Engineering bromodomains with a photoactive amino acid by engaging 'Privileged' tRNA synthetases

Article abstract of DOI:10.1039/c9cc09891g

Site-specific placement of unnatural amino acids, particularly those responsive to light, offers an elegant approach to control protein function and capture their fleeting 'interactome'. Herein, we have resurrected 4-(trifluoromethyldiazirinyl)-phenylalan

Full text of DOI:10.1039/c9cc09891g

ChemComm
View Article Online
View Journal
COMMUNICATION
Engineering bromodomains with a photoactive amino
acid by engaging ‘Privileged’ tRNA synthetases†
Cite this:DOI: 10.1039/c9cc09891g
Shana Wagner, Babu Sudhamalla,‡ Philip Mannes, Sushma Sappa, Sam Kavoosi,
Received 20th December 2019,
Accepted 19th February 2020
Debasis Dey, Sinan Wang and Kabirul Islam
*
DOI: 10.1039/c9cc09891g
rsc.li/chemcomm
Site-specific placement of unnatural amino acids, particularly those
responsive to light, offers an elegant approach to control protein
function and capture their fleeting ‘interactome’. Herein, we have
resurrected 4-(trifluoromethyldiazirinyl)-phenylalanine, an underutilized
photo-crosslinker, by introducing several key features including easy
synthetic access, site-specific incorporation by ‘privileged’ synthetases
and superior crosslinking efficiency, to develop photo-crosslinkable
bromodomains suitable for ‘interactome’ profiling.
Dynamic biological processes rely on transient protein–protein
and protein-nucleic acid interactions occurring in precise space
and time. Gaining molecular insights into such processes
requires characterization of the specific interacting partners
in a physiologically relevant environment. Proteins and oligo-
nucleotides carrying photoactive building blocks have been
developed to capture transient interactions.1 A particularly
important approach includes site-specific introduction of a
photo-crosslinkable amino acid (PCAA) or a photo-controlled
chemical crosslinker by the read-through of amber stop codon
using an evolved synthetase-tRNA pair (Fig. 1A).2–5
While PCAAs 4-benzoylphenylalanine (BzF) 1, 4-azidophenyl-
alanine (AzF) 2 and 4-(trifluoromethyldiazirinyl)-phenylalanine
(tmdF) 3 have been used to probe protein interactions, each
possesses distinct advantages and limitations (Fig. 1B–D).3 1 pro-
vides the maximum degree of crosslinking owing to its regeneration
potential in aqueous medium; however, having a larger size, it often
interferes with the binding of interacting partners. In contrast, 2 is
closest in size to phenylalanine and tyrosine, but it suffers in
crosslinking efficiency due to intramolecular insertion of the nitrene
intermediate resulting in ring expansion. Compared to 1 and 2,
Fig. 1 Structure and reactivity of photo-crosslinkable amino acids (PCAAs).
(A) Incorporation of PCAAs such as 1, 2 and 3 at a specific site into proteins is
achieved using amber suppressor mutagenesis to form a covalent bond
with interacting partners upon photo-irradiation. (B) In presence of UV light,
BzF generates a diradical which, in addition to undergoing crosslinking with
interactome, could regenerate BzF through hydration. (C) AzF generates
highly unstable nitrene which primarily decomposes to a ring-expanded
inactive intermediate. (D) Upon photo-irradiation, tmdF furnishes a carbene
stabilized by the phenyl and trifluoromethyl groups.
PCAA 3, which generates a carbene upon irradiation, is the least utilized photo-crosslinker mainly because of the challenges in
synthetic access and site-specific incorporation into proteins.
Department of Chemistry, University of Pittsburgh, Pittsburgh, PA 15260, USA.
E-mail: kai27@pitt.edu
Furthermore, a comparative analysis of crosslinking efficiency of
these PCAAs in the context of protein–protein interaction is
largely unexplored.
We recently developed a chemoproteomic approach called
‘interaction-based protein profiling’ (IBPP) by introducing 2
into bromodomain-containing protein BRD4 that recognizes
† Electronic supplementary information (ESI) available. CCDC 1978317. For ESI
and crystallographic data in CIF or other electronic format see DOI: 10.1039/
c9cc09891g
‡ Present address: Department of Biological Sciences, Indian Institute of Science
Education and Research-Kolkata, Mohanpur, West Bengal, 741246, India.
Chem. Commun.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Communication
ChemComm
acetylated histones as well as non-histone proteins such as synthetases are capable of charging their cognate tRNAs with
ILF3, GATA1, RelA to regulate gene expression.6,7 Interestingly, multiple unnatural amino acids (UAAs), one at a time, while
BRD4 carrying 1 failed to bind its cognate interacting partners, maintaining high fidelity by not recognizing the canonical amino
such as histone H4, likely due to increased steric bulk of 1. To acids.15 Polysubstrate specificity (often called ‘permissivity’) of such
expand the repertoire of PCAAs and compare their crosslinking a ‘privileged’ synthetase is the result of lack of evolutionary pressure
efficiency, we sought to introduce 3 in a site-specific manner against other UAAs during in vitro selections to evolve the
into a set of bromodomain-containing proteins.
synthetase.16,17 We envisioned exploiting the ‘permissivity’ of
Schultz and colleagues have introduced 3 into a Z domain the M. jannaschii synthetases developed by the Schultz group to
protein using a pBK-based expression vector carrying the evolved introduce 3 into bromodomain-containing proteins.
synthetase under the constitutively active glnS promoter.8 The
We selected a panel of four engineered synthetases and their
suppressor tRNA and gene of interest were inserted into pLei cognate tRNAs cloned in the pEVOL-based expression system.
vector under the lpp promoter. Schultz group later developed a These evolved pairs have been shown to incorporate bulky para-
pEVOL-based single vector carrying both the evolved synthetase substituted tyrosine analogues including 4-acetylphenylalanine
and the suppressor tRNA under constitutive as well as inducible (AcF), 4-cyanophenylalanine (CNF), BzF 1 and AzF 2 into
promoters.9,10 It has been shown that pEVOL-based vector leads proteins.9,18,19 Given the structural similarity of 3 to these analogues,
to 250% increase in expression of the desired protein compared we hypothesized that 3 could be accepted by some of these
to the pBK-based construct. We reasoned that employing a ‘privileged’ synthetases to acylate the orthogonal tRNAs with 3. For
pEVOL-based expression system for 3 would simplify its incor- an initial screening, we used BRD4-W81TAG mutant. Under the
poration into proteins and accelerate crosslinking-based ‘inter- optimized expression condition, we observed efficient read-through
actome’ characterization.
of the amber codon by the CNF and AzF synthetases only in the
We also recognized that previous attempts to access 3 involved presence of 3 (Fig. 2A and Fig. S2, ESI†).
up to twelve chemical steps, occasionally with an enzymatic hydro-
It has been shown that CNF-specific tRNA synthetase exhibits
lysis of an acyl group, to furnish the amino acid.11–13 We realized broader substrate scope due to a larger binding pocket, while
that use of compound 3 as a routine photo-crosslinker would maintaining its ability to discriminate against the 20 canonical
benefit from an improved synthesis. Our scheme commenced from amino acids.15 To further understand the binding of 3 into this
toluene 4 that was converted to trifluorodiaziridine 8 via the ‘privileged’ synthetase, we performed a docking study using
intermediacy of 5, 6 and 7 in four steps following reported methods Autodock energy minimization software. For this purpose, we
(Scheme 1).11,12 Although oxidation of 8 to diazirine 9 can be used the coordinates from CNF-bound synthetase crystal struc-
accomplished in multiple ways, in our hands, the previous ture (PDB 3qe4). In the most stable conformations, both the
approaches resulted in inconsistent yields often with no product enantiomers of 3 occupy the CNF binding site and forms an
formation. We found that pyridinium dichromate (PDC) oxidation array of polar contacts, similar to CNF, with residues in the
of 8 was smooth and furnished 9 in excellent yield. We modified active site (Fig. 2B, C and Fig. S3, ESI†). The CF3 group is situated
the subsequent steps, particularly the C–C bond formation between in the hydrophobic pocket created by L32, V65, M107, G158 and
10 and protected glycine 11 under kinetically controlled condition
to afford protected, fully-assembled diazirine 12.14 The final depro-
tection proceeded smoothly to furnish (Æ)-3 in eight linear steps
with 10% overall yield with high purity as judged by analytical data
(Scheme 1 and Fig. S10–S22, ESI†). We confirmed the structure of
(Æ)-3 by X-ray crystallography (CCDC 1978317†) (Fig. S1, ESI†).
After accessing 3 in a large quantity, we set out to uncover a
pEVOL-based system for site-specific incorporation of 3 into
proteins. It has been demonstrated that certain evolved
Fig. 2 Efficient incorporation of 3 into BRD4 bromodomain by ‘privileged’
synthetases. (A) Expression of BRD4-W81tmdF mutant by a set of evolved
Tyr
M. jannaschii TyrRS-tRNACUA
pairs either in absence or presence of
tmdF 3 as judged by Coomassie blue staining (N = 2). (B) Energy-
minimized structure of S enantiomer of 3 in the catalytic pocket of evolved
CNF synthetase. (C) Polar contacts between tmdF 3 and the active site
residues. (D) Binding of tmdF in the conserved, extended hydrophobic
pocket of evolved synthetase.
Scheme 1 Synthesis of (Æ)-tmdF 3.
Chem. Commun.
This journal is © The Royal Society of Chemistry 2020

View Article Online
ChemComm
Communication
W2950tmdF mutants dissociated from the peptide with Kd of
309 Æ 23 mM and 231 Æ 21 mM, respectively (Fig. 3C and Fig. S7,
ESI†).
Next, to examine photo-crosslinking efficiency of the BPTF
analogues, we synthesized tetraacetylated histone H4 peptide
14 carrying a tetramethylrhodamine (TAMRA) moiety at the
N terminal (Fig. 4A and Fig. S6, ESI†). The tracer peptide was
incubated separately with wild type and mutant BPTF proteins
and then exposed to UV light at 365 nm wavelength. Subsequent
in-gel fluorescence experiment revealed a protein band corres-
ponding to the crosslinked species of BPTF mutant and the
peptide only in the presence of UV light (Fig. S8, ESI†). This
crosslinking ability of the mutants must have been acquired
through incorporation of tmdF in the binding pocket because
fluorescent labeling of wild type BPTF was undetectable in spite
of its strong affinity towards the peptide (Fig. S8, ESI†).
A careful time-dependent experiment revealed that tmdF-
containing BPTF underwent enhanced crosslinking compared
to the AzF counterpart over a period of photo-irradiation
(Fig. 4B, C and Fig. S8, ESI†). This is likely due to an extremely
Fig. 3 Incorporation of 3 into BPTF bromodomain by evolved synthetases.
(A) Domain structure of BPTF. (B) Bacterial expression of BPTF-W2950tmdF
mutant as observed with Coomassie blue staining and confirmed by ESI-
HRMS. (C) Dissociation constants (Kd) of BPTF and its W2950tmdF mutant
from tetra-acetylated H4Kac4 peptide 13 as measured by ITC.
A159 (Fig. 2D). The mode of binding is consistent with the short-lived nitrene intermediate generated from photolysis of
observation that these residues are important for accommodating the azido moiety, making pAzF a less effective crosslinker
UAAs carrying para-substituted phenylalanine analogues.15
(Fig. 1C).3 In contrast, tmdF generates a benzyl radical which
We next examined generality of the approach and established 3 could gain further stability by the trifluoromethyl group
as an efficient photo-crosslinker. The bromodomain and PHD (Fig. 1D). Furthermore, TAMRA-attached peptide 15 with a
finger containing transcription factor (BPTF) has been shown to sequence identical to 14, but lacking the acetyl moieties, did
remodel chromatin structure.20 The bromodomain and the PHD not undergo crosslinking with the BPTF mutants (Fig. S8, ESI†).
finger bind acetylated histone H4 and methylated histone H3, To the best of our knowledge, this is the first demonstration of
respectively (Fig. 3A). We recently developed a BPTF bromodomain comparative crosslinking efficiency of 2 and 3 in context of
analogue carrying AzF 2 at 2950th position and demonstrated protein–protein interaction.
its ability to crosslink histone variant H2A.Z.21 We sought to
We next sought to probe if the engineered BPTF is capable of
incorporate 3 into this bromodomain using the ‘privileged’ interacting with a full-length histone protein. To access site-
synthetase, examine the binding potential of engineered BPTF specifically acetylated histone H4, we employed a thermal thiol–
to histone protein and evaluate the effectiveness of 3 as a
photo-crosslinker.
An amber stop codon (TAG) was introduced to replace
W2950 in the BPTF bromodomain. This residue is situated in
the aromatic cage known to house acetylated lysines. We
observed efficient incorporation of 3 into W2950 site by both
CNF- and AzF-specific synthetases (Fig. 3B and Fig. S4, ESI†).
Integrity of the mutant protein was confirmed by LC-HRMS
(Fig. 3B and Fig. S5, ESI†). This result demonstrates that
incorporation of 3 into multiple proteins (BRD4, BPTF and
likely others) is achievable using the ‘privileged’ synthetases
evolved to incorporate related UAAs.
We then tested the BPTF analogue for its ability to bind
interacting partners. To compare the binding and crosslinking
efficiency, we also generated W2950AzF mutant carrying 2
using AzF-specific synthetase and characterized by LC-HRMS
Fig. 4 Crosslinking of BPTF analogues with histone peptide. (A) Schematic
(Fig. S5, ESI†). Using a synthesized tetracetylated histone H4
depicting binding of engineered BPTF proteins to TAMRA-H4Kac4 peptide
peptide 13 (Fig. 3C and Fig. S6, ESI†), we measured dissociation 14 followed by crosslinking. Fluorescently labeled crosslinked BPTF was
visualized by in-gel fluorescence. (B) In-gel fluorescence showing cross-
constants (Kd) of the BPTF proteins from the peptide using
isothermal titration calorimetry (ITC). Wild-type BPTF bound
linking of BPTF analogues to the TAMRA-H4Kac4 peptide. Coomassie
staining of the same gel showed presence of proteins in all the samples
(for biological replicates, see Fig. S8C, ESI†). (C) Relative quantification of
crosslinking, based on the results shown in B and Fig. S8C (ESI†), repre-
the peptide with a Kd value of 185 Æ 10 mM, closely agreeing
with reported values,6,20 while the mutants displayed varied
degrees of binding (Fig. 3C and Fig. S7, ESI†). W2950AzF and sented as normalized fluorescence intensity (error bar = standard deviation).
Chem. Commun.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Communication
ChemComm
acetylated histone protein and its superior crosslinking ability over
arylazide functionality. These improvements coupled with the
recent finding that tmdF could be introduced into mammalian
cells23 would accelerate the application of this particular amino
acid for profiling transient interacting partners of the bromo-
domains and other chromatin modifiers.
We thank the University of Pittsburgh, the National Institutes
of Health (R01GM123234) and the National Science Foundation
(MCB-1817692) for financial support; S. Geib for crystal structure
analysis, Dr D. Chakraborty and members of our laboratory for
editing of the manuscript. P. Mannes is supported by an NIH
training grant (T32GM008208).
Conflicts of interest
There are no conflicts to declare.
Fig. 5 Crosslinking of BPTF-W2950tmdF with full-length histone. (A) Scheme
showing the synthesis of H4 carrying site-specific acetylated thialysine (KCac)
using thermal thiol–ene reaction (ref. 22). (B) ESI LC-HRMS spectra of full-
length H4KC8ac. (C) Western blot showing crosslinking of BPTF-W2950tmdF
to H4KC8ac and H4KC12ac (dotted box) using anti-H4 and anti-6xHis anti-
bodies. (D) Crosslinking of hyperacetylated H4 isolated from HEK293T cells
with BPTF mutants as confirmed by H4 antibody.
Notes and references
1 N. D. Pham, R. B. Parker and J. J. Kohler, Curr. Opin. Chem. Biol.,
2013, 17, 90–101.
2 H. Neumann, P. Neumann-Staubitz, A. Witte and D. Summerer,
Curr. Opin. Chem. Biol., 2018, 45, 1–9.
3 T. A. Nguyen, M. Cigler and K. Lang, Angew. Chem., Int. Ed., 2018, 57,
14350–14361.
4 Y. Tian, M. P. Jacinto, Y. Zeng, Z. Yu, J. Qu, W. R. Liu and Q. Lin,
J. Am. Chem. Soc., 2017, 139, 6078–6081.
5 J. Liu, S. Li, N. A. Aslam, F. Zheng, B. Yang, R. Cheng, N. Wang,
S. Rozovsky, P. G. Wang, Q. Wang and L. Wang, J. Am. Chem. Soc.,
2019, 141, 9458–9462.
6 B. Sudhamalla, D. Dey, M. Breski, T. Nguyen and K. Islam, Chem.
Sci., 2017, 8, 4250–4256.
7 J. P. Lambert, et al., Mol. Cell, 2019, 73(621–638), e617.
8 E. M. Tippmann, W. Liu, D. Summerer, A. V. Mack and P. G. Schultz,
ChemBioChem, 2007, 8, 2210–2214.
9 T. S. Young, I. Ahmad, J. A. Yin and P. G. Schultz, J. Mol. Biol., 2010,
395, 361–374.
10 A. Chatterjee, S. B. Sun, J. L. Furman, H. Xiao and P. G. Schultz,
Biochemistry, 2013, 52, 1828–1837.
ene reaction to produce acetylated thialysine and synthesized full-
length H4Kc8ac and H4Kc12ac proteins (Fig. 5A, B and Fig. S9,
ESI†).22 The synthetic full-length H4 proteins were allowed to bind
with W2950tmdF mutant followed by UV-irradiation. The cross-
linked species were separated in SDS-PAGE and visualized using
anti-H4 and anti-6xHis antibodies. We observed chemiluminescent
signals at a molecular weight higher than that of the individual
proteins (H4 and BPTF) when irradiated with UV light, confirming
successful crosslinking (Fig. 5C). No crosslinking was observed
in absence of photo-irradiation or when the individual proteins 11 M. Nassal, J. Am. Chem. Soc., 1984, 106, 7540–7545.
12 G. Baldini, B. Martoglio, A. Schachenmann, C. Zugliani and
were exposed to UV light. Finally, we extracted endogenous hyper-
J. Brunner, Biochemistry, 1988, 27, 7951–7959.
acetylated histones from HEK39T cells and subjected them to
13 K. Masuda, A. Koizumi, T. Misaka, Y. Hatanaka, K. Abe, T. Tanaka,
crosslinking with the BPTF analogues. Western blotting with anti-
H4 antibody indeed showed crosslinking, albeit to a lesser extent,
for both the mutants with W2950tmdF having marginally increased
M. Ishiguro and M. Hashimoto, Bioorg. Med. Chem. Lett., 2010, 20,
1081–1083.
14 H. Nakashima, M. Hashimoto, Y. Sadakane, T. Tomohiro and
Y. Hatanaka, J. Am. Chem. Soc., 2006, 128, 15092–15093.
crosslinked signal compared to the W2950AzF variant (Fig. 5D). 15 D. D. Young, T. S. Young, M. Jahnz, I. Ahmad, G. Spraggon and P. G.
Schultz, Biochemistry, 2011, 50, 1894–1900.
16 S. J. Miyake-Stoner, C. A. Refakis, J. T. Hammill, H. Lusic,
Collectively, these results demonstrate that the BPTF analogue with
site-specifically introduced PCAA 3 is capable of interacting and
J. L. Hazen, A. Deiters and R. A. Mehl, Biochemistry, 2010, 49,
establishing covalent bond with full-length histone H4 upon
irradiation by UV light.
1667–1677.
17 A. L. Stokes, S. J. Miyake-Stoner, J. C. Peeler, D. P. Nguyen,
R. P. Hammer and R. A. Mehl, Mol. BioSyst., 2009, 5, 1032–1038.
18 J. W. Chin, S. W. Santoro, A. B. Martin, D. S. King, L. Wang and
P. G. Schultz, J. Am. Chem. Soc., 2002, 124, 9026–9027.
19 J. W. Chin, A. B. Martin, D. S. King, L. Wang and P. G. Schultz, Proc.
Natl. Acad. Sci. U. S. A., 2002, 99, 11020–11024.
20 A. J. Ruthenburg, H. Li, T. A. Milne, S. Dewell, R. K. McGinty,
M. Yuen, B. Ueberheide, Y. Dou, T. W. Muir, D. J. Patel and
C. D. Allis, Cell, 2011, 145, 692–706.
In this work, we established tmdF as an excellent photo-
sensitive amino acid which has remained largely underutilized in
‘interactome’ profiling. We repurposed a set of ‘privileged’ synthe-
tases to incorporate tmdF into bromodomain-containing proteins.
This approach to expand substrate polyspecificity of an already
evolved synthetase has the benefit of accessing engineered transla-
tional machinery without having to perform additional steps
21 G. T. Perell, N. K. Mishra, B. Sudhamalla, P. D. Ycas, K. Islam and
W. C. K. Pomerantz, Biochemistry, 2017, 56, 4607–4615.
involved in directed evolution. We also improved the existing 22 A. Dhall, S. Wei, B. Fierz, C. L. Woodcock, T. H. Lee and
C. Chatterjee, J. Biol. Chem., 2014, 289, 33827–33837.
23 N. Hino, M. Oyama, A. Sato, T. Mukai, F. Iraha, A. Hayashi,
synthetic strategy towards tmdF by modifying a few critical steps
and reducing the number of protection-deprotection maneuvers.
H. Kozuka-Hata, T. Yamamoto, S. Yokoyama and K. Sakamoto,
Finally, we demonstrated the utility of tmdF group in crosslinking
J. Mol. Biol., 2011, 406, 343–353.
Chem. Commun.
This journal is © The Royal Society of Chemistry 2020