Download probing protein structure and interactions using functionalized

PROBING PROTEIN STRUCTURE AND INTERACTIONS USING FUNCTIONALIZED
NAPHTHALIMIDES
L. Kelly, B. Abraham, M. Mullan
Department of Chemistry and Biochemistry, University of Maryland, Baltimore County
iNTRODUCTION.
Laser Flash Photolysis Studies. Transient spectroscopy was used to identify
the specific amino acids that are oxidized by the naphthalimide excited states.
The transient spectrum shown in Figure 2 illustrates that the oxidation of both
tyrosine and tryptophan is mediated by the 1,8-naphthalimide triplet excited
states (eq 1). When the experiment is conducted in an air-saturated aqueous
solution, the spectra of the amino acid radicals are cleanly observed.
Biomolecular interactions, particularly those involving nucleic acids
and proteins with each other or with small molecules, are ubiquitous and
critical to maintaining structure and function. Work in our laboratory has
focused on using water soluble 1,8-naphthalimides and 1,4,5,8-naphthalene
diimides as protein structural probes. The compounds initiate a number of
photochemical reactions. It is of interest to develop small molecules that will
cleave, crosslink, or photoaffinity label proteins and DNA. In this study, the
fundamental reactions with amino acids and proteins are studied to
understand the mechanisms of protein modification.
The rate constants for triplet-mediated electron transfer were measured
using individual amino acids. Tyrosine and tryptophan were found to be the
only amino acids that could be oxidized by the triplet state of compound (II).
The native proteins, bovine serum albumin (BSA) and lysozyme also quenched
the triplet state. Radicals were observed as lysozyme quenching products.
The rate constants for the reaction shown in eq 1 were determined from the
bimolecular quenching plots shown in Figure 3. The rate constants and radical
(cage escape) yields are summarized in Table 1 (ref. 2).
a
5
15 s
(a)
0.04
0.03
Naphthalene imides and diimides are synthesized from commercially available
aromatic anhydrides and primary amines. To date, the anionic, cationic, and
polyamine derivatives shown in Figure 1 have been synthesized.
0.5 s
0.02
0.01
0.00
3
-0.01
-0.02
300
R
350
400
450
500
550
II* + AA
600
N
O
II.- + AA.+ (eq 1)
O
Delta A
R
0.006
(III): R = -CH2CH2 N+
(b)
15 s
kq
(x10-9 M-1s-1)
0.01
Tryptophan 2.60 ± 0.16
0.52 ± 0.04
Tyrosine
0.89 ± 0.03
0.33 ± 0.05
BSA
0.42 ± 0.02
Not
observed
Lysozyme
0.76 ± 0.01
0.22 ± 0.01
-0.01
0
40
80
time(s)
5
2x10
0
0
100
200
300
400
[AA] M
(d)
FIGURE 3. Bimolecular quenching plots for reaction
of the triplet excited states of compound (II) with
tryptophan (red) and tyrosine (blue). Inset: Tripletstate decay with added (0 – 50 M) tryptophan.
0.004
Radical
Yield
0.02
II
0.008
FIGURE 1. Synthetic Structural
Probes
0.03
0.00
5
4x10
+
(II): R = -CH2CH2COO-
N
kq
O2
 (nm)
(I): R = -CH2CH2 N
O
(c)
kobs (s )
Delta A
EXPERIMENTAL RESULTS.
A (480 nm)
6x10
-1
0.05
O
Amino Acid
0.04
TABLE 1. Summary of bimolecular rate constants
for and radical yields from the reaction of 3II* with
amino acids and proteins.
0.002
(IV): R = -CH2CH2COO0.000
300
O
350
400
500
550
600
 (nm)
O
N
450
R
FIGURE 2. Transient absorption spectrum of (II) (10 M) measured in an
air-saturated aqueous solution of (a) 2.25 mM tyrosine and (b) tryptophan.
The spectrum observed at long times (15 s) shows similar spectral
features as the known spectrum of the tyrosyl and tryptophan radicals ((c)
and (d), respectively (From ref. 1)).
Photochemistry of Naphthalene Diimides. The redox properties of the
naphthalene imides and diimides are known. 1,4,5,8-naphthalene diimides ((III)
and (IV)) have one-electron reduction potentials that are ca. 0.5 V more positive
than their 1,8-naphthalimide counterparts (I and II). Thus, it was of interest to
compare their photoreactivity with amino acids and proteins.
The fluorescence emission spectra of compounds III and IV were compared.
These are shown in Figure 4. As seen from the figure, the relative intensity of
compound IV (and other carboxylated naphthalene diimides) is significantly
smaller than that of compound III. The quantum yields are indicated on the figure.
O
0.4
0 min
5 min
10 min
15 min
20 min
0.5
Absorbance
0.8
0.3
A (280 nm)
Intensity (A.U.)
0.4
0.6
0.4
0.3
0.2
0.1
0 min
5 min
10 min
15 min
20 min
0.3
200
0.2
250
300
350
400
Drug
0.2
NH
O
300
350
400
OH
 (nm)
NDI
450
500
10
15
Preliminary evidence (Figure 5) suggests that carboxyalkyl-functionalized 1,4,5,8naphthalene diimide derivatives may be well-suited for photoaffinity labeling reagents.
These compounds have molar extinction coefficients in excess of 20,000 M-1cm-1. The
proposed mechanism for protein photoaffinity labeling is given in Scheme III.
20
Retention Time (minutes)
550
Ph
Ph Drug
450
0.0
400
recombination
 (nm)
5
N
H
Drug
Drug
FIGURE 7. Benzophenone-initiated protein photoaffinity labeling.
0.0
250
(III) (f = 0.0009)
0.0
350
+
0.1
450
O
N
H
0.0
0.1
0.2
OH
h
0.4
Absorbance
1.0
O
O
0.5
BSA
(IV) (f = 0.017)
Photoaffinity Labels. Benzophenone-ligand (drug) conjugates have been used to probe
protein-ligand interaction sites. Production of the n-p* excited state, followed by
hydrogen atom abstraction and radical recombination, gives rise to a covalent adduct
between the protein and benzophenone. This is shown in Figure 7. Protein degradation
and fragment analysis is then used to identify the ligand-protein interaction site.
Unfortunately, benzophenone has a very small molar extinction coefficient above 300 nm
and requires a high concentration to produce a reasonable absorption cross-section.
 (nm)
O
FIGURE 4. Fluorescence emission spectra of aqueous buffered solutions
of compounds III and IV. The solutions were optically matched at the
excitation wavelength of 382 nm. The absolute fluorescence quantum
yields are given.
The data shown in Figure 4, along with other experiments (ref. 3) indicate
that the singlet state of compound III is rapidly quenched via intramolecular
electron transfer (Scheme I). By analogy to published work on phthalimides, we
propose that, following intramolecular electron transfer, homolytic bond cleavage
produces CO2 and the carbon-centered radical. Preliminary evidence (Figure 5)
suggests that this carbon-centered radical may covalently attach to a protein.
FIGURE 5. HPLC chromatograph showing separation of (III) from the protein BSA. Upon
irradiation ( > 320 nm; 0 – 20 min as indicated) of (III) in the presence of BSA, the peak at ca. 15
minutes shows the UV absorption spectrum of (III). Irradiation of (III), followed by mixing with a
solution of BSA, did not indicate attachment of the naphthalene diimide to the protein.
CO2O
N
O
*
Prospects as Protein Structural Probes. Compounds (I) – (IV) have shown diverse
photochemical reactivity with proteins. As shown in Figures 2 and 3, the
naphthalene imide (II) predominantly reacts with proteins and amino acids via oneelectron amino acid oxidation in the absence of oxygen. Gel electrophoresis
confirms that oxidative cross-linking is the major protein modification product
(Figure 6). The reaction is believed to occur following oxidation of either tyrosine
or tryptophan, deprotonation to product the neutral radical, and radical-radical
recombination.
CO2
-
O
N
-
O
O
N
O
N
R
O
O
N
R
O
O
N
O
*
N
H
CO2
-O
N
O
-O
N
O
O
-O
N
O
radical
rec.
SCHEME III.
C-H-abs
O
N
R
O
O
N
R
O
O
N
R
O
O
N
R
O
CONCLUSION. We have shown that naphthalene imides and diimides can be lightactivated to initiate a variety of protein modification event. Photoinduced
cleavage, cross-linking, and affinity-labeling are useful tools to probe the structure
and dynamics of proteins. The development of new photoaffinity labeling agents
with large extinction coefficients offers the possibility of using these reagents at
M concentrations.
14 kD
O
6.5 kD
-CO2
CO2-
N
H
27 kD
1
1
ACKNOWLEDGMENTS AND REFERENCES. This work was supported by NSF
Grant CHE-9984874
Radical reaction products?
O
N
R
O
FIGURE 6. Separation of protein photoproducts using gel electrophoresis (20% SDS/PAGE, post-stained
with coomassie blue). Solutions contain 80 M of lysozyme and 50 M of (II). In all cases, lanes 1 and 2
contain only lysozyme (no (II)); samples run in lanes 1 and 3 have been kept in the dark. Lane 2 has been
exposed to light for the entire duration of photolysis. Lanes 4, 5, 6, 7 represent lysozyme samples that
have been irradiated for 0.5, 1, 2, and 3 hours respectively, under nitrogen-saturated conditions.
(1)Wagenknecht, H. A.; Stempt, E. D.A.; Barton, J. K. Biochemistry 2000, 39, 54835491.
(2) Abraham, B.; Kelly, L. A. J. Phys. Chem. B 2003, 107, 12534 – 12541.
(3) Abraham, B.; McMasters, S.; Mullan, M. A.; Kelly, L. A. J. Am. Chem. Soc. 2004,
126, 4293-4300.