Critical protein-protein interactions are ubiquitous in biology. To provide a new method to detect these interactions, we designed and synthesized fluorinated bromopyronins as molecular probes. These electrophilic compounds rapidly react with amines via a S(N)Ar mechanism to form modestly electrophilic aminopyronin fluorophores. To investigate whether proteins modified with aminopyronins might selectively transfer these fluorophores between proximal lysine residues at protein-protein interfaces, immunoglobulin-G (IgG) was conjugated to fluorinated pyronins and added to unlabeled Protein A (SpA) from S. aureus. Analysis by gel electrophoresis and mass spectrometry revealed transfer of this fluorophore from IgG to specific lysines of its binding partner SpA but not to bovine serum albumin (BSA) as a nonbinding control. Examination of an X-ray structure of IgG bound to SpA revealed that the fluorophore was selectively transferred between amino groups of lysines that reside within ~10 Å at the protein-protein interface. To evaluate whether this approach could be used to identify interactions with endogenous cellular proteins, pyronin-modified Rnase A was added to crude extracts of human HeLa cells. Analysis of interacting proteins by gel electrophoresis revealed the endogenous ribonuclease inhibitor as the primary cellular target. Given that proximal lysine residues frequently reside at protein-protein interfaces, this method may facilitate identification of diverse protein-protein interactions present in complex biological matrices.
Critical protein-protein interactions are ubiquitous in biology. To provide a new method to detect these interactions, we designed and synthesized fluorinated bromopyronins as molecular probes. These electrophilic compounds rapidly react with amines via a S(N)Ar mechanism to form modestly electrophilic aminopyronin fluorophores. To investigate whether proteins modified with aminopyronins might selectively transfer these fluorophores between proximal lysine residues at protein-protein interfaces, immunoglobulin-G (IgG) was conjugated to fluorinated pyronins and added to unlabeled Protein A (SpA) from S. aureus. Analysis by gel electrophoresis and mass spectrometry revealed transfer of this fluorophore from IgG to specific lysines of its binding partner SpA but not to bovineserum albumin (BSA) as a nonbinding control. Examination of an X-ray structure of IgG bound to SpA revealed that the fluorophore was selectively transferred between amino groups of lysines that reside within ~10 Å at the protein-protein interface. To evaluate whether this approach could be used to identify interactions with endogenous cellular proteins, pyronin-modified Rnase A was added to crude extracts of humanHeLa cells. Analysis of interacting proteins by gel electrophoresis revealed the endogenous ribonuclease inhibitor as the primary cellular target. Given that proximal lysine residues frequently reside at protein-protein interfaces, this method may facilitate identification of diverse protein-protein interactions present in complex biological matrices.
Protein–protein
interactions
control diverse biological processes. To detect these interactions,
numerous methods have been used including a variety of fluorescence-based
techniques.[1] If a protein is covalently
labeled with a fluorophore, interaction with a second protein can
potentially alter the intensity, wavelength, or polarization of the
fluorophore, enabling detection of the interaction. However, this
approach generally requires that the primary protein be labeled with
a single fluorophore on a unique residue that is sensitive to complexation.
If proteins are more extensively labeled by conjugation to multiple
lysines or other residues, spectroscopic changes at the fluorophore
most affected by the protein–protein interaction can be difficult
or impossible to detect above background fluorescence.Proximity-driven transfer
of fluorinated pyronin fluorophores between
lysine residues for detection of protein–protein interactions.We report here a new method for
detection of protein–protein
interactions. This method allows lysine residues on the surface of
a protein of interest to be randomly or extensively conjugated to
novel fluorinated fluorophores. These fluorophores are designed to
form stable protein conjugates. However, the high effective concentration
of proximal lysines at interfaces of protein–protein complexes
provides a driving force for reversible exchange between lysines of
interacting proteins. This transfer of a fluorophore from a donor
to an acceptor protein can potentially be detected by gel electrophoresis
and/or proteomics methods. Several methods for proximity-driven alkylation
of cysteine residues,[2−5] acylation of lysine residues,[6−8] and other reactions of proteins[9−11] have been described and reviewed,[12,13] but proximity-driven
exchange of fluorophores between Lys residues at protein–protein
interfaces has not been previously reported. This new approach is
illustrated in Figure 1.
Figure 1
Proximity-driven transfer
of fluorinated pyronin fluorophores between
lysine residues for detection of protein–protein interactions.
We recently published[14] the synthesis
of fluorinated xanthones such as 1–3 (Scheme 1) via iterative nucleophilic aromatic
substitution. To create novel reagents for protein modification, we
treated these xanthones with triflic anhydride and tetrabutylammonium
bromide to generate the fluorinated bromopyronins 4–6. These compounds are stable solids, but they
readily react with amines to yield fluorinated aminopyronins such
as the ethylamine derivatives 7–9. Similar to other known pyronins,[15,16] these compounds
are highly fluorescent (Figure 2). Furthermore, 7–9 exhibit large Stokes shifts of ∼100
nm, which can beneficially limit self-quenching when multiple fluorophores
are conjugated to biomolecules,[17] and high
quantum yields of ∼0.8 were measured for both 7 and 8. The absorbance spectra of 7–9 were insensitive to changes in pH in the range of 4–9,
but deprotonation was observed under more basic conditions, and pKa values of 9.9 to 11.2 were measured (Figure 2, pH-dependent spectra are provided in Figure S4
of the Supporting Information). Analysis
of reactivity by mass spectrometry revealed that aminopyronins 7–9 are highly stable in buffered aqueous
solutions under physiological conditions (pH 7.4). Moreover, treatment
of 7–9 (50 μM) with a high
concentration (10 mM) of methylamine in buffered water (pH 7.4) resulted
in <1% conversion to 10–12 (Scheme 1) after 96 h at 37 °C. However, under the same
conditions in nonbuffered methanol/water (9:1), the addition of excess
methylamine (10 mM) resulted in clean conversion to 10–12, with half-times of 1.9 to 3.8 h under pseudo-first-order
conditions (Figure 2C), resulting from nucleophilic
aromatic substitution. Under these conditions, the more hindered pyronin 9 was ∼2-fold more reactive than 7 or 8, and this increased reactivity was correlated with a lower
quantum yield and decreased pKa, consistent
with less conjugation of the t-Bu-substituted amine.
Scheme 1
Synthesis of Bromopyronins
(4–6)
and Aminopyronins (7–12)
Figure 2
(A) Absorbance and fluorescence emission
spectra of pyronins 5 and 8 in PBS (pH 7.0).
(B) Physicochemical
properties of 4–9 in PBS. (C) Kinetics
of conversion of 7–9 to 10–12 upon reaction with excess methylamine.
We reasoned that the high stability of aminopyronins 7–9 under physiological conditions, in conjunction
with their high reactivity toward amines under more basic conditions,
might be useful for detection of protein–protein interactions.
Exchange of these fluorophores between proximal lysine residues at
protein–protein interfaces, driven by the high effective concentration
of donor and acceptor lysines, might reveal interacting proteins by
the labeling of protein partners.(A) Absorbance and fluorescence emission
spectra of pyronins 5 and 8 in PBS (pH 7.0).
(B) Physicochemical
properties of 4–9 in PBS. (C) Kinetics
of conversion of 7–9 to 10–12 upon reaction with excess methylamine.To evaluate this hypothesis in
a model system, we conjugated bromopyronins 4–6 to lysines of human IgG protein (150
kDa) to afford conjugates 13–15 (Figure 3) at fluorophore:antibody ratios of 3–4.
These conjugates were purified by size-exclusion chromatography and
added to an aqueous buffer (pH 7.4) containing both Protein A (SpA,
42 kDa) from S. aureus and bovineserum albumin (BSA,
66 kDa). SpA binds the Fc region of human IgG with high affinity,[18] whereas BSA was included as a nonbinding but
Lys-containing internal control. As shown in Figure 3, transfer of the fluorescent tag from IgG to SpA, but not
BSA, was observed both qualitatively and quantitatively by SDS-PAGE.
Although all three of these conjugates showed some transfer to SpA
after 16 h, 14 labeled SpA to the greatest extent (Figure 3), and this conjugate was the focus of subsequent
studies because it was more reactive than 13 and exhibited
a higher quantum yield compared with 15. Moreover, in
further support of the mechanism shown in Figure 1, when lysines of IgG were conjugated to 5-carboxyfluorescein
or fluorescein isothiocyanate (FITC) through more stable
amide or thiourea bonds, no transfer of fluorophore from IgG to SpA
was observed after 16 h (Figure S6, Supporting
Information). These controls further confirm that the unique
reactivity of these pyronin fluorophores toward amines is critical
for detection of protein–protein interactions.
Figure 3
Analysis of transfer of pyronins between
proteins. Human IgG was
conjugated to pyronins 4 (A), 5 (B), and 6 (C), added at 5 μM to PBS (pH 7.4) containing unlabeled
SpA (5 μM) and BSA (5 μM), allowed to react at 37 °C
for 0–16 h, and analyzed by SDS-PAGE with fluorescence detection.
Quantification of the fluorescence of the SpA bands is shown on the
bottom.
SpA contains
five homologous domains with an affinity for the constant
Fc region of most IgG subclasses.[18] This
bacterial protein also binds the variable Fab fragments of some IgGs.[19] However, conjugation of 5 to purified
Fc and Fab regions of human IgG revealed that the labeled Fc fragment
exclusively transferred the pyronin to SpA (Figure S5, Supporting Information). To further examine the
mechanism of transfer of pyronins from IgG to SpA, after the fluorophore
was transferred, unreacted conjugate 14 was removed from
labeled SpA using resin-bound SpA, and modification of specific SpA
residues was analyzed by tryptic digestion and MS-based protein sequencing.
Remarkably, as shown in Figure 4, among the
118 total lysine residues of mature SpA, only four homologous residues,
Lys-69, Lys-130, Lys-188, Lys-246, predominantly reacted with the
pyronins of IgG. To investigate the molecular basis of this pattern
of reactivity, we examined an X-ray crystal structure of the Fc region
of human IgG bound to Fragment B of SpA.[20] As shown in Figure 5, this structure revealed
that the amino groups of Lys-246 of SpA and Lys-288 of IgG uniquely
reside within 5–9 Å, depending on side-chain rotamer conformations,
and represent the most proximal lysines in the complex. This close
proximity enforced by the protein–protein interaction explains
the highly selective reactivity of residues in Fragments A, D, and
E that are homologous to SpALys-246.
Figure 4
Analysis of fluorophore transfer from IgG conjugate 14 to SpA by tryptic digestion and MS/MS. The amino acid sequence of
mature SpA (Uniprot ID P02976) is shown with fragments A–E
aligned by homology. Lysine residues are colored red. Four dominant
pyronin-modified tryptic peptides, identified with >99% probability,
are shown boxed, with arrows illustrating sites of lysine modification.
The underlined residues in fragment B represent the residues visible
in the X-ray structure of the IgG Fc-SpA complex (PDB ID 1FC2).
Figure 5
(A) X-ray structure of the Fc region of human
IgG (blue) bound
to Fragment B of SpA (red, PDB ID 1FC2). Carbohydrates of the IgG
were omitted for clarity. Lysine residues are shown as CPK models.
(B) Mechanism of molecular transfer of pyronin between lysine residues
of SpA and IgG.
Analysis of transfer of pyronins between
proteins. Human IgG was
conjugated to pyronins 4 (A), 5 (B), and 6 (C), added at 5 μM to PBS (pH 7.4) containing unlabeled
SpA (5 μM) and BSA (5 μM), allowed to react at 37 °C
for 0–16 h, and analyzed by SDS-PAGE with fluorescence detection.
Quantification of the fluorescence of the SpA bands is shown on the
bottom.To examine whether this
approach might allow detection of protein–protein interactions
in complex biological matrices such as a cell extract, we conjugated
bovine ribonuclease (Rnase) to bromopyronin 5. Rnase
binds tightly to Rnase inhibitor, and a structure of a complex formed
by these proteins is shown in Figure 6.[21] This fluorescent Rnase conjugate was added both
to purified humanribonuclease inhibitor and to a crude extract of
humanHeLa cells, to promote transfer of the fluorophore to interacting
proteins. Although the low solubility of Rnase limited the extent
of labeling to 0.3–0.6 fluorophores per protein, reducing the
sensitivity of the assay in this case, analysis by SDS-PAGE revealed
transfer of the fluorophore to both recombinant human Rnase inhibitor
and endogenous human Rnase inhibitor in HeLa extracts (Figure 6). These proteins, and oligomeric forms of the Rnase
conjugate,[22] were identified by comigration
with authentic standards. Comparison of the limited number of fluorescent
bands with the abundance of total proteins stained by coomassie blue
in the cell extract demonstrated that fluorinated pyronin conjugates
can exhibit high selectivity for the detection of specific interacting
proteins.
Figure 6
(A) X-ray structure of
bovine Rnase A (red) bound to porcine Rnase
inhibitor (blue, PDB ID 1DFJ) illustrating the proximity of lysine
residues (CPK models) in the complex. (B–C) SDS-PAGE gel imaged
by fluorescence (B) followed by staining of all proteins in the same
gel with coomassie (C). (Lane i) protein MW marker. (ii) Pyronin-conjugated
bovine Rnase A alone. (iii) Reaction of pyronin-conjugated bovine
Rnase A (20 μM) with recombinant human Rnase inhibitor (20 μM)
for 18 h at 37 °C. (iv) Reaction of pyronin-conjugated bovine
Rnase A (20 μM) with a crude HeLa cell extract for 18 h at 37
°C. The arrows show specific proteins identified by comparison
with authentic standards.
Analysis of fluorophore transfer from IgG conjugate 14 to SpA by tryptic digestion and MS/MS. The amino acid sequence of
mature SpA (Uniprot ID P02976) is shown with fragments A–E
aligned by homology. Lysine residues are colored red. Four dominant
pyronin-modified tryptic peptides, identified with >99% probability,
are shown boxed, with arrows illustrating sites of lysine modification.
The underlined residues in fragment B represent the residues visible
in the X-ray structure of the IgG Fc-SpA complex (PDB ID 1FC2).(A) X-ray structure of the Fc region of human
IgG (blue) bound
to Fragment B of SpA (red, PDB ID 1FC2). Carbohydrates of the IgG
were omitted for clarity. Lysine residues are shown as CPK models.
(B) Mechanism of molecular transfer of pyronin between lysine residues
of SpA and IgG.(A) X-ray structure of
bovineRnase A (red) bound to porcine Rnase
inhibitor (blue, PDB ID 1DFJ) illustrating the proximity of lysine
residues (CPK models) in the complex. (B–C) SDS-PAGE gel imaged
by fluorescence (B) followed by staining of all proteins in the same
gel with coomassie (C). (Lane i) protein MW marker. (ii) Pyronin-conjugated
bovineRnase A alone. (iii) Reaction of pyronin-conjugated bovineRnase A (20 μM) with recombinant human Rnase inhibitor (20 μM)
for 18 h at 37 °C. (iv) Reaction of pyronin-conjugated bovineRnase A (20 μM) with a crude HeLa cell extract for 18 h at 37
°C. The arrows show specific proteins identified by comparison
with authentic standards.We demonstrated that fluorinated pyronin fluorophores offer
unique
tools for detection of interactions between biomolecules. In a buffered
aqueous solution (pH 7.4), amino derivatives of fluorinated pyronins
can be highly stable, even in the presence of a large excess of other
soluble amines. However, the high effective concentration that results
from approximation of lysine residues at protein–protein interfaces
enables transfer of these fluorophores between proteins, providing
a means to identify interacting partners. Given that lysine is one
of the most common amino acids on protein surfaces,[23] and these residues are often in close proximity at protein–protein
interfaces, this method may be useful for the discovery of novel interactions
and factors that affect these interactions.
Authors: M Graille; E A Stura; A L Corper; B J Sutton; M J Taussig; J B Charbonnier; G J Silverman Journal: Proc Natl Acad Sci U S A Date: 2000-05-09 Impact factor: 11.205
Authors: Zhe Gao; Krishna K Sharma; Angelo E Andres; Brandon Walls; Fadel Boumelhem; Zachary R Woydziak; Blake R Peterson Journal: RSC Med Chem Date: 2022-03-07
Authors: William L Harryman; James P Hinton; Cynthia P Rubenstein; Parminder Singh; Raymond B Nagle; Sarah J Parker; Beatrice S Knudsen; Anne E Cress Journal: Biochim Biophys Acta Date: 2016-09-24