2.2. Interfaces revealed by UAA incorporation and crosslinking
coupled to MS
While the UAA-mediated photo-crosslinking technique facilitates a scan
of interfacial residues on one side of the biomolecular complex,
specific locations of the crosslinks made at the other binding molecules
remain elusive. This is because photo-reactive UAA can be
site-specifically incorporated in the target protein but its reactivity
upon irradiation towards a binding partner protein is non-specific.
Analytic improvements made through enzymatic fragmentation of the
crosslinked complex followed by fingerprinting by high-resolution mass
spectrometry opened up a new arena of probing two-sided binding
interfaces of a protein complex with better resolution and fidelity
(Figure 2B). For example, to investigate the mechanistic details of
lipopolysaccharide transport through the integral membrane protein
LptD/E complex, LptE was site-specifically modified with an UAA
photo-crosslinker at various positions. The covalent intermolecular
complexes LptD/E formed upon irradiation were observed in Western
blotting and then subjected to tryptic digestion followed by
high-resolution two-dimensional MALDI-TOF/TOF spectrometry. The UAA
incorporated at the 150th position of LptE was found
to be linked to S538 in LptD, specifically suggesting that a putative
extracellular loop of the LptD β-barrel spanning a residue S538 was a
major motif of interaction with LptE (Freinkman et al., 2011). In case a
ligand is a small molecule that is amenable to derivatization with a
photo-reactive group, the interface mapping can be performed without
generating multiple protein mutants bearing a photo-reactive UAA.
Diazirine-bearing cholesterol acting as a photoactivatable crosslinker
as well as a cofactor for the glycine receptor was employed to
investigate how the landscape encompassing contact residues between the
cholesterol and the glycine receptor would vary depending on a
cholesterol concentration. Crosslinks obtained after trypsin digestion
were analyzed by the multidimensional MS to identify specific amino
acids of the glycine receptor linked to the cholesterol with high
sensitivity and subsequently map the cholesterol-protein interface
altering as a function of cholesterol composition (Ferraro and Cascio
2018). A bifunctional UAA bearing a diazirine, for photo-crosslinking,
and a terminal alkyne group, for bioorthogonal tagging, has been
reported (Yang et al., 2020). A ligand harboring the bifunctional UAA
captured a target protein by photo-crosslinking and then could be tagged
with an affinity motif such as biotin by alkyne-azide click chemistry.
This strategy might enable the enrichment of ligand-target peptide
conjugates after digestion and purification for high resolution MS,
allowing efficient mapping of biological interfaces on both sides of a
ligand and a target.
Although the mass spectrometry greatly enhances the abundance of
structural information obtained from the crosslinking, it often suffers
from a high rate of false-positive hits and from missed true-positive
hits due to complicated mass spectra to interpret, in particular when
the photo-crosslinking is performed under living conditions, i.e., with
high contamination backgrounds. To increase the confidence level of data
analyses, a novel photo-reactive UAA bearing a transferable chemical
label has been developed. The UAA introduced site-specifically to a
target protein, upon interaction with a binding partner, makes a
crosslink and then undergoes an oxidative cleavage leaving the label on
the interface of a binding partner. This technique enables the
identification of true-positive mass fragments with high confidence and
the exact contact residues on both sides of interfaces simultaneously
(He et al., 2017; Yang et al., 2016; Yang et al., 2017).
As demonstrated in studies reviewed above, the UAA-mediated crosslinking
method genetically installs a reactive chemical handle directly into a
predicted interfacial residue of interest, and induces interfacial
crosslinkage of a protein complex in a controlled manner. As it require
no exogenous chemical reagents at the time of crosslinking, diffusional
and steric limitations that otherwise would adversely affect data
fidelity can be avoided. Despite such unique advantages it provides,
however, some cares need to be taken to address a potential risk of poor
expression yield of UAA-incorporated proteins and/or structural
perturbation of a native interface by UAA mutation. Abovementioned and
related studies are summarized in the header row of ‘photo-crosslinking
UAA’ (Table 1).