A three-component Mannich-type reaction for selective tyrosine bioconjugation

Article abstract of DOI:10.1021/ja0439017

A new selective bioconjugation reaction is described for the modification of tyrosine residues on protein substrates. The reaction uses imines formed in situ from aldehydes and electron-rich anilines to modify phenolic side chains through a Mannich-type electrophilic aromatic substitution pathway. The reaction takes place under mild pH and temperature conditions and can modify protein substrates at concentrations as low as 20 μM. Using an efficient fluorescence-based assay, we demonstrated the reaction using a number of aldehydes and protein targets. Importantly, proteins lacking surface-accessible tyrosines remained unmodified. It was also demonstrated that enzymatic activity is preserved under the mild reaction conditions. This strategy represents one of the first carbon-carbon bond-forming reactions for protein modification and provides an important complement to more commonly used lysine- and cysteine-based methods. Copyright

Full text of DOI:10.1021/ja0439017

Published on Web 11/19/2004
A Three-Component Mannich-Type Reaction for Selective Tyrosine
Bioconjugation
Neel S. Joshi, Leanna R. Whitaker, and Matthew B. Francis*
Department of Chemistry, UniVersity of California, Berkeley, California 94720-1460, and
Materials Sciences DiVision, Lawrence Berkeley National Laboratory, Berkeley, California 94720
Received October 6, 2004; E-mail: francis@cchem.berkeley.edu
The broad utility of bioconjugates for the study of biological
systems and the construction of new materials1 creates an ever-
increasing need for new protein modification reactions. As a result,
many strategies have appeared that can modify native protein
functional groups (most commonly lysine and cysteine side chains)
in a chemoselective fashion.2 However, the high frequency of
lysines on protein surfaces severely limits the site selectivity of
reactions targeting this residue, typically leaving cysteine modifica-
tion as the only effective way to functionalize specific locations.3
To complement this technique, we have focused on the develop-
ment of new protein modification reactions that target aromatic
amino acids,4 as these residues are displayed with intermediate
frequency on protein surfaces and can be introduced genetically
without changing the overall charge state or redox sensitivity. Of
these, tyrosine residues provide particularly attractive targets, as
the reactivity of the phenolic side chains is largely orthogonal to
that of cysteine, lysine, and carboxylate-containing residues.
Furthermore, these residues are often partially “buried” in the
surface of the proteins because of the amphiphilic nature of the
phenolic group. This close association with the topography of
protein surfaces results in varying levels of accessibility for the
reactive sites, suggesting that one of several tyrosine residues could
be targeted selectively through proper reagent design.
Existing methods for tyrosine modification typically occur
through electrophilic aromatic substitution pathways, as exemplified
by iodination5 and diazonium coupling reactions.4a,6,7 In terms of
carbon-carbon bond-forming strategies, formaldehyde-based cross-
linking techniques have been reported to proceed through the
reaction of tyrosines with imines formed with lysine residues,8 albeit
with no control of site selectivity. These reactions are typically
carried out using high concentrations of formaldehyde or with
significant heating and thus are of limited utility for applications
that require preservation of secondary and tertiary protein structure.
As a more synthetically useful strategy, we have developed a new
method that can modify proteins using imines formed in situ from
aldehydes and electron-rich anilines (eq 1).9 This three-component
Mannich-type coupling reaction is highly selective for tyrosine side
chains and proceeds under mild conditions.
Figure 1. Modification of tyrosine residues with imine substrates. (a) A
20 µM solution of R-chymotrypsinogen A was exposed to 1a and 2 (both
at 25 mM) in aqueous buffer, pH 6.5, at room temperature for 18 h. (b)
Following removal of the small molecules via gel filtration, ESI-MS analysis
indicated the formation of the anticipated Mannich adduct (expected m/z
) 26410). (c) In the absence of the aniline component, no reaction products
were observed. (d) A gel-based assay was used to screen for reactivity using
a variety of aldehyde and protein substrates. These reactions were carried
out as described above, but at 37 °C. The top pane indicates the total protein
content by Coomassie blue staining, and the bottom pane indicates the
fluorescence of attached chromophore 2.
protein was observed (Figure 1c). Optimization experiments
determined that maximum reactivity can be obtained in a pH range
of 5.5-6.5, with substantially lower product formation occurring
at pH values above 8. Following removal of unreacted chromophore
2 using dialysis, UV analysis indicated that 60% conversion was
achieved.
To determine the reaction chemoselectivity, the products of this
and analogous reactions employing 4-chloroaniline (3e) were
subjected to trypsin digestion. Through analysis of the resulting
peptide fragments by MALDI-MS, Y146 was determined to be the
primary site of modification. As Friedel-Crafts reactions have
rarely been demonstrated in aqueous solution,11 we confirmed the
product structure by carrying out an analogous reaction using
p-cresol as a small molecule tyrosine mimic.13
As an example of this reactivity, a 20 µM solution of chymo-
trypsinogen A was exposed to 25 mM formaldehyde (1a) and 25
mM 210 for 18 h at room temperature (Figure 1a). Subsequent
analysis using ESI-MS indicated the formation of a new species
into which both reactive components were incorporated with equal
stoichiometry (Figure 1b). In an analogous experiment conducted
in the absence of the aniline component, no modification of the
The fluorescent nature of 2 allowed the use of an efficient gel-
based assay to screen additional aldehyde components and protein
targets (Figure 1d). In addition to formaldehyde, pyruvaldehyde
9
15942
J. AM. CHEM. SOC. 2004, 126, 15942-15943
10.1021/ja0439017 CCC: $27.50 © 2004 American Chemical Society

C O M M U N I C A T I O N S
Scheme 1
One feature of this reaction that stands in contrast to other
tyrosine modification strategies is the relatively low pH at which
it takes place. We hypothesize that at pH 6.5, a cyclic transition
state occurs in which the phenolic group donates a hydrogen bond
to the imine nitrogen (Scheme 1). This would serve to increase the
electrophilicity of the imine group while simultaneously activating
the aromatic ring as a nucleophile. It is also possible that
hydrophobic attractions between the imine substrate and the tyrosine
ring accelerate the reaction by increasing the effective concentration.
However, it should be noted that imine formation in water involves
numerous equilibrating species, and thus the reaction could proceed
through a more complex pathway.
Figure 2. Modification sites and activity assay for chymotrypsinogen A.
(a) Of the accessible tyrosines (green), Y146 was identified as the primary
modification site. (b) Reaction profile for proteolysis of Suc-Gly-Gly-Phe-
NHC6H4NO2 for unmodified chymotrypsinogen (red) and the bioconjugate
prepared from 1a and 3a (blue) after both enzymes were activated with
trypsin.
Table 1. Modification of Tyrosine Residues on
Chymotrypsinogena
aniline
X-substituent
unmod(%)
+
1modb
+
2modb
+
3modb
2
-rhodamine dye
-CH2CH2OH
-CH2CH2NH2
-CH3
34
20
27
38
57
53
57
91
100
66
45
47
46
43
36
33
9
0
35
26
16
0
11
10
0
0
0
0
0
0
0
0
0
0
As a result of these studies, an efficient new tyrosine modification
method has emerged. The reaction is operationally simple to effect,
and the mild conditions under which it proceeds ensure compat-
ibility with a wide range of protein targets. The three-component
nature of the reaction allows for the installation of multiple
functional groups, and the selectivity of the reaction for tyrosine
residues provides a complementary strategy to cysteine- and lysine-
based methods. Current efforts are focused on the full exploration
of the substrate scope as well as the elucidation of the mechanistic
aspects. Experiments examining the ability of the reaction com-
ponents to select between multiple tyrosines on a single protein
surface are also underway.
3a
3b
3c
3d
3e
3f
-OCH3
-Cl
-CO2H
3g
3h
-F
-NO2
0
0
a Conditions: 200 µM R-chymotrypsinogen A, 25 mM formaldehyde,
and 25 mM aniline in 100 mM phosphate buffer, pH 6.5, room temperature,
18 h. The X-substituents refer to the 4-position of the anilines, as indicated
in eq 1. Product distributions were determined from ESI-MS analyses. b In
percent.
Acknowledgment. We gratefully acknowledge the University
of California, Berkeley, and the DOE Nanoscale Science, Engineer-
ing, and Technology Program for funding. N.S.J. acknowledges
the Saegebarth family for generous fellowship support. We also
thank Waters, Inc. for access to a Q-TOF Micro instrument, and
(1b), (E)-crotonaldehyde (1c), and 2-furaldehyde (1e) also yielded
appreciable levels of reactivity; however, reactions carried out with
glyoxylic acid (1d), benzaldehyde, and propionaldehyde did not
afford modification products. In all cases, no cross-linking of the
proteins was observed.
In addition to chymotrypsinogen A, other proteins bearing
surface-accessible tyrosine residues have been modified, including
lysozome and RNase A (Figure 1d). In contrast, no reactivity has
been observed for horse heart myoglobin, a protein that lacks
surface-accessible tyrosine residues.
Jacob M. Hooker for his analysis expertise.
Supporting Information Available: Full experimental procedures
and characterization data are available for all compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
References
In terms of reagent scope, successful reactions have been
achieved with a range of electron-rich anilines (Table 1). As
determined for the modification of chymotrypsinogen A, 4-alkyl-
substituted anilines (3a-c) generally afforded high levels of
reactivity and yielded both singly and doubly modified protein
products. Chymotrypsinogen A bears three potentially reactive
tyrosine residues on its surface (Figure 2a),12 which can account
for its overmodification. Efforts to quantify the reactivity levels at
each site are underway. Aniline 3b is of particular interest, as the
aliphatic amine of this substrate can readily be coupled to NHS-
esters, effectively converting the large number of commercially
available lysine-reactive compounds into reagents for tyrosine
modification. Other para-functionalized anilines (3d-f) also par-
ticipated in the reaction, although lower levels of conversion were
obtained. To date, reactions employing electron-deficient anilines
and aliphatic amines have not been successful. This is significant,
as the amino groups of lysine residues also do not participate in
the reaction under these conditions.
(1) For lead references, see: (a) Niemeyer, C. M. Angew. Chem., Int. Ed.
2001, 40, 4128-4158. (b) Seeman, N. C.; Belcher, A. M. Proc. Natl.
Acad. Sci. U.S.A. 2002, 99, 6451-6455.
(2) For example, see: Hermanson, G. T. Bioconjugate Techniques, 1st ed.;
Academic Press: San Diego, CA, 1996.
(3) For an example of site-selective binding to native protein functional groups,
see: Griffin, B. A.; Adams, S. R.; Jones, J.; Tsien, R. Y. Fluorescent
Labeling of Recombinant Proteins in Living Cells with FlAsH. In
Applications of Chimeric Genes and Hybrid Proteins; Thorner, J., Emr,
S. D., Abelson, J. N., Eds.; Methods in Enzymology 327; Academic
Press: San Diego, CA; 2000, pp 565-578.
(4) (a) Hooker, J. M.; Kovacs, E. W.; Francis, M. B. J. Am. Chem. Soc. 2004,
126, 3718. (b) Antos, J. M.; Francis, M. B. J. Am. Chem. Soc. 2004, 126,
10256.
(5) (a) Adams, H. R.; Paik, C. H.; Eckelman, W. C.; Reba, R. C. J. Labelled
Compd. Radiopharm. 1982, 19, 1477-1478. (b) Eckelman, W. C.; Adams,
H. R.; Paik, C. H. Int. J. Nucl. Med. Biol. 1984, 11, 163-166.
(6) (a) Higgins, H. G.; Fraser, D. Aust. J. Sci. Res., Ser. A 1952, 5, 736-753.
(b) Shin, J. A. Bioorg. Med. Chem. Lett. 1997, 7, 2367.
(7) For alternative oxidative strategies, see: (a) Brown, K. C.; Yang, S. H.;
Kodadek, T. Biochemistry 1995, 34, 4733-4739. (b) Meunier, S.; Strable,
E.; Finn, M. G. Chem. Biol. 2004, 11, 319-326.
(8) Fraenkel-Conrat, H.; Olcott, H. S. J. Biol. Chem. 1948, 174, 827-843.
(9) For an example of a lanthanide-promoted phenol modification with imines
in organic solvents, see: Huang, T. S.; Li, C. J. Tetrahedron Lett. 2000,
41, 6715.
(10) Nguyen, T.; Francis, M. B. Org. Lett. 2003, 5, 3245-3248.
(11) For a notable example, see: Breslow, R. Acc. Chem. Res. 2004, 37, 471.
(12) Pjura, P. E.; Lenhoff, A. M.; Leonard, S. A.; Gittis, A. G. J. Mol. Biol.
2000, 300, 235-239.
To demonstrate that this bioconjugation method preserves the
native function of biological targets, the catalytic activity of
chymotrypsinogen A was evaluated after modification with 1a and
3a. It was found that little if any reactivity was lost even though
80% of the protein had been functionalized (Figure 2b).13
(13) See Supporting Information for details.
JA0439017
9
J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004 15943