [198] .Another
major study used a pathway-specific approach with liquid
chromatography/tandem MS to characterize the inter-
actors of 32 TNF- a /NF k B pathway components in
mammalian cells under endogenous signaling conditions
[187] . Interactors were identified at baseline and under
TNF- a stimulus, revealing 221 interactions. RNAi was
then used to determine their influence on signaling output.
This study demonstrated the power of pathway-directed
proteomics in endogenous signaling contexts. One limi-
tation of this study, however, was the lack of rigorous
quantitation of the assembly of signaling complexes. Most
signaling complexes are highly dynamic, with compo-
nents often held in inactive complexes that can change
dramatically following stimulation. For example, Raf and
KSR are held in separate inactive complexes bound to
PP2A core components and 14-3-3 proteins; following
stimulation, Ras induces the recruitment of PP2A regu-
latory subunits to Raf, dephosphorylation of 14-3-3
binding sites, release of 14-3-3 proteins, membrane
recruitment, KSR and Raf co-localization, Raf phos-
phorylation of MEK, and MEK phosphorylation of MAPK
[199,200] .
RNAi and MS can also be combined by first starting
with RNAi and then following up with MS, as has been
demonstrated for RTK/ERK signaling at baseline and
under insulin stimulation [84,85] . All of the major known
components of the pathway were tagged. In addition,
a control cell line was engineered to subtract common
interactors/contaminants. Altogether, 54 339 peptides
were identified representing 12 208 proteins, encompass-
ing an unfiltered network of 5009 interactions among 1188
individual proteins. To provide a ranked list of novel
pathway interactors, filtering out sticky proteins found in
control preparations and providing a probability that the
observed interactor is real, the significance analysis of
interactome (SAINT) method was applied to the PPI
dataset [181] .UsingaSAINTcut-offof0.83andafalse
discovery rate (FDR) of 10%, a filtered PPI network of
386 interactions among 249 proteins surrounding the
canonical components of the RTK/Ras/ERK signaling
pathway was generated [84] . In this network canonical
baits have multiple common interactors, as would be
expected from a well-connected signaling pathway (as
opposed to unbiased PPI mapping of random protein
165] . PPI networks derived
from Y2H methods are composed of binary (direct) interac-
tions, whereas those derived from AP/MS techniques can be
both direct and indirect, as they identify protein complexes.
Large-scale Y2H studies have been conducted with
proteins from Helicobacter pylori [166] , yeast [167
e
169] ,
e
C. elegans [21,170,171] , Drosophila [172
174] , and
e
human [175
179] . Strikingly, the three large-scale
Drosophila Y2H mapping studies failed to fully recapitu-
late known signaling pathways. For example, querying the
combination of these studies for Raf reveals only interac-
tions with CG15422, Ras, Rhomboid, and Rap2L (http:
//itchy.med.wayne.edu/PIM2/PIMtool.html), neglecting to
identify most known targets, scaffolds, and co-regulators of
Raf activity. Thus, these 'proteome-scale' approaches,
although they identified highly abundant or strongly inter-
acting cellular components, failed to identify many inter-
actors of signaling components
e
most likely because of
the absence of endogenous signaling contexts [180] .For
this reason, MS-based approaches have become more
popular, especially as the difficulty of implementation and
costs have dropped dramatically.
Comprehensive MS-based PPI mapping has been
applied in yeast [181
e
185] . A global protein kinase and
phosphatase interaction network identified 1844 interac-
tions between 887 proteins [181] . The success of MS
approaches has been aided by the increased sensitivity of
MS technology and implementation of tandem affinity
purification (TAP) of protein complexes [186] . Recently,
tandem affinity purification followed by MS has been used
to isolate protein complexes from Drosophila tissue culture
cells and tissues (http: //flybase.org/). ~5000 Drosophila
proteins were fused to a FLAG-HA tag so that the fusion
proteins could be expressed and recovered with their
interacting partners from cells, or from whole transgenic
flies. In addition to proteome-scale AP/MS, a number of
smaller studies have been conducted in human cells on
signaling pathways such as TNF- a and Wnt [187,188] ,
biological processes such as autophagy [189] , protein
families such as the de-ubiquitinating enzymes [190] , and
protein complexes such as the RNA
e
polymerase II and
e
PP2A complexes [191,192] .
Both Y2H and AP/MS PPI mapping methods have
been applied to the characterization of cellular networks
with disease relevance, such as virus
host interactions
e
[193
197] . These proteomic studies have confirmed that
cellular processes take place within large networks of
interconnected proteins.
e
Search WWH ::




Custom Search