Unnatural amino acid mutagenesis for site-specific incorporation of keto and azido functionalities into G protein-coupled receptors

1
Unnatural amino acid mutagenesis for
site-specific incorporation of keto and azido
functionalities into G protein-coupled receptors
Shixin Ye, Thomas Huber, Amy Grunbeck, Thomas P.
Sakmar Laboratory of Molecular Biology
Biochemistry, The Rockefeller University, New
York, NY 10065
ABSTRACT
Results
CONCLUSIONS (chemical labeling)
Introduction The insertion of unnatural amino
acids into proteins using amber stop codon
suppression has shown promise as a technique for
probing protein structures.  To investigate
applications to studies of G protein-coupled
receptors, we have developed methods that allow
incorporation of each of three tyrosine analogues
p-acetyl-phenylalanine (Acp),
p-benzoyl-phenylalanine (Bzp), and
p-azido-phenylalanine (Azp) into GPCRs
site-specifically (1) at high yields in mammalian
cell culture.  The unique keto and azido
functionalities allow specific attachment of tags
and fluorophores into GPCRs by hydrazone (2) and
Staudinger-Bertozzi ligation (3) respectively
under physiological conditions.  Together with
cysteine-specific labeling methods, the hydrazone
and Staudinger-Bertozzi ligations will make it
possible to introduce pairs of fluorophores into
GPCRs in a general way. The benzoylphenone moiety
generates a reactive species by irradiation with
UV light, which would covalently crosslink to any
nearby protein. Therefore, the binding interface
of GPCR and its G protein can be systematically
mapped.
(A) Purified luciferase (wt and Y70Acp mutant) on
Ni-NTA (left panel) or 1D4 resins (right panel)
was reacted with 1mM biocytin-hydrazide. The
eluents were analyzed by western blotting.
(A)
Table 1.
Using luciferase and rhodopsin as model systems,
we have demonstrated that hydrazone and
Staudinger-Bertozzi ligations allow covalent
attachment of labels under physiological
conditions. Both proteins are functional after
12hr reactions. We have also noticed that
wild-type luciferase or rhodopsin in the absence
of a keto moiety reacted with hydrazide and
caused nonspecific attachment of labels. The
chemical nature of the observed nonspecific
reaction is due to naturally presence of
keto/aldehyde groups in proteins (4). Various
proteins can carry different levels of
keto/aldehyde groups (table 1). The heavier the
protein, the higher the chance it gets one keto
or aldehyde group present. This nonspecific
labeling reaction has been overlooked because
model proteins being studied for the hydrazone
ligation so far are mainly small molecular weight
proteins (5-20kD). Together with
cysteine-specific labeling methods, the hydrazone
and Staudinger-Bertozzi ligations will make it
possible to introduce pairs of fluorophores into
GPCRs in a general way.  This is a prerequisite
for single molecule fluorescent resonance energy
transfer (smFRET) studies, which will yield
receptor dynamic information not readily
available by other experimental methods.
Comparisons of these ligations are discussed in
table 2.
(B)
(B) Purified luciferase (wt and Y70Acp mutant) on
Ni-NTA resins was reacted with biocytin-hydrazide
at various concentration (left panel) for 24hr.
Purified luciferae (wt and Y70Azp mutant) on
Ni-NTA resins was reacted with biotin-phosphine
at 0.1mM for 24hr (right panel). The eluents were
analyzed by western blotting.
Table 2.
Amber stop codon suppression in GPCRs
a. General scheme
(C)
(C) Purified luciferase (wt and Y70Acp mutant) on
1D4 resins was reacted with biocytin-hydrazide or
biotin-PEO4-HNAA at various concentrations. The
eluents were analyzed by western blotting.
Photo-crosslinking with Bzp
Mapping the binding interface between two
proteins by the site-directed mutagenesis with
Bzp. A. The amber stop codon suppression method
introduces Bzp at any location in the putative
binding interface. B. Under UV-light, Bzp moiety
generates free radical species that covalently
attaches to a nearby protein. C. The crosslinked
product can be identified by conventional
immunoblotting. D. Systematically introducing
Bzp at the binding interface will help to
identify the key amino acids involved in binding.
(D)
b. Three tyrosine analogues
(A) In vitro labeling of functional rhodopsin
mutants containing Acp at positions 29, 102, or
274. UV-vis spectra before (solid lines) and
after photobleaching (dashed lines) of rhodopsins
treated with fluorescein hydrazide. (B)
Difference spectra (dark minus light spectra) of
labeled rhodopsin mutants generated from (A).
(C) Stoichiometric ratios of fluorescein/rhodopsi
n were determined after normalizing the
amount of rhodopsin based on the absorbance at
500 nm. The molar extinction coefficients used
for calculations were 42,000 M-1cm-1 for
rhodopsin and 93,200 M-1cm-1 for fluorescein.
F/R (fluorescein/rhodopsin molar ratio). (D)
Fluorescein detection by a Typhoon 9400 Image
Scanner using an excitation/emission filter set
optimized for fluorescein.
Establishing the crosslinking procedure in a
model GPCR - rhodospin.
(B)
(A)
anti-1D4 detection rhodopsin
biotinylatedGt? peptide
Membrane prep
wash
SDS-PAGE
Western blotting
Rho-Bzp mutants
streptavidin detection rhodopsin crosslinked to
the Gt? peptide
UV light Crosslinking
Chemical Labeling with Acp and Azp

1. Crystal structure of opsin/Gt? peptide complex
   (adapted from ref 5).
2. Crosslinking of Rho-Bzp mutants to the Gt?
   peptide. Rhodopsin mutants containing Bzp at
   T229 and T243 positions form covalent bonds with
   Gt? peptide after photo-crosslinking with UV
   light. The biotinylated Gt? peptide is detected
   by HRP-streptatvidin (top panel), and rhodopsin
   is detected by anti-1D4 antibody (bottom panel).

a. Reaction schemes
b. Biotin reagents
(A) Biotin reagents for hydrazone ligation
CONCLUSIONS (photo-crosslinking)

1. We use rhodopsin as a model system to establish
   the crosslinking conditions with Bzp.
   Site-specific incorporation of a benzophenone
   moiety into rhodopsin at specific locations (e.g.
   T229 and T243), followed by irradiation with UV
   light generates a reactive species that
   crosslinks it to the Gt? peptide.
2. This method will provide an alternative strategy
   to map the binding interface of a GPCR and its G
   protein. Evidence has accumulated that
   dimerization plays an important role in the GPCR
   signaling process. Homodimerization and
   heterodimerization of GPCRs can modify the
   physiological response through interaction or
   activation of its neighbor.

Human IgG (Sigma) as a model protein testing the
presence of naturally occurring hydrazide
reactive groups. IgG at 1mg/ml in 100mM
phosphate, pH6.0, 150mM NaCl, react with 1mM
fluorescien hydrazide (Invitrogen C356) in the
absence and in the presence of additives for 24h
incubation, RT. Reaction mixture (200ul) is
purified with Sephadex G50 (5ml bed, 6cm length).
Fractions are collected at 200ul for 20 times.
Fractions containing purified IgG are identified
by photometric plate reader (SpectraMax 250) and
pooled for further analysis. Quantify F/P ratio
is done by the 2nd derivative-analysis of
measured spectra (494nm fluorescien)/(301nm
protein).
(B) Biotin reagents for Staudinger-Bertozzi
ligation
REFERENCES

1. Ye, S.X. et al. J. Biol. Chem. 283 1525-1533
   (2008).
2. Cornish, V.W., Hahn, K.M., Schultz, P.G. J. Am.
   Chem. Soc. 118 8150-8151 (1996).
3. Agard, N.J., Baskin, J.M., Prescher, J.A., Lo.
   A., Bertozzi, C.R. ACS Chem. Biol. 1 644-648
   (2006).
4. Ahn, B., Rhee, S.G., Studtman, E.R. Analytical
   Biochem. 161 245-7 (1987).
5. Scheerer, P. et al. Nature 455 497-503 (2008).

c. Luciferase as a model protein to study the
labeling chemistry