Interactome Mass Spectrometry: Proximity-dependent Biotinylation for Mapping Protein Complexes


>>Hi, welcome to the next in our OSTRD Series
on cutting edge methodologies. Today, we’re going to talk about proximity dependent biotinylation
for the characterization of protein complexes and the mapping of organelles. A standard
experiment for characterization of protein protein complexes involves pull down of a
base coupled with mass spectrometry to identify the prey that copurifies with it. This is
a mass spec variant of a standard IP or pull down experiment. These are nice experiments,
they’re relatively straightforward, routinely used, if one can do an IP Western adding it
into a mass spec detection isn’t very difficult. And they’re also readily coupled to quantitative
methods to get relative information on changes in enrichment. However, there are some disadvantages.
For example, you can lose interactions where an antibody binds if you’re doing amino precipitation.
The antibody and protein interactors are going to compete for binding to the same sites and
the antibody is generally going to win. These experiments don’t work well for hard to purify
interactors, those that are transient or form unstable complexes. Also, for enzymes, because
the substrates of these enzymes are often low affinity interactions, they’re meant to
be released and not stable. And also, for baits that have poor solubility such as proteins
from membranes. Standard IP mass spec experiments are also subject to post-lysis artifacts.
These are interactions that form only in the life state; they don’t normally form under
physiological conditions between 2 proteins that aren’t normally close together in a cell.
So, there are new ways that have been developed to try and overcome some of these disadvantage
is and help to find better protein complex information. The class of experiment that
we’re going to talk about today is proximity-dependent biotinylation or PDB experiments. In these
experiments, there’s expression of a fusion protein of the bait with an enzyme that catalyzes
activation of biotin or phenolic biotin derivative. These are either going to be biotin ligases
or peroxidases. Expression of the bait fusion with the relevant context and using a biotin
substrate leads to biotinylation of proteins that are close by in the cell. Lysis of the
cell can then happen under potentially harsh conditions, but you can get full solubilization
of membrane proteins and capture of biotinylated proteins by streptavidin pull down. This means
that unstable or transient complexes could still be tagged and are still able to be pulled
down and so they can find interactors regardless of their affinity or their presence in the
membrane. So, let’s look at some of these examples. Biotin ligases are commonly used
such as the BirA or BioID tags. In these methods, they use activation of biotin to biotin oil
AMP. These enzymes have been modified so they’re unable to catalyze direct transfer to a specific
residue but instead they release a cloud of activated biotin which reacts with epsilon
amines on nearby proteins within approximately 10 nanometers of labeling radius. And the
other side is peroxidases such as the Apex and Apex 2 enzymes which catalyzed tyrosine
attack by phenolic compounds which are coupled to biotin which are activated by peroxide.
These have a slightly larger labeling radius of about 20 nanometers. These are similar
though so the major difference in the 2 methods is the time required for sufficient signal
for mass spec analysis. BioID experiments are generally performed with labeling of 12
to 24 hours, although the new turbo ID and many turbo tags can generate strong signals
reportedly in minutes. In contrast, the Apex 2 methods which use peroxide are definitely
performed on the time scale of minutes, so really the approach to go with depends on
the type of experiment and what type of labeling speed you need. Just to go again into an example
as you can see on this slide, what you have is either BioID or Apex where there are proteins
in complex. When they’re activated, those proximal proteins get biotinylated and that
can then be pulled down on streptavidin sepharose for mass spectrometry analysis. Since these
streptavidin biotin interactions is so tight, these experiments can be performed under harsh
to denaturing conditions for the cell lysis and also understand stringent washing conditions
to help bring down the background. Looking here at the cell, you can see again this is
the scale in which things happen. The nucleolus is on the scale of 2 to 3 microns so that
when we’re getting labeling radii of 10 to 20 nanometers this really is a small labeling
radius and so we can get good modification within protein complexes or within organelles.
There are some, of course, important controls and things to consider with these types of
experiments just as with any experiment. First is the background. There are, of course, natively
biotinylated proteins in the cell so one has to consider the ability to identify those
as well. They’re also going to be proteins that are promiscuously biotinylated due to
some mislocalization of the fusion protein as well as proteins that will nonspecifically
interact with the resin. So, because of these proteins that contribute to the background,
controls are very important. Controls to determine endogenous biotinylations such as the bait
protein without addition of the enzyme or untransfected cells are often useful. Controls
of the conditions that look at promiscuous biotinylation are also helpful. In this case,
that would be just the enzyme alone or effused to an irrelevant protein that’s unrelated
to what you’re studying. And furthermore, replicate analyses really help. These background
proteins are going to tend to be low or random presence so that when doing replicate experiments
true interactors should come out time and time again. There are also some caveats to
these experiments that are important to keep in mind. The residence time of proteins in
the complex and the number of solvent-exposed reactive residues will somewhat determine
how detectable proteins can be. Labeling also depends on the protein and complex side. So,
for example, if you have a very large complex, not all members will be labeled by a single
bait and so multiple different baits may be required. Further, for bait proteins that
are embedded in membranes or localized to organelles, abundant proteins that are present
in those locations but are not necessarily direct interactors will likely still be present
and may dominate that identification list. Just because there are a lot of them, they’re
going to tend to be able to be labeled. So, in that case, an extra experimental control
in which proteins with the same localization are used but different interactors or even
a mutant form of the bait protein may help to distinguish those possibilities. So, this
is an example using BioID where the authors were using this method to look at proteins
associating with mRNA biology looking at specifically proteins in RNA granules either stress granules
or processing bodies. Using the methodology, they were able to find over 7,000 unique interactions
with nearly 1800 proteins. Analysis of the correlated patterns between the preys uncovered
spatial organization of RNA regulatory structures and enabled definition of core components
of both the stress granules and the processing bodies. So, the method here was really applied
nicely to these different subcellular localization groups and able to identify what actually
constitutes the different bodies. In contrast, this is an example using Apex to study g-protein
coupled receptor signaling. First, the authors established that the Apex method could be
used to characterize complexes of their GPCR after treatment with the agonist and with
the antagonist and identified approximately 1200 protein interactors in the agonists-treated
cells including small g-proteins. None of those interactors were enriched in the antagonist-treated
cells. So, knowing that the method worked, they could then use it to do a time-resolved
experiment where they were able to track the interactions at 10 different time points using
10-second intervals for the first 50 seconds and then longer intervals up to 30 minutes.
In addition, they had one sample with no ligand treatment at all as a control. And so, using
the Apex method and the fact that it works over very short time scales, they were able
to track proximity kinetics for over 1,000 proteins. They were able to look at enzycytosis
and subcellular localization over longer time scales looking at the changes in the interactors
known to be involved in the process or in the localization. And so, this is a very unique
experiment to be able to look at the effects of protein signaling complexes over time,
look at how the complexes form and fall apart and it’s really a method that can only be
done with this type of approach and experimental design. So it’s just something to think about
when planning your experiment to talk to someone to ask what their preferences are and what
methods could be best applied to your own experiments. And for that I would suggest
in the NCI intramural program you can always talk to me, Lisa Jenkins or my colleague Thorkell
Andresson up in Frederick and we would be happy to discuss with you about your phosophopeptide
needs and your analysis desires. Within the NIH community you can also visit the CREx
for more information about available resources. Thank you very much for your time attention. So, for more information about these types
of methods and to think about how you could use them in your own research, please feel
free to contact your local mass spectrometerist. Within the NCI, you can talk to me, Lisa Jenkins
or my colleague, Thorkell Andresson in Frederick. Within the wider NIH community, I would direct
you to visit the [inaudible] for more information about available resources. Thank you very
much for your attention.>>U.S. Department of Health and Human Services,
National Institutes of Health, National Cancer Institute, cancer.gov 1-800-4CANCER. Produced
August 2019.

, , , , , ,

Post navigation

Leave a Reply

Your email address will not be published. Required fields are marked *