Cell polarization is a cell property
that is fundamental for overall development. It is required for vital processes,
such as cell migration, asymmetric cell division and specification into
functionally different domains. During polarization cellular components or
functions are distributed asymmetrically along an axis of polarity.
Interactions between proteins play a
vital role in most, if not all, cellular processes. This is no different for
the polarization of cells. Previous studies have identified several proteins
and protein complexes that can drive cell polarization. For example, the
PAR-3-PAR-6-aPKC and Crumbs-SDT-PATJ protein complexes together promote apical
domain identity in polarization. These protein complexes are called cortical
polarity complexes and interact with several other cell components, such as
Rho-family GTPases, when establishing polarity.
State of the art
It is known that mutual exclusion is
a key mechanism in establishing polarity by cortical polarity regulators.
However, on the detailed mechanisms involved in the segregation of polarity
regulators into different domains knowledge is still lacking. Also, little is
known about the mechanisms through which cortical polarity is integrated with
cellular events, such as cytoskeletal rearrangement and functional
specialization of membrane domains. To fully understand the establishment of
polarity, extensive knowledge on the proteins involved and the interactions
between them is needed. Koorman et al. (2015) studied this establishment of
polarity in Caenorhabditis elegans. They
identified interactions underlying polarity establishment by using large-scale
yeast two-hybrid (Y2H) screens and phenotypic profiling. These findings can be
of large value for future research on cell polarity and the understanding of
this essential process.
To create a map of protein
interactions involved in polarity establishment, Koorman et al. (2015) selected
69 proteins that were already known to be involved in this process. Of all
these proteins, a Y2H bait construct was created by cloning the full-length
open reading frame into a Gal4 AD vector. Also, several fragment bait
constructs were cloned for each protein as this was expected to increase
interaction detectability. Screening all the bait constructs in two GAL4 AD
libraries and a C. elegans AD-cDNA
library and some additional steps to increase accuracy, resulted in a C. elegans polarity interaction network
(CePIN). This network contains 439
interactions between 296 proteins. Of these interactions 54 were reported
previously, of which 19 have been studied in detail. The majority of the
interactions were detected using fragments only, confirming the expected
increase of detectability.
They also assessed the quality of
the CePIN by examining whether the
interacting proteins also had a functional association by looking for other
shared characteristics. The results show that interacting protein bars were
enriched for similar GO terms and for presence in WormNet. This indicates
functional interactions between C.
Also, numerous interactions were
retested and validated with co-affinity purifications from human embryonic-kidney-293
cells, mammalian protein-protein interaction trap (MAPPIT) and comparisons with
three previously published C. elegans MAPPIT
analyses. Together they confirmed the quality of the CePIN.
The functional associations were
further examined by combining protein interaction data with phenotypic data.
For this they studied the phenotypical effects of RNAi in nine different
strains (consisting of different epithelial tissues, processes and neurons) expressing
different fluorescently tagged proteins which are involved in numerous
polarity-related processes. For example, defects in excretory canal polarity
after RNAi were examined by using a strain expressing VHA-5::GFP. All the 69
bait proteins were screened for 40 possible defects across all nine marker
strains. For 44 bait proteins, RNAi caused a detectable defect in at least one
of the marker strains. Next, the binding partners of the 44 bait proteins were
screened specifically in the strains in which the defect was detected. This
phenotypic profiling resulted in the identification of 100 physically
interacting protein pairs.
Next, these results were evaluated
in three ways. It was shown that an overlap in phenotype predicts functional
association and that it correlates with protein interaction. This was done with
hierarchical clustering to cluster the bait genes by phenotypic resemblance and
GO similarity respectively. The results show that an overlap in phenotype is positively
correlated with high GO similarity.
It was also examined whether
interacting proteins with overlapping phenotype were more likely to be
previously described in literature. The results showed that 9% of the
interactions with overlap in phenotype were already in literature, compared to
2% of the remaining interactions. All these results together indicate that
phenotypic profiling makes it possible to detect interactions important in vivo and to cell polarity.
Finally, a last validation of the approach
used by Koorman et al. (2015) to identify functionally relevant interactions
was done. For this, they focussed on the interaction between PAR-6 and PAC-1,
proteins which are both involved in radial polarization. The results show that
a 100-amino-acid sequence just downstream of the PAC-1 RhoGAP domain was able
to bind to PAR-6. This interaction was confirmed by co-affinity purification
from mammalian HEK293 cells. Also, when expressed together in HeLa cells PAC-1
and Par-6 co-localized in a pattern obviously distinct from the pattern when
expressed individually. Then, by deleting the PAR-6-binding domain from PAC-1,
Koorman et al. (2015) determined the functional importance of the interaction.
The results show that the lack of the binding site led to a failure of the
radial localization pattern of PAR-6. This indicates that a direct interaction
between the two is needed in order for the establishment of polarization.
Discussion and future developments
In this study Koorman et al. (2015)
created a polarity interaction network in C.
elegans, containing 439 interactions between 296 proteins. This led to the identification of 385
new interactions of which 12 were studied in detail. The quality of this CePIN has been validated with several
computational and experimental approaches. Together these results show that the
smaller-scale targeted approach used in this study can be used to identify more
important interactions, which are not present in high-throughput interaction
Also, phenotypical analysis show
that overlapping phenotypes are correlated with the presence of physical
interaction, GO similarity and previous descriptions of interactions in
literature. This indicates that combining phenotypic data with interaction data
gives a better prediction of functional relationships than interaction data
alone. Therefore, future research should focus on identifying additional pairs
with overlapping phenotypes, which can be done by using additional marker
Overall, the CePIN created provides much information on polarity-related
interactions. This information can be of great value in future studies
focussing on the mechanisms involved in cell polarity. Considering the fact
that many mechanisms involved in polarity establishment are conserved between C. elegans and for example humans and
flies, the CePIN can be used in a
large variety of future research models.