Efficient light-dependent DNA repair requires a large cofactor separation

Article abstract of DOI:10.1021/ja990466l

DNA photolyases are repair enzymes which split (repair) UV-induced cyclobutane DNA lesions. Critical steps in the light-driven repair reaction are the absorption of light by a deazaflavin or methenyl tetrahydrofolate cofactor and the transfer of the excit

Full text of DOI:10.1021/ja990466l

7318
J. Am. Chem. Soc. 1999, 121, 7318-7329
Efficient Light-Dependent DNA Repair Requires a Large Cofactor
Separation
Robert Epple and Thomas Carell*
Contribution from the Laboratorium fu¨r Organische Chemie, Eidgeno¨ssische Technische Hochschule,
UniVersita¨tstrasse 16, CH-8092 Zu¨rich, Switzerland
ReceiVed February 15, 1999
Abstract: DNA photolyases are repair enzymes which split (repair) UV-induced cyclobutane DNA lesions.
Critical steps in the light-driven repair reaction are the absorption of light by a deazaflavin or methenyl
tetrahydrofolate cofactor and the transfer of the excitation energy to a reduced and deprotonated FADH-
cofactor, which initiates an electron transfer to the dimer lesion. Although most efficient energy transfer requires
a close cofactor arrangement, there is a separation of >17 Å between the cofactors in photolyases. To determine
the effect of the large cofactor distance on the repair efficiency, a systematic study with model compounds
was performed. A series of compounds were synthesized which contain a model DNA lesion covalently
connected to a flavin and a deazaflavin. While the flavin-dimer lesion distance was kept constant in all model
compounds, the flavin-deazaflavin distance was incrementally increased. Investigation of the dimer cleavage
efficiency shows that compounds with a large cofactor separation possess a low energy-transfer efficiency but
split the dimer most efficiently within a few minutes. Model compounds with a close cofactor orientation
feature a highly efficient energy transfer from the deazaflavin to the flavin. They are, however, unable to
perform the repair of the dimer lesion. At very short cofactor distances, the light-driven repair process is fully
inhibited. This is explained by a competitive electron transfer between both cofactors, which hinders the electron
transfer to the dimer lesion and hence the dimer splitting. The presented data suggest that the large cofactor
separation (17 Å) found in photolyases is a critical parameter that determines the DNA repair efficiency by
photolyases.
Introduction
defense system against UV-induced DNA damage in many
organisms.11,12
Photoinduced electron-transfer reactions play a key role in
biology.1,2 They are the basis for the photosynthetic process in
which chlorin, pheophytine, and quinone chromophores orches-
trate the conversion of sunlight into chemical energy.3,4 Pho-
tolyase DNA repair enzymes utilize a reduced riboflavin
(FADH-), a ribodeazaflavin (F0), or a methenyl tetrahydrofolate
(MTHF) for the catalytic, light-dependent repair of the most
abundant UV-induced genome lesions.5-7 A similar set of
cofactors was recently discovered in cryptochrome photorecep-
tors,8 which control key developmental events in plants, such
as hypocotyl elongation,9 and also set the circadian clock.10
DNA photolyases are repair enzymes that represent a major
UV irradiation of cells induces a [2π + 2π] cycloaddition of
pyrimidines located above each other in the DNA double strand.
The resulting cis-syn pyrimidine dimers are responsible for cell
death and the degeneration of cells into tumor cells.13,14 DNA
photolyases specifically recognize the cis-syn dimer lesions
and split the dimers in a light-dependent reaction.15 The
key mechanistic steps of the repair reaction are depicted in
Figure 1.
The repair mechanism of F0-containing photolyases (type II)
includes the absorption of light by the F0 cofactor and an energy
transfer from the F0 to a reduced and deprotonated FADH-.16-18
The flavin donates an electron to the cyclobutane pyrimidine
dimer lesion, which cleaves spontaneously as its radical anion.6
Two recent X-ray crystal structures of photolyases reveal a large
distance between the light-harvesting and the redox-active
cofactor.19,20 Despite their need to interact efficiently, the
cofactors were found to be separated by more than 17 Å, which
* To whom correspondence should be addressed. Tel.: 0041-1-632-2944.
Fax: 0041-1-632-1109. E-mail: tcarell@org.chem.ethz.ch.
(1) Gray, H. B.; Winkler, J. R. Annu. ReV. Biochem. 1996, 65, 537-
561.
(2) Langen, R.; Chang, I.-J.; Germanas, J. P.; Richards, J. H.; Winkler,
J. R.; Gray, H. B. Science 1995, 268, 1733-1735.
(3) Clayton, R. K. Photosynthesis: Physical Mechanisms and Chemical
Patterns; Cambridge University Press: London, 1980.
(4) Wasielewski, M. R. Chem. ReV. 1992, 92, 435-461. Kurreck, H.;
Huber, M. Angew. Chem., Int. Ed. Engl. 1995, 34, 849-866.
(5) Sancar, A. Biochemistry 1994, 33, 2-9.
(11) Friedberg, E. C.; Walker, G. C.; Siede, W. DNA repair and
mutagenesis; ASM Press: Washington, DC, 1995.
(12) Carell, T. Chimia 1995, 49, 365-373.
(13) Taylor, J.-S. J. Chem. Educ. 1990, 67, 835-841.
(14) Taylor, J.-S. Acc. Chem. Res. 1994, 27, 76-82.
(15) Carell, T. Angew. Chem., Int. Ed. Engl. 1995, 34, 2491-2494.
(16) Kim, S.-T.; Heelis, P. F.; Okamura, T.; Hirata, Y.; Mataga, N.;
Sancar, A. Biochemistry 1991, 30, 11262-11270.
(17) Kim, S.-T.; Heelis, P. F.; Sancar, A. Biochemistry 1992, 31, 11244-
11248.
(6) Heelis, P. F.; Hartman, R. F.; Rose, S. D. Chem. Soc. ReV. 1995,
289-297.
(7) Begley, T. P. Acc. Chem. Res. 1994, 27, 394-401. Carell, T.; Epple,
R. Eur. J. Org. Chem. 1998, 1245, 5-1258.
(8) Ahmad, M.; Cashmore, A. R. Nature 1993, 366, 162-166.
(9) Malhotra, K.; Kim, S.-T.; Batschauer, A.; Dawut, L.; Sancar, A.
Biochemistry 1995, 34, 6892-6899.
(18) Jorns, M. S.; Ramsey, A. J. Biochemistry 1992, 31, 8437-8441.
(19) Park, H.-W.; Kim, S.-T.; Sancar, A.; Deisenhofer, J. Science 1995,
268, 1866-1872.
(10) Miyamoto, Y.; Sancar, A. Proc. Natl. Acad. Sci. U.S.A. 1998, 95,
6097-6102.
10.1021/ja990466l CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/03/1999

Efficient Light-Dependent DNA Repair
J. Am. Chem. Soc., Vol. 121, No. 32, 1999 7319
active form in type II DNA photolyases. Comparison of the
cleavage results for all model compounds allowed the deter-
mination of two main parameters. It was found that the flavin-
deazaflavin distance strongly influences the repair rate in a
counterintuitive way, and we could determine that the depro-
tonation of the deazaflavin helps to increase the repair efficiency.
Results and Discussion
Synthesis of the Model Compounds. All model compounds
1a/b-4a/b and 5 were chemically assembled, as depicted in
Scheme 1, from the cyclobutane uracil dimer building block 6
and the N(3)-alkyl-N(10)-aminoethylflavins 7 and 8, the flavin
amino acid 9, the N(3)-pentyl-N(10)-aminoethyldeazaflavin 10,
and the deazaflavin amino acid 11 (Schemes 2 and 3). The
synthesis of the uracil dimer 6, of the flavin 8, and of the
reference model compound 5 was achieved as recently described
in detail.26,30 All other cofactor derivatives were prepared as
depicted in Schemes 2 and 3. The flavin and the deazaflavin
peptides 12-14, required for the preparation of the model
compounds 1-3, were prepared by liquid- and solid-phase
peptide synthesis (Scheme 4).31
For the preparation of the flavin amino acid 9 and the pentyl-
substituted flavin 7 (Scheme 2), the tert-butyloxycarbonyl (Boc)-
protected aminoethyl-substituted flavin 1532 was alkylated with
tert-butyl bromoacetate to give the Boc-protected flavin amino
acid tert-butyl ester 16. Acidic cleavage of the tert-butyl-
containing protection groups yielded the flavin amino acid 17,
which was converted into the fluorenylmethoxycarbonyl (Fmoc)-
protected amino acid 9, ready for the incorporation into peptides
using solid-phase synthesis, by treatment with N-(9-fluorenyl-
methoxycarbonyloxy)succinimide (Fmoc-OSu) and K2CO3.
Alkylation of 15 with pentyl bromide in dimethyl-
formamide (DMF) with Cs2CO3 as the base afforded the pentyl-
substituted flavin derivative 18, which was deprotected with
trifluoroacetic acid (TFA) to give 7.
The deazaflavins 10 and 11 (Scheme 3) were prepared by
substitution of 6-chlorouracil 19 with the mono-Boc-protected
ethylenediamine 20.33 The product 21 was condensed to the
deazaflavin 22 with 2,4-(dibenzyloxy)benzaldehyde (23).34
Alkylation of the deazaflavin 22 with pentyl bromide in the
presence of Cs2CO3 in DMF yielded the pentyl-substituted
deazaflavin 24. Cleavage of the Boc group with TFA afforded
10. Alkylation of 22 with tert-butyl bromoacetate to 25 and
acidic cleavage of the tert-butyl-containing protecting groups
yielded the deazaflavin amino acid 26. Reaction of 26 with
Fmoc-OSu and K2CO3 furnished the deazaflavin amino acid
11, ready for incorporation into peptides using solid-phase
synthesis.
For the synthesis of the flavin-deazaflavin peptide 12
(Scheme 4), the carboxylic acid 9 was activated with benzo-
triazol-1-yloxytris(dimethylamino)phosphoniumhexafluorophos-
phate (BOP) and reacted with the N(3)-pentyl-subtituted dea-
zaflavin 10 in DMF to yield 27. Cleavage of the Fmoc group
in 27 was performed with diethylamine directly before the
coupling with the dimer diacid 6, without purification and
characterization of the intermediate amine 12.
Figure 1. Schematic representation of the mechanism of the type II
photolyase-initiated DNA repair process. The deazaflavin (F0) absorbs
light and transfers the energy to a reduced and deprotonated flavin
(FADH-). This initiates an electron transfer to a cyclobutane pyrimidine
dimer lesion (T)T). The dimer cleaves spontaneously as its radical
anion. EET ) Excitation energy transfer.
led to the hypothesis that the cofactor-cofactor interaction has
never been optimized during evolution.19,21 Moreover, two
invariant basic amino acid residues at the bottom of the F0
binding pocket in all type II photolyases ensure the de-
protonation of the F0 cofactor (O- state) for an unknown
reason.20
We recently communicated the synthesis of chimaeric flavin-
and deazaflavin-containing pyrimidine dimer model com-
pounds,22-24 which are able to mimic the F0 f FADH- energy
transfer and the FADH- f dimer electron-transfer process of
the photolyase repair reaction.25 In this publication, we report
the investigation of the potential reasons for the structural
peculiarities found in photolyases. With the set of novel bis-
cofactor model compounds presented in Scheme 1, the effects
of the cofactor-cofactor distance and the deazaflavin depro-
tonation on the energy-transfer efficiency and repair rate were
studied. All model systems contain a flavin (Fl) and a deazafla-
vin (dFl) covalently attached to a cis-syn-cyclobutane uracil
dimer.26 The flavin-dimer distance was kept constant in all
model compounds, which ensures a constant electron-transfer
efficiency from the flavin to the dimer, while the flavin-
deazaflavin distance was systematically changed. The distance
is shortest in the model compounds 1a and 1b and largest in
the model compounds 2a and 2b, which contain four “semirigid”
proline spacers.27,28 In all model compounds of series a (1a-
4a), the deazaflavin is fixed in the OH form due to benzylation
of the 8-OH group. Cleavage of this protection group by
catalytic hydrogenation and adjustment of the pH to >8
converted the deazaflavin (pKa ≈ 6)29 into the deprotonated form
present in 1b-4b (series b). This “quinoid” deazaflavin is the
(20) Tamada, T.; Kitadokoro, K.; Higuchi, Y.; Inaka, K.; Yasui, A.; de
Ruiter, P. E.; Eker, A. P. M.; Miki, K. Nature Struct. Biol. 1997, 11, 887-
891.
(21) Heelis, P. F. J. Photochem. Photobiol., B 1997, 38, 31-34.
(22) Eker, A. P. M.; Dekker, R. H.; Berends, W. Photochem. Photobiol.
1981, 33, 65.
(23) Rokita, S. E.; Walsh, C. T. J. Am. Chem. Soc. 1984, 106, 4589-
4595.
(24) Jorns, M. S. J. Am. Chem. Soc. 1987, 109, 3133-3136.
(25) Epple, R.; Carell, T. Angew. Chem., Int. Ed. Engl. 1998, 37, 938-
941.
(30) Epple, R.; Wallenborn, E.-U.; Carell, T. J. Am. Chem. Soc. 1997,
119, 7440-7451.
(31) Carell, T.; Schmid, H.; Reinhard, M. J. Org. Chem. 1998, 63, 8741-
8747.
(32) Butenandt, J.; Eker, A. P. M.; Epple, R.; Gramlich, V.; Wallenborn,
E.-U.; Carell, T. Chem. Eur. J., accepted for publication.
(33) Pfleiderer, W.; Nu¨bel, G. Liebigs. Ann. Chem. 1959, 631, 168-
174. Petry, C., Dissertation, Heidelberg, 1993.
(34) Kimachi, T.; Tanaka, K.; Yoneda, F. J. Heterocycl. Chem. 1991,
28, 439-443.
(26) Carell, T.; Epple, R.; Gramlich, V. HelV. Chim. Acta 1997, 80,
2191-2203.
(27) McCafferty, D. G.; Friesen, D. A.; Danielson, E.; Wall, C. G.;
Saderholm, M. J.; Erickson, B. W.; Meyer, T. J. Proc. Natl. Acad. Sci.
U.S.A. 1996, 93, 8200-8204.
(28) Slate, C. A.; Striplin, D. R.; Moss, J. A.; Chen, P.; Erickson, B.
W.; Meyer, T. J. J. Am. Chem. Soc. 1998, 120, 4885-4886.
(29) Walsh, C. Acc. Chem. Res. 1986, 19, 216-221.

7320 J. Am. Chem. Soc., Vol. 121, No. 32, 1999
Epple and Carell
Scheme 1. Synthesis of the Model Compounds 5 and 1a/b to 4a/ba
a Conditions: (a) BOP, NEt3, DMF, room temperature; (b) H2, Pd/BaSO4, acetic acid, room temperature.
Scheme 2. Synthesis of the Flavin Building Blocks 7 and 9
Required for the Preparation of the Model Compounds:
Depiction of the Flavin Compound 8a
Scheme 3. Synthesis of the Deazaflavin Building Blocks 10
and 11a
a Conditions: (a) Br-C5H11, DMF, Cs2CO3; (b) TFA; (c) Br-CH2-
COO-t-Bu, Cs2CO3, DMF; (d) Fmoc-OSu, K2CO3, 4 °C.
a Conditions: (a) n-butanol, reflux; (b) DMF, 120 °C, 20 h; (c) Br-
C5H11, DMF, Cs2CO3; (d) TFA; (e) Br-CH2COO-t-Bu, Cs2CO3, DMF;
(f) Fmoc-OSu, K2CO3, 4 °C.
The synthesis of the deazaflavin peptides 13 and 14 required
coupling of the deazaflavin amino acid 11 to a Rink-Amide
MBHA resin to yield the resin 28.35 This was performed in
1-methyl-2-pyrrolidone (NMP) as the solvent, with diisopro-
pylethylamine (DIEA) as the base, after in situ activation of
the carboxylic acid group of 11 with N-hydroxybenzotriazole/
2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium tetrafluo-
roborate (HOBT/TBTU).36 Subsequent cleavage of the Fmoc
protection group with piperidine in DMF and stepwise coupling
of either two or four Fmoc-protected proline units to the
deazaflavin-loaded resin 28 gave the resins 29 and 30, respec-
tively. The final cleavage of the Fmoc group in 29 and 30 with
piperidine and of the peptides from the resins with TFA afforded
the crude peptides 13 and 14. Both peptides were obtained after
reversed-phase HPLC purification in excellent yields and purity
as yellow powders.
The synthesis of the model compounds (Scheme 1) required
activation of the two carboxylic acid groups of 6 with BOP in
DMF.37 After addition of 1 equiv of the building blocks 8 or
12 and DIEA, the mixture was allowed to react for 1 h.
Subsequent addition of an excess of pentylamine yielded the
model compounds 5 and 1a, respectively. Reaction of the BOP-
activated uracil dimer 6 with 1 equiv of the deazaflavin peptides
13 or 14, followed by addition of the flavin 8, furnished the
model compounds 2a and 3a. Reaction of the BOP-activated
dimer 6 with the pentyl-substituted flavin 7, followed by addition
of the deazaflavin 10, yielded the model compound 4a. All
model compounds were isolated either by flash chromatography
(5, 1a, and 4a) on silica-H or by preparative reversed-phase
HPLC on a C18-column with a water/acetonitrile gradient (2a,
3a). The compounds were obtained as yellow powders. For the
synthesis of the model compound in series b (1b-4b), which
possess a debenzylated “quinoid” deazaflavin, the corresponding
(35) Grant, G. A. Synthetic Peptides: A User’s Guide; W. H. Freeman
and Co.: New York, 1992.
(36) Knorr, R.; Trzeciak, A.; Bannwarth, W.; Gillessen, D. Tetrahedron
Lett. 1989, 30, 1927-1930.
(37) Castro, B.; Evin, G.; Delve, C.; Seyer, R. Synthesis 1977, 413-
469.

Efficient Light-Dependent DNA Repair
J. Am. Chem. Soc., Vol. 121, No. 32, 1999 7321
Scheme 4. Synthesis of the Flavin-Deazaflavin Peptide 12
Needed for the Preparation of the Model Compounds 1a and
1b and of the Deazaflavin-Proline Peptides 13 and 14
Required for the Preparation of the Model Compounds 2a/b
and 3a/ba
a Conditions: (a) 20% piperidine in DMF; (b) 11, HOBT, TBTU,
DIEA; (c) Fmoc-Pro, HOBT, TBTU, DIEA; (d) TFA/H2O/triisopro-
pylsilane (TIS) (95:2.5:2.5), 1 h; (e) BOP, NEt3, DMF; (f) 40% Me2NH
in H2O, DMF, 20 min.
Figure 2. (A) UV spectra of the model compounds 1a-4a (series a)
prior to the reduction with dithionite (solid line) and after the addition
of sodium dithionite (dotted line). (B) UV spectra of the model
compounds 1b-4b (series b) prior to the reduction with dithionite (solid
line) and after the addition of sodium dithionite (dotted line). All UV
data were recorded at 10-5 M in ethylene glycol. (C) Action spectra of
the reference compound 5 (2), of the model compounds in series a
(b, represented by 4a), and of the model compounds of series b (9,
represented by 4b).
model compounds 1a-4a were dissolved in acetic acid. A small
amount of Pd/BaSO4 catalyst was added, and the suspension
was stirred under a hydrogen atmosphere until complete
debenzylation was achieved. During the hydrogenolysis, both
chromophores were reduced as well. Stirring of the reaction
solution in air, however, caused complete reoxidation of the
flavin and the deazaflavin cofactors. Decomposition of the model
compounds was not observed (thin-layer chromatography). The
model compound 4b was prepared in quantitative yield on a
larger scale to allow its isolation and complete characterization.
All other model compounds 1b-3b were characterized by UV
and fluorescence spectroscopy without prior purification. The
purity of the prepared model compounds 1b-3b was determined
by analytical reversed-phase HPLC before the measurements
and found to be >96%.
Sample Preparation. To quantify the efficiency of the dimer
splitting in all model compounds 1a/b-4a/b, selective reduction
of the flavin unit in the presence of a deazaflavin chromophore
either in its OBn- or its “quinoid” O- form had to be performed.
This was accomplished with sodium dithionite. The model
compounds were dissolved (10-5-10-6 M) in ethylene glycol
(3 mL) in a UV cuvette, stoppered with a rubber septum, and
DIEA (10 µL) was added to ensure basic conditions. The assay
solution was subsequently purged with nitrogen. The dithionite
solution was then carefully added with a microsyringe until all
flavin was reduced. The selective reduction could be monitored
by UV spectroscopy (Figure 2A,B). The nonreduced model
compounds in series a possess absorption bands around 390
and 410 nm, due to the benzylated deazaflavin chromophore
(Figure 2A). The oxidized flavin possesses absorption maxima
at 350 and 450 nm.38-40 In the reduced state, the absorption
around 450 nm disappears, and the absorption at 350 nm is
(38) Bruice, T. C. Acc. Chem. Res. 1980, 13, 256-262.
(39) Walsh, C. Acc. Chem. Res. 1980, 13, 148-155.
(40) Dudley, K. H.; Ehrenberg, A.; Hemmerich, P.; Mu¨ller, F. HelV.
Chim. Acta 1964, 47, 1354-1383.

7322 J. Am. Chem. Soc., Vol. 121, No. 32, 1999
Epple and Carell
cence data of all model compounds 1-4 clearly show a strongly
reduced deazaflavin fluorescence in all model compounds,
compared to the reference deazaflavins 25 and 31. In agreement
with the r-6 distance dependence of energy-transfer rates,
according to Fo¨rster theory,44 we observe the strongest fluo-
rescence reduction (95% deazaflavin fluorescence quenching)
and hence the best energy transfer in the model compounds 1a
and b, which feature the closest cofactor arrangement. The most
intense deazaflavin fluorescence, and consequently the least
efficient energy transfer, was determined for the model com-
pounds 3a/b and 2a/b, in which both cofactors are separated
by two or four proline units, respectively. Although the proline
spacer present only “semirigid” bridging units, the average
distance between the two cofactors is clearly increased if proline
units are inserted between the deazaflavin and the dimer. In
the model compounds 3a/b and 2a/b, stacking of the two
chromophores on top of each is not possible, which excludes
that the observed fluorescence quenching is due to formation
of a cofactor-cofactor complex. In summary, the data suggest
that the energy flux from the deazaflavin to the reduced and
deprotonated flavin is most efficient at small cofactor-cofactor
distances and decreases in efficiency as the cofactor separation
is increased.
Action Spectra and Investigation of the Splitting Reaction.
To firmly establish that the fluorescence reduction of the
deazaflavins in the model compounds is caused by an intramo-
lecular energy transfer to the FlH- unit, action spectra (splitting
rates depending on the irradiation wavelengths) of the model
compounds were recorded. Previous measurements showed that
deazaflavins alone are unable to initiate the splitting of
cyclobutane pyrimidine dimers.25 If, however, an energy transfer
from the deazaflavin to the reduced and deprotonated flavin
occurs, a significant shift in the action spectrum should be
observed. With our model compounds, all action spectra were
recorded in ethylene glycol, and the splitting rates were
determined as described, in detail.30 Briefly, the model com-
pounds were dissolved, reduced with dithionite, and irradiated
at various wavelengths between 360 and 470 nm. HPLC-analysis
of the reaction solutions after certain irradiation time intervals
allowed the determination of the conversion rate at each
wavelength by quantification of the amount of starting material
and product.
Three representative action spectra of the model compounds
5, 4a, and 4b are depicted in Figure 2C. The reference
compound 5, which contains only a flavin cofactor and no
deazaflavin, exhibits maximal absorption at 370 nm. This is
also the wavelength at which maximal repair is observed. The
model compound 4a and all other series a model compounds,
in contrast, show the most intense absorption around 390 and
410 nm due to the presence of the benzylated deazaflavin. In
the action spectra of these model compounds, maximal repair
is observed around 400 nm, which proves that the light that is
absorbed by the deazaflavin chromophore is transferred to the
reduced and deprotonated flavin unit, where it is used to drive
the cleavage reaction. Model compound 4b and all model
compounds of series b exhibit maximal repair at 430 nm. This
is, again, the wavelength at which the deprotonated deazaflavin
absorbs most intensely. The result again shows that the light
that is absorbed by the “quinoid” deazaflavin is used to drive
the splitting reaction. The action spectra consequently support
the hypothesis that both types of deazaflavins act in our model
compounds solely as light-harvesting photoantennas. Absorption
of light by the deazaflavin and subsequent energy transfer yields
Figure 3. (A) Fluorescence spectra of the reference compound 25 (s)
and of the model compounds 1a-4a (1a, ‚‚-‚‚; 4a, ‚-‚; 3a, ‚‚‚; 2a,
- - - ). (B) Fluorescence spectra of the reference compound 31 (s) and
of the model compounds 1b-4b (1b, ‚‚-‚‚; 4b, ‚-‚; 3b, ‚‚‚; 2b, - - -).
All fluorescence spectra were recorded in ethylene glycol at 10-6
concentration.
M
strongly reduced. The absorption bands of the deazaflavin unit,
however, remain intact, indicating the successful reduction of
only the flavin unit. Figure 2B shows the reduction of the model
compounds of series b. These model compounds possess a
“quinoid” deazaflavin, which absorbs most intensely at 430 nm.
Again, only the absorption around 450 nm of the oxidized flavin
disappears upon addition of the dithionite solution. The absorp-
tion at 430 nm remains, which proves that the “quinoid”
deazaflavin stays in its oxidized state. The obtained UV spectra
of the reduced solutions of all series b model compounds are
in perfect agreement with the UV data of type II DNA
photolyases in the active state.41 This result underlines that these
model compounds perfectly simulate the spectroscopic proper-
ties of these enzymes.41,42
dFl f FlH- Energy-Transfer Efficiency. The efficiency
of the dFl f FlH- energy transfer was investigated in all model
compounds using steady-state fluorescence spectroscopy (Figure
3A,B). The benzylated deazaflavins exhibit a strong fluorescence
at 430 nm (Figure 3A). All model compounds in series b possess
a strong fluorescence band at 470 nm (Figure 3B). If energy
donation from the deazaflavin chromophore to an energy
acceptor occurs, this is expected to lead to a quenching of the
deazaflavin fluorescence. Increasing energy-transfer efficiency
should yield decreasing fluorescence intensities.43 The fluores-
(41) Eker, A. P. M.; Hessel, J. K. C.; Dekker, R. H. Photochem.
Photobiol. 1986, 44, 197-205.
(42) Yasui, A.; Eker, A. P. M.; Yasuhira, S.; Yajima, H.; Kobayashi,
T.; Takao, M.; Oikawa, A. EMBO J. 1994, 13, 6143-6151.
(43) Becker, H. G. O. Einfu¨hrung in die Photochemie, 3. Auflage ed.;
Deutscher Verlag der Wissenschaften: Berlin, 1991.
(44) Fo¨rster, T. Discuss. Faraday Soc. 1959, 27, 7-17.

Efficient Light-Dependent DNA Repair
J. Am. Chem. Soc., Vol. 121, No. 32, 1999 7323
an excited FlH-* chromophore, which initiates an electron
transfer to the dimer lesion. All three action spectra are in full
agreement with the dFl f FlH- energy-transfer model.
Since the energy-transfer efficiency is maximal in the model
compounds 1a and 1b, which possess the closest cofactor-
cofactor arrangement, one would predictsbased on our knowl-
edge about the flavin-deazaflavin systemsmost efficient dimer
splitting in these model systems.
To investigate this hypothesis, the splitting efficiencies of
all model compounds were investigated upon irradiation with
the wavelength at which maximal repair occurs. The model
compounds of series a were irradiated at 400 nm, and the
splitting reactions of all model compounds belonging to series
b were investigated at 430 nm. The obtained splitting rates for
the model compounds are depicted in Figure 4A for series a
and in Figure 4B for series b. It is evident that all model
compounds possess largely differing cleavage rates. The quan-
tum yields for the dimer splitting were calculated from these
rate data and the cofactor absorptions. The quantum yields are
listed in Table 1, together with the estimated averaged cofactor
distances, which were obtained from molecular modeling
investigations. The obtained distance data were verified by short-
time spectroscopic measurements.45
The presentation of the repair rate data shows, in sharp
contrast to our expectations, that the model compounds 1a and
1b, which possess the closest cofactor-cofactor distance and
the most efficient energy transfer, are unable to perform an
acceptable repair. Both compounds possess long half-life times
for the splitting reaction of 78 (1a) and 32 min (1b), as
extrapolated from the data shown in Figure 4. The reference
compound 5, in contrast, possesses a significantly shorter half-
life time of t1/2 ) 18 min at 366 nm (data not shown) and 32
min at 400 nm irradiation, despite the lack of a second
deazaflavin cofactor. Only at 430 nm is this compound slower
compared to 1b because the flavin possess only negligible
absorption at this wavelength. The deazaflavin consequently
inhibits the repair process! During the irradiation of 1a and 1b,
no products other than the expected cleavage compounds were
detected. This excludes photodecomposition as a possible
explanation for the observed very slow cleavage rates. Inves-
tigation of the whole model compound series showed that all
the model compounds which possess the largest cofactor-
cofactor distance cleave the dimer unit upon irradiation most
efficient. Both model compounds (2a and 2b), which possess
four proline spacer, split the dimer lesions by a factor of 40 (!)
faster than 1a and 1b. The model compounds 2a and 2b possess
t1/2 values of only 2 and 1 min, respectively. All model
compounds which possess intermediate cofactor-cofactor dis-
tances feature intermediate half-life times. t1/2 values of 13 and
8 min were measured for the model compounds 4a and 4b. The
two proline-containing model compounds, 3a and 3b, require
9 and 3 min to reach the 50% cleavage level.
Figure 4. (A) Repair rates determined for the model compounds 1a-
4a in series a (1a, [; 4a, 2; 3a, b; 2a, 9; 5, ±). (B) Repair rates
determined for the model compounds 1b-4b in series b (1b, [; 4b,
2; 3b, b; 2b, 9; 5, ±). The model compounds were measured in
ethylene glycol as the solvent and were reduced (flavin only) with
sodium dithionite. The solutions were irradiated with monochromatic
light at 400 (A) and 430 nm (B). The amount of photocleaved product
versus the irradiation time was determined by reversed-phase HPLC
as reported earlier in detail.30 (a) Irradiation, (b) reoxidation, c) HPLC
analysis.
The quantum yields for the splitting reaction of all model
compounds were determined by ferrioxalate actinometry.46 In
agreement with the slow dimer splitting, we measured very low
quantum yields of approximately 1% for the model compounds
1a and 1b. These are the lowest values measured in the whole
series. The model compounds 2a and 2b, which repair the dimer
most efficiently and possess the largest cofactor-cofactor
distance, show the best repair quantum yields of Φ ≈ 10%.
The model compounds 3a/b and 4a/b possess intermediate
quantum yields, consistent with the observed rate data.
In summary, the cleavage data unequivocally show that the
repair activity is strongly influenced by the distance between
(45) Michel-Beyerle, M. E.; Carell, T., unpublished results. The deaza-
flavin is also able to perform an energy transfer to the oxidized flavin.
This energy transfer was investigated by short-time laser fluorescence
spectroscopy. The distances between the cofactors were estimated from the
fluorescence lifetimes according to the Fo¨rster theory.
(46) Hatchard, C. G.; Parker, C. A. Proc. R. Soc. A 1956, 235, 518-
536.

7324 J. Am. Chem. Soc., Vol. 121, No. 32, 1999
Epple and Carell
Table 1. Repair Quantum Yields, Φrepair,a of 1a-4a (Series a)
Measured upon Irradiation at 400 nm, of 1b-4b (Series b)
Measured upon Irradiation at 430 nm, and of the Reference
Compound 5 Measured at 400 and 430 nmb
dependence and decrease more rapidly at larger distances.1,4,49
The increasing through-bond distances between the flavin and
the deazaflavin cofactors in the model compounds 3a/b and 2a/
b, which possess two or even four proline spacers,27,28 respec-
tively, should affect the short-circuit intercofactor electron
transfer more strongly than the beneficial energy transfer from
the deazaflavin to the FlH-. This is indeed observed: the repair
rate rises in both series of model compounds by a factor of
approximately 50 from t1/2 ) 78 min (1a) and t1/2 ) 32 min
(1b) to about 1-2 min in 2a and 2b, while at the same time,
the energy-transfer efficiency shrinks to about one-quarter in
2a and 2b compared to that in 1a and 1b (Figure 3A,B). The
electron-transfer model could also explain why all model
compounds in series b are more “repair-active” compared to
their counterparts in series a. In the models of series b, the
deazaflavin is deprotonated and negatively charged, which
should attenuate its electron-accepting properties. We measured
almost maximal repair efficiency in series b already with
compound 3b (60% efficiency compared to 2b). In series a, in
contrast, compound 3a is still a factor of 3 slower compared to
2a. The electron-transfer model is also consistent with the
available redox potentials. The light-excited, reduced, and
deprotonated flavin possesses an estimated redox potential of
-2.8 V.50 The potential for the reduction of the dimer unit is
approximately -2.2 V,50 and the reduction potentials of the
deazaflavins are likely to be between -0.8 and -1.5 V.51 An
electron transfer from the light excited, reduced, and deproto-
nated flavin to the deazaflavin is, consequently, thermodynami-
cally more favorable than that to the dimer unit.
approximate
Fl-dFl
distance (Å) (min)
model
compound
flavin
state
deazaflavin
state
t1/2
Φ
2a
3a
4a
1a
5
2b
3b
4b
1b
5
FlredH- dFlox-OBn
FlredH- dFlox-OBn
FlredH- dFlox-OBn
FlredH- dFlox-OBn
FlredH- dFlox-OBn
FlredH- dFlox-O-
FlredH- dFlox-O-
FlredH- dFlox-O-
FlredH- dFlox-O-
FlredH- dFlox-O-
19
16
12
7
-
19
16
12
7
2
9
13
78
32
1
3
8
32
50
0.10
0.03
0.02
0.005
0.08
0.10
0.06
0.02
0.01
0.08
-
a Photon flux of the light source was determined by ferrioxalate
b
actinometry. t1/2 is the half-life time for the cleavage of the model
compounds. The average cofactor distance was determined on the basis
of computer modeling (Macromodel, AMBER* force field). The
distance estimates are supported by data from short-time fluorescence
spectroscopic measurements.45 Estimated error in ΦRepair ) (20%.
Estimated error for the average Fl-dFl distance ) (2 Å.
Summary and Conclusion
These studies reveal that the repair activity in our model
compounds critically depends on the cofactor-cofactor distance
and the cofactors’ redox and protonation states. At large
cofactor-cofactor distances, the deazaflavin functions exclu-
sively as a photoantenna and channels light energy to the
reduced flavin, where it is used to drive the electron transfer to
the dimer unit. The dimer undergoes spontaneous cleavage as
its radical anion. At shorter distances, the light-channeling
process is more efficient, but the repair activity is strongly
reduced. All our data, in combination with the available redox
potentials, suggest a competitive light-induced electron transfer
from the reduced flavin to the deazaflavin. This process
represses the splitting reaction at close cofactor-cofactor
distances. To achieve maximal repair efficiency, both cofactors
have to be separated until this competitive process becomes
negligible. It is very likely that photolyases faced the same
problem and that they had to evolve strategies to circumvent
the possibility of intrinsic electron transfer between the essential
cofactors. From our data and based on the knowledge of DNA
photolyase crystal structures, we argue that they may have
solved this problem by tuning two main parameters: (1) The
deazaflavin is deprotonated to attenuate its electron acceptor
function. (2) The flavin and the deazaflavin distance is extended
until the electron donation to the deazaflavin is efficiently
suppressed. In addition, deazaflavin-containing photolyases
Figure 5. Schematic representation of the model of the electron-transfer
possibilities between a FADH-*, a deazaflavin (F0), and a thymine
dimer (T)T) after excitation of the FADH-. EET ) Excitation energy
transfer.
the cofactors. Increasing flavin-deazaflavin distances yields
decreasing half-life times for the splitting reaction. The cleavage
data show that the presence of a deazaflavin cofactor in close
proximity to the electron-donating FlH-* inhibits the cleavage
reaction. At larger cofactor-cofactor distances, the splitting rate
increases, although the energy flux from the deazaflavin to the
FlH- is reduced.
This counterintuitive result shows that a large cofactor
separation is required if efficient dimer splitting is desired. The
results suggest that the deazaflavin cofactor can intercept the
electron transfer from the flavin to the dimer lesion. At close
cofactor distances, such as in 1a/1b, this property fully inhibits
the repair reaction. Due to the strong distance dependence of
the inhibitory effect, we believe that the deazaflavin functions
as an internal electron acceptor and induces a short-circuit
electron transfer between the reduced flavin and the deazaflavin
at short distances. If the deazaflavin is positioned too close to
the reduced and deprotonated flavin, it functions as an alternative
electron acceptor, as shown in Figure 5. This electron-transfer
model is supported by all fluorescence and quantum yield data.
In contrast to the through-space energy-transfer process, which
is effective over relatively large distances (100 Å),44,47,48
through-bond electron-transfer reactions possess an e-r distance
(47) Lankiewicz, L.; Malicka, J.; Wiszk, W. Acta Biochim. Pol. 1997,
44, 477-490.
(48) Van der Meer, B. W.; Coker, G., III; Chen, S.-Y. Resonance Energy
Transfer; VCH Verlagsgesellschaft: Weinheim, 1994.
(49) Marcus, R. A.; Sutin, N. Biochim. Biophys. 1985, 811, 265-322.
(50) Scannel, M. P.; Fenick, D. J.; Yeh, S.-R.; Falvey, D. E. J. Am. Chem.
Soc. 1997, 119, 1971-1977.
(51) van der Plas, H. C.; Link, P. A. J. Org. Chem. 1986, 51, 1602-
1606.

Efficient Light-Dependent DNA Repair
J. Am. Chem. Soc., Vol. 121, No. 32, 1999 7325
1289m, 1230m, 1206m, 1175m, 1072w, 1015w, 885w, 807w. 1H NMR
(400 MHz, (CD3)2SO): δ 0.88 (t, J ) 6.7 Hz, 3 H), 1.22 (s, 9 H), 1.32
(m, 4 H), 1.58 (m, 2 H), 2.41 (s, 3 H), 2.50 (s, 3 H), 3.41 (q, J ) 5.9
Hz, 2 H), 3.88 (t, J ) 7.4 Hz, 2 H), 4.67 (t, J ) 5.9 Hz, 2 H), 6.96 (t,
J ) 5.9 Hz, 1 H), 7.85 (s, 1 H), 7.93 (s, 1 H). 13C NMR (100 MHz,
(CD3)2SO): δ 13.78, 18.68, 20.76, 21.85, 27.00, 27.93 (3C), 28.55,
36.95, 40.77, 44.00, 77.80, 116.16, 130.94, 131.41, 134.18, 135.72,
135.93, 146.43, 148.96, 154.63, 155.74, 159.35. MS (FAB): m/z 456
(100, [M + 1]+), 356 (53), 313 (71), 243 (36), 198 (28). Anal. Calcd
for C24H33N5O4 + 0.5H2O (464.57): C, 62.05; H, 7.38; N, 15.07.
Found: C, 62.21; H, 7.23; N, 15.04.
10-(2-Aminoethyl)-7,8-dimethyl-3-pentyl-3H,10H-benzo[g]pteri-
dine-2,4-dione, TFA Salt (7). A solution of the Boc-protected flavin
18 (180 mg, 0.40 mmol) in TFA/H2O (95:5) (5 mL) was stirred at
room temperature for 90 min. The reaction solution was evaporated to
dryness in vacuo. The product 7 was filtered off after the addition of
Et2O. 7 was obtained as a yellow powder and used without further
purification (180 mg, 97%). Rf (CHCl3/MeOH/AcOH, 5:1:1): 0.38.
Mp: 196-198 °C. IR (KBr): 3433w, 3100w, 3033w, 2958w, 2856w,
1689s, 1660s, 1584s, 1549s, 1461m, 1436m, 1411w, 1349w, 1254m,
1201m, 1133m, 1133m, 1020w, 798w, 722w. 1H NMR (500 MHz,
(CD3)2SO): δ 0.88 (t, J ) 6.9 Hz, 3 H), 1.32 (m, 4 H), 1.59 (m, 2 H),
2.41 (s, 3 H), 2.51 (s, 3 H), 3.24 (t, J ) 6.5 Hz, 2 H), 3.89 (t, J ) 7.4
Hz, 2 H), 4.88 (t, J ) 6.5 Hz, 2 H), 7.84 (s, 1 H), 7.97 (s, 1 H), 8.02
(br s, 2-3 H). 13C NMR (125 MHz, (CD3)2SO): δ 13.80, 18.70, 20.58,
21.85, 26.98, 28.54, 36.48, 40.84, 41.11, 115.63, 130.47, 131.26, 134.14,
136.04, 136.36, 146.98, 149.61, 154.66, 159.25. HRMS (FAB) for
C19H26N5O2: (MH+) calcd 356.2086, found 356.2092.
feature, despite the large cofactor-cofactor distance, an excellent
energy-transfer efficiency (98%). This transfer efficiency is far
higher than that in our best compounds, 2a and 2b, which
possess a random cofactor orientation. This suggests that
photolyases also optimized the cofactors’ orientations, to make
the energy transfer most efficient even at the required large
cofactor distance of 17 Å. Our data show that the large
cofactor-cofactor distances observed in photolyases are an
essential structural feature that will undoubtedly strongly
determine their catalytic efficiency. The distance is not, as
previously hypothesized,52 “not optimized during eVolution”.
Although the strict requirement for a large cofactor separation
was shown with flavin- and deazaflavin-containing model
compounds, we believe that similar reasons apply to the flavin-
folate system in folate-containing type I photolyases. These types
of photolyases also feature a surprisingly large cofactor-
cofactor distance, and they possess additionally an unfavorable
cofactor-cofactor orientation.
Experimental Part
General Methods. All materials were obtained from commercial
suppliers and were used without further purification. Solvents of
technical quality were distilled prior to use. The aqueous buffers were
prepared using deionized water. For reactions under an inert gas
atmosphere, nitrogen of standard quality was used. For analytical thin-
layer chromatography, precoated silica gel plates (Merck 60-F254) were
used. Staining of amino compounds was performed with ninhydrin.
Flash chromatography was performed using silica gel (Merck, 0.040-
0.063 mm) and silica gel-H (Fluka, 0.005-0.040 mm). Melting points
are uncorrected and were determined on a Bu¨chi Smp 20. IR spectra
were recorded on a Perkin-Elmer 1600 FT-IR using KBr pellets or
CHCl3 solutions. UV spectra were recorded on a Varian Cary 5 UV-
vis spectrophotometer in 1-cm quartz cuvettes. Fluorescence spectra
were measured on a Spex 1680 0.22m double spectrometer with a
bandwidth of 3.4 nm in 1-cm quartz cuvettes. NMR spectra were
recorded on a Varian Gemini 200 (200 MHz (1H), 50 MHz (13C)), a
Varian Gemini 300 (300 MHz (1H), 75 MHz (13C)), and a Bruker AM-
500 (500 MHz (1H), 125 MHz (13C)). The chemical shift (δ) is reported
in ppm downfield from tetramethylsilane (TMS, δ ) 0 ppm).
Alternatively, the resonances of residual solvent protons were used as
the reference. EI mass spectra and FAB mass spectra were measured
by the staff of the mass spectrometry facilities of the ETH Zurich on
a Hitachi-Perkin-Elmer VG TRIBRID with 70-eV ionization energy
(EI) and on a ZAB-2 SEQ with 3-nitrobenzyl alcohol as the matrix
(FAB). MALDI-TOF spectra were measured on a Bruker Reflex
spectrometer. ESI spectra were recorded on a Finnigan TSQ 7000.
Elemental analyses were performed at the microanalysis laboratory of
the ETH Zurich. HPLC chromatograms were obtained with a Knauer
HPLC instrument (HPLC pumps 64, Knauer variable-wavelength UV
detector, Knauer degasser) using HPLC grade solvents.
10-[2-({[(tert-Butyl)oxy]carbonyl}amino)ethyl]-7,8-dimethyl-3-
pentyl-3H,10H-benzo[g]pteridine-2,4-dione (18). 1-Bromopentane
(588 mg, 3.89 mmol, 0.48 mL) was added with a syringe slowly to a
suspension of the flavin compound 15 (500 mg, 1.30 mmol) and dry
Cs2CO3 (0.6 g, 1.83 mmol) in DMF (50 mL, dried over 4-Å molecular
sieves). The reaction mixture was stirred at room temperature for 3 h.
The reaction mixture was diluted with CHCl3 (200 mL) and washed
three times with H2O (3 × 80 mL). The organic phase was separated,
dried with MgSO4, filtered, and evaporated to dryness. The product 18
was obtained after column chromatography on silica gel-H (d ) 3 cm,
l ) 18 cm) with CHCl3/MeOH (20:1) as the eluent and recrystallization
from MeOH/H2O as yellow needles (140 mg, 24%). Rf (CHCl3/MeOH,
20:1): 0.46. Mp: 228-230 °C. IR (KBr): 3329w, 2931w, 2867w,
1706m, 1691m, 1644s, 1583s, 1547s, 1450m, 1433m, 1366m, 1336m,
tert-Butyl 2-{10-[2-({[(tert-Butyl)oxy]carbonyl}amino)ethyl]-7,8-
dimethyl-2,4-dioxo-3H,10H-benzo[g]pteridin-3-yl}acetate (16). A
suspension of the flavin 15 (1 g, 2.59 mmol), Cs2CO3 (1 g, 3.07 mmol),
and tert-butyl bromoacetate (1.5 g, 7.69 mmol) in DMF (100 mL) was
stirred for 5 h at room temperature. The reaction mixture was diluted
with CHCl3 (250 mL) and washed three times with H2O (150 mL).
The organic phase was separated, dried with MgSO4, filtered, and
concentrated in vacuo to dryness. The crude product 16 was precipitated
through the addition of acetone/H2O and filtered off. 16 was obtained
after column chromatography on silica gel-60 (d ) 3 cm, l ) 18 cm,
CHCl3/MeOH, 15:1) and recrystallization from EtOAc/Ether as yellow
needles (0.93 g, 75%). Rf (CHCl3/MeOH, 10:1): 0.73. Mp: 198-200
°C. IR (CHCl3): 3458w, 3007m, 3000m, 1742m, 1709s, 1661s, 1628w,
1
1584s, 1549s, 1504m, 1460m, 1369m, 1157s, 855w. H NMR (400
MHz, (CD3)2SO): δ 1.22 (s, 9 H), 1.43 (s, 9 H), 2.42 (s, 3 H), 2.52 (s,
3 H), 3.45 (q, J ) 5.9 Hz, 2 H), 4.54 (s, 2 H), 4.70 (t, J ) 5.9 Hz, 2
H), 6.97 (t, J ) 5.9 Hz, 1 H), 7.90 (s, 1 H), 7.97 (s, 1 H). 13C NMR
(100 MHz, (CD3)2SO): δ 18.68, 20.83, 27.63 (3C), 27.90 (3C), 36.95,
42.90, 44.39, 77.84, 81.35, 116.34, 131.00, 131.69, 134.45, 135.36,
136.13, 147.03, 149.10, 154.12, 155.75, 159.18, 166.97. MS (FAB):
m/z 500 (100, [M + 1]+), 388 (9), 344 (28), 198 (9). Anal. Calcd for
C25H33N5O6 (499.57): C, 60.11; H, 6.66; N, 14.02. Found: C, 60.16;
H, 6.48; N, 14.00.
2-[10-(2-Aminoethyl)-7,8-dimethyl-2,4-dioxo-3H,10H-benzo[g]-
pteridin-3-yl]acetic Acid, TFA Salt (17). A solution of 16 (0.8 g,
0.160 mmol) in TFA/H2O (95:5) (20 mL) was stirred for 2 h at room
temperature. The reaction mixture was subsequently evaporated to
dryness, and the product was precipitated through the addition of Et2O
to the residual oil. The product was filtered off, washed with Et2O,
and dried under high vacuum. 17 (0.73 g, quantitative) was used in
the next step without further purification. Rf (CHCl3/MeOH/AcOH, 5:1:
1): 0.45. Mp: 153-155 °C. IR (KBr): 3600-3200m, 3100-2700w,
1702m, 1661m, 1584m, 1548s, 1464w, 1352w, 1325w, 1247m, 1197m,
1
1136m, 1036w, 936w, 802w, 725w. H NMR (500 MHz, (CD3)2SO):
δ 2.43 (s, 3 H), 2.53 (s, 3 H), 3.27 (m, 2 H), 4.59 (s, 2 H), 4.91 (t, J
) 6.4 Hz, 2 H), 7.89 (s, 1 H), 8.01 (s, 1 H), 8.05 (br s, 2-3 H), 11.5-
13.5 (br s, 1 H). 13C NMR (125 MHz, (CD3)2SO): δ 18.70, 20.63,
39.00, 41.48, 42.36, 115.77, 130.78, 131.36, 134.44, 135.85, 136.39,
147.60, 149.80, 154.23, 159.08, 169.21. HRMS (FAB) for C18H18N5O6F3
(457.37): (MH+ - CF3COO) calcd 344.1359, found 344.1359.
2-{10-[2-({[(Fluoren-9-yl)methoxy]carbonyl}amino)ethyl]-7,8-
dimethyl-2,4-dioxo-3H,10H-benzo[g]pteridin-3-yl}acetic Acid (9). A
suspension of the flavin amino acid 17 (250 mg, 0.55 mmol) in aqueous
(52) Park et al. proposed, for the folate-containing type-I photolyases,
that the energy transfer within photolyases is not rate limiting. In this case,
we would expect identical repair rates for all prepared model compounds,
independent from the flavin-deazaflavin distance. This is clearly not
observed. All compounds feature largely different repair efficiencies.

7326 J. Am. Chem. Soc., Vol. 121, No. 32, 1999
Epple and Carell
K2CO3 solution (9%, 10 mL) was stirred and cooled to 0 °C. Then,
Fmoc-OSu (300 mg, 0.89 mmol), dissolved in 2 mL of DMF, was
added. The solution was stirred for 90 min and allowed to warm to
room temperature during this time. The reaction mixture was diluted
with H2O (250 mL) and acidified (pH ) 3.5) through the addition of
a 10% citric acid solution (100 mL). This solution was extracted three
times with CHCl3 (3 × 300 mL). The combined organic phases were
separated, dried with MgSO4, filtered, and concentrated in vacuo to
dryness. Et2O was added to the residual oil, and the reaction product
was filtered off (9, 250 mg, 80%). Rf (CHCl3/MeOH/AcOH, 5:1:1):
0.80. Mp: 179-181 °C. IR (KBr): 3600-3200m, 3056w, 2956w,
1711s, 1661s, 1583s, 1547s, 1450m, 1411w, 1350w, 1322w, 1261m,
vacuo. The solid material was recrystallized from CHCl3/MeOH. The
product 24 was obtained as yellow needles (0.4 g, 70%). Rf (CHCl3/
MeOH, 20:1): 0.67. Mp: 216-217 °C. IR (KBr): 3434m, 3044w,
2932w, 2867w, 1702m, 1638s, 1607s, 1537s, 1500m, 1457m, 1411w,
1
1367w, 1236s, 1167m, 1126w, 1006w, 956w, 798w. H NMR (500
MHz, CDCl3): δ 0.90 (t, J ) 7.1 Hz, 3 H), 1.37 (m, 4 H), 1.46 (s, 9
H), 1.70 (m, 2 H), 3.55 (m, 2 H), 4.05 (t, J ) 7.6 Hz, 2 H), 4.82 (br
s, 2 H), 5.09 (t, J ) 6.0 Hz, 1 H), 5.46 (s, 2 H), 7.13 (dd, J ) 2.1 Hz,
J ) 8.8 Hz, 1 H), 7.36 (t, J ) 7.2 Hz, 1 H), 7.42 (t, J ) 7.2 Hz, 2 H),
7.53 (d, J ) 7.2 Hz, 2 H), 7.77 (d, J ) 8.8 Hz, 1 H), 7.93 (s, 1 H),
8.77 (s, 1 H). 13C NMR (125 MHz, CDCl3): δ 14.03, 22.51, 27.63,
28.38 (3C), 29.21, 37.29, 41.35, 43.85, 71.10, 80.11, 99.39, 111.96,
116.23, 116.67, 127.68 (2C), 128.39, 128.72 (2C), 132.94, 135.87,
142.15, 143.27, 156.49, 156.69, 157.26, 162.17, 165.43. MS (FAB):
m/z 533 (100, [M + 1]+), 460 (13), 433 (14), 390 (11). Anal. Calcd
for C30H36N4O5 (532.64): C, 67.65; H, 6.81; N, 10.52. Found: C, 67.46;
H, 6.73; N, 10.45.
10-(2-Aminoethyl)-8-benzyloxy-3-pentyl-5-carba-3H,10H-benzo-
[g]pteridine-2,4-dione, TFA Salt (10). A solution of the Boc-protected
deazaflavin derivative 24 (300 mg, 0.56 mmol) in TFA/H2O (95:5)
was stirred for 2 h at room temperature. The reaction mixture was
concentrated in vacuo, and the product 10 was crystallized through
the addition of Et2O to the residual oil. Compound 10 was filtered off
and was obtained as the yellow trifluoroacetic acid salt (300 mg, 97%).
Rf (CHCl3/MeOH/HOAc, 5:1:1): 0.72. Mp: 241-243 °C. IR (KBr):
3426m, 3033w, 2944w, 1703m, 1606s, 1535s, 1500w, 1461w, 1417w,
1372w, 1254m, 1189m, 1150w, 994w, 739w. 1H NMR (500 MHz,
(CD3)2SO): δ 0.86 (t, J ) 7.0 Hz, 3 H), 1.30 (m, 4 H), 1.56 (m, 2 H),
3.20 (m, 2 H), 3.86 (t, J ) 7.4 Hz, 2 H), 4.90 (t, J ) 6.3 Hz, 2 H),
5.38 (s, 2 H), 7.33 (dd, J ) 1.9 Hz, J ) 8.9 Hz, 1 H), 7.37-7.45 (m,
4 H), 7.53 (d, J ) 7.2 Hz, 2 H), 7.98 (br s, 3 H), 8.19 (d, J ) 8.9 Hz,
1 H), 8.97 (s, 1 H). 13C NMR (125 MHz, (CD3)2SO): δ 13.82, 21.84,
27.08, 28.59, 36.41, 40.18, 41.47, 70.53, 100.33, 111.62, 114.03, 116.33,
128.29 (2C), 128.38, 128.59 (2C), 134.21, 135.78, 142.03, 142.37,
155.65, 156.80, 161.43, 164.62. HRMS (FAB) for C25H28N4O3: (MH+)
calcd 433.22397, found 433.2247.
1
1231m, 1206m, 1011w, 761w, 742w. H NMR (400 MHz, (CD3)2-
SO): δ 2.29 (s, 3 H), 2.33 (s, 3 H), 3.51 (q, J ) 5.9 Hz, 2 H), 4.05 (t,
J ) 7.0 Hz, 1 H), 4.21 (d, J ) 7.0 Hz, 2 H), 4.53 (s, 2 H), 4.73 (t, J
) 5.9 Hz, 2 H), 7.38-7.95 (m, 11 H), 11.5-13.5 (br s, 1 H). 13C NMR
(100 MHz, (CD3)2SO): δ 18.57, 20.66, 37.06, 42.21, 43.62, 46.49,
65.63, 115.92, 119.98 (2C), 124.89 (2C), 126.91 (2C), 127.50 (2C),
130.99, 131.53, 134.36, 135.46, 136.06, 140.56 (2C), 143.66 (2C),
146.98, 149.10, 154.13, 156.41, 159.14, 169.23. HRMS (FAB) for
C31H27N5O6: (MH+) calcd 566.2040, found 566.2065.
6-{[2-({[(tert-Butyl)oxy]carbonyl}amino)ethyl]amino}-1H,3H-
pyrimidine-2,4-dione (21). The mono-Boc-protected ethylenediamine
2033 (6.5 g, 41 mmol, 1.5 equiv) was added to a solution of
6-chlorouracil 19 (4 g, 27 mmol) in n-butanol (100 mL). This solution
was heated for 4 h at reflux. The solution was cooled to room
temperature, and the n-butanol was distilled off in vacuo. The residual
material was dissolved in boiling water and then stored at 4 °C for 12
h. The colorless precipitate was filtered off and was once more
recrystallized from water. The product 21 was obtained as a colorless
microcrystalline powder (4.8 g, 65%). Rf (CHCl3/MeOH, 10:1): 0.18.
Mp: 221-223 °C. IR (KBr): 3310s, 3222s, 3100m, 2978m, 1726s,
1683s, 1596s, 1539s, 1449m, 1389m, 1364m, 1333m, 1277m, 1244m,
1223m, 1170m, 1106w, 1039w, 1017w, 987w, 964w, 829w, 806w,
1
767w, 548m. H NMR (400 MHz, (CD3)2SO): δ 1.38 (s, 9 H), 3.05
(m, 4 H), 4.48 (s, 1 H), 6.08 (br s, 1 H), 6.89 (br s, 1 H), 9.98 (br s,
1 H), 10.12 (br s, 1 H). 13C NMR (100 MHz, (CD3)2SO): δ 28.17
(3C), 38.81, 41.09, 72.54, 77.81, 150.80, 154.08, 155.74, 164.24. MS
(FAB): m/z 271 (96, [M + 1]+), 215 (22), 171 (31). Anal. Calcd for
C11H18N4O4 (270.29): C, 48.88; H, 6.71; N, 20.73. Found: C, 48.98;
H, 6.83; N, 20.70.
tert-Butyl 2-{10-[2-({[(tert-Butyl)oxy]carbonyl}amino)ethyl]-8-
benzyloxy-2,4-dioxo-5-carba-3H,10H-benzo[g]pteridin-3-yl}-
acetate (25). tert-Butyl bromoacetate (1.1 mL, 7.3 mmol) was added
dropwise to a suspension of the deazaflavin 22 (1 g, 2.2 mmol) and
dry Cs2CO3 (1 g, 3.06 mmol) in DMF (50 mL, dried over molecular
sieves (4 Å)) at 60 °C. The reaction mixture was stirred for 5 h at 60
°C. The reaction mixture was subsequently diluted with CHCl3 (200
mL), and the organic phase was washed three times with H2O (3 ×
100 mL). The organic phase was separated, dried with MgSO4, filtered,
and concentrated in vacuo. The residual material was subjected to
column chromatography on silica gel-60 (d ) 3 cm, l ) 18 cm) with
CHCl3/MeOH (20:1) as the eluent. The product 25 was recrystallized
from MeOH and obtained as yellow needles (0.62 g, 49%). Rf (CHCl3/
MeOH, 20:1): 0.74. Mp: 219-221 °C. IR (KBr): 3423m, 3044w,
2978w, 2933w, 1735m, 1691m, 1640s, 1609s, 1567m, 1537s, 1498s,
1456m, 1412m, 1370m, 1322w, 1289w, 1244s, 1189m, 1157m, 1039w,
8-Benzyloxy-10-[2-({[(tert-butyl)oxy]carbonyl}amino)ethyl]-5-
carba-3H,10H-benzo[g]pteridine-2,4-dione (22). A solution of 2,4-
(dibenzyloxy)benzaldehyde (23) (2.1 g, 6.6 mmol) and the uracil
derivative 21 (1 g, 3.7 mmol) in DMF (50 mL) was stirred for 20 h at
120 °C. The reaction solution was cooled to 4 °C, and the precipitated
yellow reaction product 22 was isolated by filtration. The product was
washed with Et2O and was recrystallized twice from MeOH. 22 was
obtained as a yellow, microcrystalline powder (0.9 g, 52%). Rf (CHCl3/
MeOH, 10:1): 0.41. Mp: 183-185 °C. IR (KBr): 3419m, 3133w,
2967w, 2811w, 1699s, 1661m, 1603s, 1561m, 1527s, 1492m, 1456m,
1405m, 1367m, 1242s, 1188m, 1167m, 1144m, 983w, 796w, 579w.
1H NMR (500 MHz, (CD3)2SO): δ 1.28 (s, 9 H), 3.36 (q, J ) 6.5 Hz,
2 H), 4.67 (br s, 2 H), 5.41 (s, 2 H), 7.23 (m, 2 H), 7.38 (m, 1 H), 7.44
(m, 2 H), 7.53 (d, J ) 7.1 Hz, 2 H), 7.71 (s, 1 H), 8.10 (d, J ) 9.0 Hz,
1 H), 8.90 (s, 1 H), 10.97 (s, 1 H). 13C NMR (125 MHz, (CD3)2SO):
δ 27.97 (3C), 36.81, 44.18, 70.34, 77.99, 99.91, 111.87, 114.85, 115.90,
127.91 (2C), 128.20, 128.51 (2C), 133.46, 135.89, 141.19, 142.87,
156.06, 156.50, 157.67, 162.24, 164.26. MS (FAB): m/z 463 (100,
[M + 1]+), 389 (8), 363 (21). Anal. Calcd for C25H26N4O5 + H2O
(480.53): C, 62.49; H, 5.87; N, 11.66. Found: C, 62.39; H, 5.87; N,
11.66.
8-Benzyloxy-10-[2-({[(tert-butyl)oxy]carbonyl}amino)ethyl]-3-
pentyl-5-carba-3H,10H-benzo[g]pteridine-2,4-dione (24). 1-Bro-
mopentane (0.4 mL, 3.31 mmol) was added dropwise to a suspension
of the deazaflavin 22 (0.5 g, 1.08 mmol) and dry Cs2CO3 (0.5 g, 1.53
mmol) in dry DMF (50 mL, dried over molecular sieves (4 Å)) at 60
°C. The reaction mixture was stirred for 2 h at this temperature. The
reaction mixture was diluted with CHCl3 (200 mL), and the organic
phase was washed three times with H2O (3 × 100 mL). The organic
phase was separated, dried with MgSO4, filtered, and concentrated in
1
990w, 937w, 828w, 796w. H NMR (500 MHz, CDCl3): δ 1.47 (s, 9
H), 1.48 (s, 9 H), 3.55 (q, J ) 7.3 Hz, 2 H), 4.73 (s, 2 H), 4.83 (br s,
1 H), 5.06 (t, J ) 6.1 Hz, 1 H), 5.47 (s, 2 H), 7.15 (dd, J ) 2.2 Hz, J
) 8.8 Hz, 1 H), 7.37 (m, 1 H), 7.42 (m, 2 H), 7.54 (d, J ) 7.3 Hz, 2
H), 7.77 (d, J ) 8.8 Hz, 2 H), 7.98 (d, J ) 1.5 Hz, 1 H), 8.80 (s, 1 H).
13C NMR (125 MHz, CDCl3): δ 28.10 (3C), 28.38 (3C), 37.26, 42.99,
44.05, 71.17, 77.29, 80.16, 81.97, 99.44, 111.56, 116.25, 116.96, 127.70
(2C), 128.41, 128.73 (2C), 133.02, 135.83, 142.56, 143.46, 156.71,
156.73, 161.98, 165.66, 167.42. MS (FAB): m/z 577 (100, [M + 1]+),
521 (9), 503 (10), 434 (21), 421 (44). Anal. Calcd for C31H36N4O7
(576.65): C, 64.57; H, 6.29; N, 9.72. Found: C, 64.42; H, 6.25; N,
9.65.
tert-Butyl 2-{10-[2-({[(tert-Butyl)oxy]carbonyl}amino)ethyl]-8-
hydroxy-2,4-dioxo-5-carba-3H,10H-benzo[g]pteridin-3-yl}acetate (31).
The benzylated compound 25 (300 mg, 0.52 mmol) was dissolved at
room temperature in acetic acid (30 mL). A suspension of Pd/BaSO4
(15 mg) in acetic acid (3 mL) was added, and the reaction mixture
was stirred in an H2 atmosphere for 20 h. The reaction mixture was
filtered through Celite, and the solution was evaporated in vacuo.

Efficient Light-Dependent DNA Repair
J. Am. Chem. Soc., Vol. 121, No. 32, 1999 7327
Diethyl ether was added to the residual oil to precipitate the product
31, which was filtered off and dried in vacuo (250 mg, quantitative).
Mp: 152-154 °C. IR (KBr): 3426m, 2978w, 2922w, 1744m, 1701m,
1639s, 1606s, 1578m, 1534m, 1511s, 1472m, 1394m, 1367m, 1267m,
1228m, 1161s, 1039w, 967w, 939w, 856w, 806w, 794w. 1H NMR (400
MHz, (CD3)2SO): δ 1.28 (s, 9 H), 1.42 (s, 9 H), 3.38 (q, J ) 6.0 Hz,
2 H), 3.40 (br s, 1 H), 4.50 (s, 2 H), 4.63 (t, J ) 6.0 Hz, 2 H), 7.04-
7.09 (m, 2 H), 7.26 (s, 1 H), 8.04 (d, J ) 8.9 Hz, 1 H), 8.90 (s, 1 H).
13C NMR (100 MHz, (CD3)2SO): δ 27.59 (3C), 27.96 (3C), 36.99,
42.23, 44.08, 77.75, 81.04, 100.96, 109.28, 115.27, 115.53, 134.07,
142.26, 143.55, 155.22, 155.76, 156.15, 161.46, 165.48, 167.27. HRMS
(FAB) for C24H30N4O7 (486.53): (M + 1+) calcd 487.2193, found
487.2192.
[10-(2-Aminoethyl)-8-benzyloxy-2,4-dioxo-5-carba-3H,10H-ben-
zo[g]pteridin-3-yl]acetic Acid, TFA Salt (26). A solution of 25 (680
mg, 1.2 mmol) in 15 mL of TFA/H2O (95:5) was stirred at room
temperature for 3 h. The reaction mixture was concentrated in vacuo,
and Et2O was added to the residual oil. 26 was filtered off, washed
with Et2O, and dried under high vacuum. 26 was obtained as a yellow
powder (640 mg, quantitative). Rf (CHCl3/MeOH/HOAc, 5:1:1): 0.09.
Mp: 251-253 °C. IR (KBr): 3422m, 3400-2300m, 1689m, 1603s,
1534s, 1494m, 1467m, 1416w, 1367w, 1317w, 1260m, 1194m, 1172m,
1133m, 1033w, 983w, 944w, 833w, 794w, 721w. 1H NMR (400 MHz,
(CD3)2SO): δ 3.26 (t, J ) 6.1 Hz, 2 H), 4.56 (s, 2 H), 4.94 (br s, 2 H),
5.40 (s, 2 H), 7.34-7.47 (m, 5 H), 7.54 (d, 2 H), 8.10 (br s, 3 H), 8.21
(d, J ) 8.8 Hz, 1 H), 9.00 (s, 1 H), 11.80-13.50 (br s, 1 H). 13C NMR
(100 MHz, (CD3)2SO): δ 36.22, 41.75, 41.80, 70.20, 100.32, 111.12,
114.40, 116.44, 128.33 (2C), 128.40, 128.61 (2C), 134.35, 135.77,
142.33, 142.87, 155.30, 156.90, 161.28, 164.91, 169.52. HRMS (FAB)
for C24H21N4O7F3 (534.45): (M+ - CF3COO) calcd 421.1512, found
421.1513.
2-{10-[2-({[(Fluoren-9-yl)methoxy]carbonyl}amino)ethyl]-8-ben-
zyloxy-2,4-dioxo-5-carba-3H,10H-benzo[g]pteridin-3-yl}acetic Acid
(11). The deazaflavin amino acid 26 (250 mg, 0.47 mmol) was
suspended in aqueous K2CO3 solution (9%, 15 mL) and cooled to 4
°C. A solution of Fmoc-OSu (300 mg, 0.89 mmol) in DMF (1 mL)
was slowly added. The suspension was stirred for another 90 min, and
the reaction mixture was allowed to warm to room temperature during
this time. The reaction mixture was diluted with water (200 mL) and
acidified (pH ) 3.5) through the addition of a aqueous citric acid
solution (10%, 100 mL). The mixture was extracted two times with
CHCl3. The combined organic phases were dried with MgSO4, filtered,
and concentrated in vacuo to dryness. The product 11 was dissolved
in a small amount of CHCl3 and precipitated through the addition of
Et2O. 11 was filtered off and dried under high vacuum (11, 220 mg,
73%). Rf (CHCl3/MeOH/AcOH, 5:1:1): 0.92. Mp: 202-204 °C. IR
(KBr): 3600-2500m, 3398m, 3333m, 3038w, 2952w, 1703s, 1605s,
1535s, 1490m, 1466m, 1417w, 1375w, 1254s, 1202m, 1152w, 996w,
792w, 742w. 1H NMR (400 MHz, (CD3)2SO): δ 3.46 (q, J ) 6.0 Hz,
2 H), 4.15 (t, J ) 7.0 Hz, 1 H), 4.27 (d, J ) 7.0 Hz, 2 H), 4.51 (s, 2
H), 4.74 (br s, 2 H), 5.30 (s, 2 H), 7.21 (dd, J ) 2.0 Hz, J ) 8.9 Hz,
1 H), 7.27 (t, J ) 7.5 Hz, 2 H), 7.35-7.43 (m, 8 H), 7.55 (d, J ) 7.5
Hz, 2 H), 7.70 (t, J ) 6.0, Hz 1 H), 7.86 (d, J ) 7.5 Hz, 2 H), 8.15 (d,
J ) 9.0 Hz, 1 H), 8.99 (s, 1 H), 12.69 (br s, 1 H). 13C NMR (100
MHz, (CD3)2SO): δ 37.09, 41.73, 43.66, 46.61, 65.71, 70.33, 99.80,
110.89, 115.05, 116.22, 120.06 (2C), 124.99 (2C), 126.99 (2C), 127.56
(2C), 128.01 (2C), 128.24, 128.50 (2C), 133.79, 135.70, 140.66 (2C),
142.49, 142.95, 143.73 (2C), 155.27, 156.28, 156.72, 161.34, 164.57,
169.58. MS (FAB) for C37H30N4O7 (642.67): 643 (100, [M + 1]+),
460 (13). HRMS (FAB) for C37H30N4O7: (MH+) calcd 643.2193, found
643.2201.
11.16 mmol) for 48 h. The solution was filtered off, and the resin was
washed three times for 3 min with NMP (5 mL) and three times for 3
min with DMF. After a negative Kaiser test, the resin was shaken with
20% piperidine in DMF (10 mL) for 15 min and washed three times
with DMF (5 mL). Fmoc-protected proline (1.38 g, 4.09 mmol) was
dissolved in NMP and shaken together with HOBt (0.73 g, 4.77 mmol)
and TBTU (1.53 g, 4.77 mmol) for 5 min. This solution and DIEA
(2.1 mL, 12.27 mmol) were added to the resin, and the slurry was
shaken for 3 h. The solution was filtered off, and the resin was washed
three times for 3 min with NMP (5 mL) and three times for 3 min
with DMF. After a negative Kaiser test, the resin was shaken with 20%
piperidine in DMF (10 mL) for 15 min and washed three times with
DMF (5 mL). The coupling of the proline was repeated either once or
three times for the synthesis of 13 and 14, respectively. The resin was
finally washed three times with DMF (5 mL), seven times with CH2Cl2
(10 mL), three times with MeOH (10 mL), and three times with Et2O
(10 mL). The resin was dried and shaken together with 15 mL of a
solution of TFA (95%), H2O (2.5%), and triisobutylsilane (TIS) (2.5%)
for 1 h. The solution was filtered off and evaporated in vacuo. The
peptides were precipitated through the addition of Et2O. The crude
peptides were purified by reversed-phase HPLC (RP18-column with a
H2O(1% TFA)/CH3CN gradient (100% water to 100% acetonitrile over
100 min). Yields: 13, 270 mg, 49%; 14, 550 mg, 50%.
Data for 13. Mp: 110-112 °C. IR (KBr): 3600-2500m, 3400m,
3200m, 3067m, 2956m, 1678s, 1650s, 1604s, 1567m, 1535s, 1494m,
1461m, 1417m, 1372m, 1254s, 1198s, 1139m, 1026w, 994w, 933w,
833w, 795w, 721w, 700w. 1H NMR (400 MHz, (CD3)2SO): δ 1.60-
1.95 (m, 6 H), 2.08 (m, 1 H), 2.26 (m, 1 H), 3.05-3.30 (m, 2 H),
3.30-3.75 (m, 4 H), 4.31 (m, 1 H), 4.44 (m, 1 H), 4.45 (s, 2 H), 4.70
(m, 2 H), 5.42 (s, 2 H), 7.06 (s, 1 H), 7.26 (dd, J ) 2.0 Hz, J ) 8.9
Hz, 1 H), 7.38-7.56 (m, 6 H), 8.15 (d, J ) 9.0 Hz, 1 H), 8.48 (br s,
1 H), 8.52 (t, J ) 5.8 Hz, 1 H), 8.95 (s, 1 H), 9.50 (br s, 1 H). 13C
NMR (100 MHz, (CD3)2SO): δ 23.36, 24.42, 27.65, 29.11, 35.46,
42.65, 42.97, 45.59, 46.69, 58.28, 59.84, 70.42, 99.78, 111.28, 115.13,
116.03, 128.26 (2C), 128.32, 128.54 (2C), 133.72, 135.81, 142.04,
142.58, 155.54, 156.16, 161.44, 164.52, 166.59, 168.88, 172.06. HRMS
(FAB) for C32H35N7O6: (MH+) calcd 614.2727, found 614.2724.
Data for 14. Mp: 186-188 °C. IR (KBr): 3600-3000m, 3430m,
2967w, 2878w, 1678s, 1644s, 1605s, 1561w, 1535s, 1489w, 1455m,
1411m, 1367w, 1311w, 1252m, 1191m, 1161m, 1133m, 1022w, 928w,
1
833w, 794w. H NMR (500 MHz, (CD3)2SO): δ 1.60-2.45 (m, 16
H), 3.05-3.70 (m, 10 H), 4.10-4.60 (m, 4 H), 4.44 (s, 2 H), 4.70 (m,
2 H), 5.44 (s, 2 H), 7.04 (s, 1 H), 7.25 (dd, J ) 2.0 Hz, J ) 8.8 Hz,
1 H), 7.37-7.55 (m, 6 H), 8.16 (d, J ) 9.0 Hz, 1 H), 8.32 (t, J ) 5.8
Hz, 1 H), 8.44 (br s, 1 H), 8.96 (s, 1 H), 9.39 (br s, 1 H). 13C NMR
(125 MHz, (CD3)2SO): δ 23.42, 24.14, 24.26, 24.40, 27.38, 27.58,
28.92, 30.67, 35.68, 42.65, 43.09, 45.74, 46.35, 46.54, 46.61, 57.46,
57.82, 58.17, 59.40, 70.35, 99.78, 111.28, 115.24, 116.01, 128.24 (2C),
128.46, 128.49 (2C), 133.62, 135.85, 142.00, 142.66, 155.51, 156.17,
161.45, 164.51, 165.98, 168.56, 168.85, 172.86. MS (FAB) for
C42H49N9O8: m/z 831 (16, [M + Na]+), 809 (100, [M + 1]+), 616 (4).
Flavin-Deazaflavin Peptide 27. The Fmoc-protected flavin amino
acid 9 (285 mg, 0.50 mmol) was stirred together with BOP (1.00 g,
2.27 mmol) in DMF (30 mL) for 10 min. The deazaflavin 10 (275 mg,
0.50 mmol) was dissolved in 5 mL of DMF and added. Twenty drops
of NEt3 were added, and the reaction mixture was stirred for 90 min at
room temperature. The reaction solution was diluted with water (100
mL) and extracted three times with CHCl3. The combined organic
phases were dried with MgSO4, filtered, and concentrated in vacuo.
The product 27 was dissolved in a small amount of CHCl3 and
precipitated with Et2O. 27 was obtained after column chromatography
on silica gel 60 (d ) 3 cm, l ) 18 cm) with CHCl3/MeOH (20:1) as
an orange-yellow powder (160 mg, 33%). Rf (CHCl3/MeOH, 10:1):
0.58. Mp: 182-184 °C. IR (KBr): 3429m, 2950w, 2933w, 2856w,
1706m, 1639m, 1607s, 1578m, 1541s, 1456m, 1406w, 1250m, 1228m,
Solid-Phase Synthesis of the Peptides 13 and 14. First, 2.2 g of a
NovaBiochem Rink-Amide MBHA resin (0.62 mmol amine/g) was
suspended in DMF (10 mL) and shaken for 2 h. The resin was shaken
with 20% piperidine in DMF (10 mL) for 15 min and washed three
times with DMF (5 mL). After a positive Kaiser test, the resin was
washed another three times with NMP (5 mL). The Fmoc-protected
amino acid 11 (1.10 g, 1.71 mmol) was dissolved in NMP and shaken
together with HOBt (0.52 g, 3.42 mmol) and TBTU (1.10 g, 3.42 mmol)
for 10 min. This mixture was added to the resin, and the slurry was
shaken after the addition of diisopropylethylamine (DIEA) (1.91 mL,
1
1022w, 739w. H NMR (500 MHz, (CD3)2SO): 0.85 (t, J ) 7.1 Hz,
3 H), 1.28 (m, 4 H), 1.53 (m, 2 H), 2.33 (s, 3 H), 2.38 (s, 3 H), 3.45
(m, 4 H), 3.83 (t, J ) 7.4 Hz, 2 H), 4.08 (t, J ) 6.9 Hz, 2 H), 4.24 (d,
J ) 6.9 Hz, 2 H), 4.45 (s, 2 H), 4.60 (m, 2 H), 4.72 (m, 2 H), 5.36 (s,
2 H), 7.21 (dd, J ) 2.0 Hz, J ) 8.9 Hz, 1 H), 7.26 (t, J ) 7.4 Hz, 2
H), 7.27 (t, J ) 7.4 Hz, 1 H), 7.33 (t, J ) 7.4 Hz, 2 H), 7.39 (t, J )

7328 J. Am. Chem. Soc., Vol. 121, No. 32, 1999
Epple and Carell
7.4 Hz, 2 H), 7.46 (d, J ) 7.0 Hz, 2 H), 7.54 (d, J ) 7.4 Hz, 2 H),
7.73 (s, 1 H), 7.85 (d, J ) 7.4 Hz, 2 H), 7.86 (s, 1 H), 7.94 (s, 1 H),
NMR (125 MHz, (CD3)2SO): δ 13.74, 13.76, 18.64, 20.76, 21.78 (2C),
26.91, 27.08, 28.51, 28.59, 35.54, 35.55, 38.53, 39.00, 40.15, 40.77,
42.82, 42.83, 47.98, 47.99, 54.69, 55.04, 100.75, 108.20, 114.70, 115.91,
130.86, 130.96, 133.84, 134.04, 135.79, 135.80, 136.16, 140.90, 143.80,
146.60, 148.91, 152.32, 152.39, 154.77, 155.87, 155.88, 159.26, 161.73,
163.90, 167.17, 167.32, 168.05, 168.34. HRMS (FAB) for
C49H55N13O11: (M + 1+) calcd 1002.4222, found 1002.07.
Model Compound 3a. BOP (210 mg, 0.452 mmol) was added to a
solution of the cis-syn-biscarboxymethyluracil dimer 6 (70 mg, 0.21
mmol) in DMF (8 mL). This solution was stirred for 10 min at room
temperature. A solution of the deazaflavin peptide 13 (130 mg, 0.21
mmol) dissolved in DMF (7 mL) was then added together with ca. 15
drops of NEt3. The solution was stirred for 90 min at room temperature.
An excess of the aminoethyl-flavin 8 (100 mg, 0.29 mmol) dissolved
in DMF (5 mL) was added, and the solution was stirred for another 90
min at room temperature. The reaction mixture was concentrated to
ca. 15 mL. This solution was subjected to reversed-phase HPLC (RP18-
column with a H2O(1% TFA)/CH3CN gradient from 100% water to
100% acetonitrile over 100 min). The model compound 3a was obtained
as an orange-colored powder (60 mg, 32%). Mp: 232-234 °C. IR
(KBr): 3423s, 3067w, 2967w, 2878w, 1700s, 1646s, 1606s, 1585m,
1537s, 1459m, 1417w, 1372w, 1333w, 1254m, 1228m, 1189w, 1017w.
NMR gave broad, unresolved signals. HRMS (FAB) for
C60H62N16O14: (M + 1+) calcd 1231.4709, found 1231.4707.
Model Compound 2a. BOP (270 mg, 0.581 mmol) was added to a
solution of the cis-syn-biscarboxymethyluracil dimer 6 (90 mg, 0.27
mmol) in DMF (8 mL). This solution was stirred for 10 min at room
temperature. Then a solution of the deazaflavin peptide 14 (220 mg,
0.27 mmol) dissolved in DMF (7 mL) was added together with ca. 20
drops of NEt3. The solution was stirred for 90 min at room temperature.
An excess of the aminoethyl-flavin 8 (130 mg, 0.37 mmol) dissolved
in DMF (5 mL) was added, and the solution was stirred for another 3
h at room temperature. The reaction mixture was concentrated to ca.10
mL. This solution was subjected to reversed-phase HPLC (RP18-column
with a H2O(1% TFA)/CH3CN gradient from 100% water to 100%
acetonitrile over 100 min). The model compound 2a was obtained as
an orange-colored powder (5 mg, 1%). Mp: 266-268 °C. IR (KBr):
3450s, 2967w, 2878w, 1700s, 1629s, 1583s, 1546s, 1455m, 1428w,
1367w, 1336m, 1261w, 1229s, 1056w, 1017w, 922w. NMR gave broad,
unresolved signals. HRMS (FAB) for C70H76N18O16: (M + 1+) calcd
1425.5765, found 1425.5790.
Model Compound 1a. A solution of the flavin-deazaflavin peptide
27 (120 mg, 0.122 mmol) in DMF (6 mL) was stirred after the addition
of a solution of 40% dimethylamine in H2O (2 mL) for 20 min at room
temperature. This solution was evaporated to dryness in vacuo, and
the remaining amine 12 (93 mg, quantitative) was dried under high
vacuum. A solution of the cis-syn-biscarboxymethyluracil dimer 6 (50
mg, 0.15 mmol) and BOP (150 mg, 0.323 mmol) in DMF (6 mL) was
prepared and stirred for 10 min at room temperature. This solution
was added to the solution of the amine 12 (93 mg, 0.122 mmol) in
DMF (5 mL). Ten drops of NEt3 were added, and the reaction mixture
was stirred for 30 min at room temperature. An excess of 1-pentylamine
was added, and the reaction mixture was stirred for another 2 h. The
solution was diluted with CHCl3 (100 mL) and washed three times
with H2O (3 × 100 mL). The organic phase was dried over MgSO4,
filtered, and concentrated in vacuo to dryness. The obtained material
was dissolved in a small amount of CHCl3 and precipitated with Et2O.
The solid material was filtered off and obtained after column chroma-
tography on silica gel-H (d ) 3 cm, l ) 20 cm, with a CHCl3/MeOH
gradient from 10:1 to 7:1) as an orange powder (28 mg, 20%). Rf
(CHCl3/MeOH, 10:1): 0.19. Mp: 227-229 °C. IR (KBr): 3436s,
3078w, 2950w, 2933w, 2856w, 1689s, 1644s, 1604s, 1578m, 1538s,
1463m, 1406w, 1372w, 1247m, 1228m, 1211w, 1183w, 1156w, 1017w.
1H NMR (500 MHz, (CD3)2SO): δ 0.84 (t, J ) 7.1 Hz, 3 H), 0.86 (t,
J ) 7.1 Hz, 3 H), 1.25 (m, 8 H), 1.36 (m, 2 H), 1.54 (m, 2 H), 2.44 (s,
3 H), 2.55 (s, 3 H), 3.00 (q, J ) 6.9 Hz, 2 H), 3.38 (d, J ) 16.5 Hz,
1 H), 3.46 (m, 2 H), 3.47 (d, J ) 16.5 Hz, 1 H), 3.50 (m, 2 H), 3.65
(m, 2 H), 3.84 (t, J ) 7.3 Hz, 2 H), 4.05 (m, 1 H), 4.11 (m, 1 H), 4.13
(d, J ) 16.4 Hz, 1 H), 4.15 (d, J ) 16.4 Hz, 1 H), 4.52 (s, 2 H), 4.63
(m, 1 H), 4.64 (m, 2 H), 4.74 (m, 1 H), 5.40 (s, 2 H), 7.23 (dd, J ) 2.0
Hz, J ) 8.9 Hz, 1 H), 7.31 (m, 3 H), 7.48 (d, J ) 6.9 Hz, 2 H), 7.77
8.12 (d, J ) 8.9 Hz, 1 H), 8.61 (t, J ) 5.8 Hz, 1 H), 8.91 (s, 1 H). 13
C
NMR (125 MHz, (CD3)2SO): δ 13.76, 18.61, 20.67, 21.78, 26.99,
28.55, 35.30, 37.10, 40.07, 42.91, 43.39, 43.68, 46.51, 65.59, 70.25,
99.44, 111.20, 115.18, 115.96, 119.96 (2C), 124.89 (2C), 126.88 (2C),
127.48 (2C), 127.73, 128.08 (2C), 128.16, 128.35 (2C), 130.98, 131.13,
133.51, 134.09, 135.75, 135.99, 136.06, 140.56 (2C), 141.91, 142.53,
143.66 (2C), 146.83, 149.03, 154.31, 155.60, 155.86, 156.41, 159.38,
161.38, 164.46, 168.28. HRMS (FAB) for C56H53N9O8: (MH+) calcd
980.4095, found 980.4092.
Model Compound 4a. BOP (300 mg, 0.646 mmol) was added to a
solution of the cis-syn-biscarboxymethyluracil dimer 6 (100 mg, 0.30
mmol) in DMF (5 mL). This solution was stirred for 10 min at room
temperature. A solution of the aminoethyl flavin 7 (150 mg, 0.32 mmol)
in DMF (3 mL) and ca. 15 drops of NEt3 were added. This solution
was stirred for 30 min at room temperature. An excess of the
aminoethyl-substituted deazaflavin 10 (230 mg, 0.42 mmol) dissolved
in DMF (3 mL) was added, and the reaction mixture was stirred for
another 30 min at room temperature. The reaction mixture was diluted
with CHCl3 (100 mL) and washed three times with H2O (3 × 100
mL). The organic phase was dried with MgSO4, filtered, and concen-
trated in vacuo. The crude product was dissolved in a small amount of
CHCl3 and precipitated with diethyl ether. Flash chromatography on
silica gel-H (d ) 3 cm, l ) 25 cm) with a CHCl3/MeOH gradient from
7.5:1 to 5:1 afforded 4a as an orange-colored powder (53 mg, 16%).
Rf (CHCl3/MeOH/HOAc, 5:1:1): 0.44. Mp: >250 °C. IR (KBr):
3414m, 3064w, 2951w, 2862w, 1700s, 1647s, 1605s, 1584m, 1538s,
1
1462m, 1410w, 1252m, 1228m, 1017w, 798w. H NMR (500 MHz,
(CD3)2SO): δ 0.84 (t, J ) 7.0 Hz, 3 H), 0.86 (t, J ) 7.0 Hz, 3 H),
1.28 (m, 8 H), 1.55 (m, 4 H), 2.37 (s, 3 H), 2.48 (s, 3 H), 3.45 (m, 4
H), 3.51 (d, J ) 16.7 Hz, 1 H), 3.57 (d, J ) 16.7 Hz, 1 H), 3.63 (m,
2 H), 3.85 (m, 4 H), 4.13 (d, J ) 16.5 Hz, 1 H), 4.14 (m, 2 H), 4.19
(d, J ) 16.5 Hz, 1 H), 4.58 (m, 2 H), 4.69 (m, 2 H), 5.44 (dd, J ) 12.2
Hz, J ) 16.4 Hz, 2 H), 7.21 (dd, J ) 2.0 Hz, J ) 8.9 Hz, 1 H), 7.37
(m, 1 H), 7.44 (m, 2 H), 7.54 (m, 2 H), 7.80 (s, 1 H), 7.89 (s, 1 H),
7.91 (s, 1 H), 8.11 (d, J ) 8.9 Hz, 1 H), 8.34 (t, J ) 6.0 Hz, 1 H), 8.59
(t, J ) 5.8 Hz, 1 H), 8.91 (s, 1 H), 10.46 (s, 1 H), 10.48 (s, 1 H). 13C
NMR (125 MHz, (CD3)2SO) δ 13.72, 13.75, 18.62, 20.72, 21.76, 21.77,
26.89, 27.00, 28.51, 28.56, 35.38, 35.59, 38.80 (2C), 40.10, 40.78,
42.76, 42.77, 48.04, 48.19, 54.84, 55.06, 70.28, 99.56, 111.28, 115.12,
115.89, 115.99, 127.88 (2C), 128.15, 128.52 (2C), 130.79, 130.97,
133.58, 134.02, 135.79, 135.81, 136.10, 141.90, 142.39, 146.61, 148.88,
152.35, 152.42, 154.74, 155.67, 155.96, 159.20, 161.35, 164.40, 167.23,
167.25, 168.36, 168.79. HRMS (FAB) for C56H61N13O11: (M + 1+)
calcd 1092.4691, found 1092.4690.
General Method for the Debenzylation of the Model Compounds
(Synthesis of 1b, 2b, 3b). The model compound (either 1a, 2a, or 3a)
was dissolved in acetic acid at room temperature (5 mL). A suspension
of 10% Pd/BaSO4 catalyst (5 mg) in acetic acid (0.5 mL) was added,
and the reaction mixture was stirred for 30 h in an H2 atmosphere (1
bar). The mixture was filtered through Celite, and the filtrate was
evaporated in vacuo to dryness. The obtained material was investigated
by fluorescence and UV-visible spectroscopy. The purity was deter-
mined by analytical reversed-phase HPLC (RP18-column with a
H2O(1% TFA)/CH3CN gradient, 100% water to 100% acetonitrile over
100 min).
For the synthesis of 4b, compound 4a (20 mg, 0.018 mmol) was
debenzylated, and the model compound 4b was isolated by preparative
reversed-phase HPLC (RP18-column with a H2O(1% TFA)/CH3CN
gradient, 100% water to 100% acetonitrile over 100 min) as a green-
colored powder (18 mg, 99%). Mp: 253-255 °C. IR (KBr): 3436s,
2952w, 2922w, 2856w, 1700m, 1628m, 1600m, 1578m, 1546s, 1467m,
1
1267m, 1222w, 1201w, 1182w, 1156w, 1028w. H NMR (500 MHz,
(CD3)2SO): δ 0.85 (t, J ) 7.0 Hz, 3 H), 0.86 (t, J ) 7.0 Hz, 3 H),
1.30 (m, 8 H), 1.55 (m, 4 H), 2.39 (s, 3 H), 2.49 (s, 3 H), 3.42 (d, J )
16.8 Hz, 1 H), 3.45 (m, 4 H), 3.49 (d, J ) 16.8 Hz, 1 H), 3.63 (m, 2
H), 3.85 (m, 4 H), 4.09 (m, 2 H), 4.13 (d, J ) 16.6 Hz, 1 H), 4.15 (d,
J ) 16.6 Hz, 1 H), 4.60 (m, 4 H), 6.83 (m, 1 H), 7.03 (m, 1 H), 7.88
(m, 1 H), 7.92 (s, 1 H), 7.93 (s, 1 H), 8.40 (t, J ) 5.8 Hz, 1 H), 8.44
(t, J ) 5.6 Hz, 1 H), 8.69 (s, 1 H), 10.44 (s, 1 H), 10.45 (s, 1 H). 13C

Efficient Light-Dependent DNA Repair
J. Am. Chem. Soc., Vol. 121, No. 32, 1999 7329
(s, 1 H), 7.92 (t, J ) 5.6 Hz, 1 H), 7.97 (s, 1 H), 8.00 (s, 1 H), 8.14 (d,
J ) 8.9 Hz, 1 H), 8.34 (t, J ) 5.8 Hz, 1 H), 8.61 (t, J ) 5.8 Hz, 1 H),
8.93 (s, 1 H), 10.39 (s, 1 H), 10.43 (s, 1 H). 13C NMR (125 MHz,
(CD3)2SO): δ 13.76, 13.78, 18.70, 20.82, 21.70, 21.77, 26.99, 28.43,
28.55, 28.66, 35.30, 35.69, 38.27, 38.40, 38.93, 40.08, 42.92, 43.04,
43.78, 48.04, 48.16, 54.55, 55.17, 70.29, 99.49, 111.23, 115.20, 115.98,
116.06, 128.12 (2C), 128.25, 128.37 (2C), 130.74, 131.09, 133.55,
134.06, 135.76, 136.13, 136.24, 141.95, 142.56, 147.11, 149.09, 152.31,
152.34, 154.53, 155.63, 155.90, 159.36, 161.40, 164.49, 166.87, 167.13,
167.44, 168.31, 168.38. HRMS (FAB) for C58H64N14O12: (M + 1+)
calcd 1149.4906, found 1149.4899.
matic light beam at the specified wavelengths. During an experiment,
6 × 50-µL aliquots of the assay solution were removed, immediately
reoxidized by the addition of oxygen, and analyzed by reversed-phase
HPLC (Nucleogel C18-column, water/methanol gradient). Integration
of the product and the starting material peaks yielded the reaction rate
per minute. The number of absorbed light quantums per minute was
determined by ferrioxalate actinometry. Both values were used to
calculate the quantum yield Φ () number of reacted molecules/number
of absorbed photons). Φ is a measure of the efficiency of the overall
light energy conversion.
Splitting Assay. For the repair assay, all model compounds were
dissolved in ethylene glycol at 10-6 M concentration, and the solution
was vigorously degassed. Addition of a sodium dithionite solution (0.05
M in water) and 10 µL of NEt3 converted the flavin moiety in all model
compounds into the reduced and deprotonated form. The cuvettes were
placed in a fluorescence spectrometer and irradiated with a monochro-
Acknowledgment. This work was supported by the Swiss
National Science Foundation. We thank Prof. F. Diederich for
his very generous support.
JA990466L