HiroshiSugiyamaDepartmentofChemistry,GraduateSchoolofScienceInstituteforIntegratedCell-MaterialSciences(iCeMS)

KyotoUniversity 1

100nm

生体分子機能論 2018-7水素引き抜き光反応

Advanced Course in Molecular Biology and Biochemistry

Basics 1 Basic elements of nucleic acids and their synthesis 2 Sequencing of DNA 3 3D structure of DNA 1 4 3D structure of DNA 2 Chemistry 5) DNA alkylation 6) Hydrogen abstraction 1 7) Hydrogen abstraction 2 8) Charge transfer Biology 9) Epigenetics 1 10) Epigenetics 2 11) ATRX

Uracil radical could read various DNA conformations

A

O

H

P-O

O

O

OpGC

N

NH

O O

O

OdGCpO

OpUGC

OdGCpOO

A

O

H

P-O

O

O

OpGC

N

NH

O O

O

OdGCpO

DCD3CDOH

D

H

OpUGC

CHO

OHdGCpO

Intrastrand Hydrogen Abstraction

(D content 93%).d(GCABrUGC)2302 nm

(100 mM)

1' ox 2' ox

+

A

The DFT-determined geometry for the transition state of the above reaction. Distances are listed in angstroms. Note the elongated C-Br and C-O bonds, and the trigonal bipyramidal structure. �

•  The transition state of a chemical reaction is defined as the state corresponding to the highest potential energy along this reaction coordinate. At this point, assuming a perfectly irreversible reaction, colliding reactant molecules always go on to form products. It is often marked with the double dagger ‡ symbol.

•  As an example, the transition state shown below occurs during the SN2 reaction of bromoethane with a hydroxyl anion:

�

177.7° 176.3°

1.2495Å 1.4947Å 1.3068Å 1.3874Å

1'-H 2'-H B3LYP / 6-311G**

2.7436Å 2.6927Å

D: 4.2 kcal/molH: 3.2 kcal/mol

D: -16.9 kcal/molH: -17.2 kcal/mol

D: 7.7 kcal/molH: 6.8 kcal/mol

H,D: -8.8 kcal/mol

kH/kD = 5.5 kH/kD = 5.7

����

��

H�������

D�������

GH GD

NH

H

OH

O AO

ONH

OP

O

OOOO

NH

H

OH

OH

AO

ONH

OP

O

OOOO

NH

H

O

OH

AO

ONH

OP

O

OOOO

H H

Ab initio Calculation of H Abstraction of Model System

Energy (au) MP2/6-31G*//UHF/6-31G*

+

ZPE (kcal/mol) UHF/6-31G*//UHF/6-31G*

H

HO H

H

H

HH

HHH

HO H

H

H

HH

HH

H

HH

HH

HO HH

H

8.5 kcal

0 kcal

α Hydrogen = 1' model

-21.3 kcal

H

HH

HH

OHH

H

H H

HH

HH

OHH

H

HH

HH

HH

OH

H

H H

β Hydrogen = 2' model

11.7 kcal

0 kcal-13.9 kcal

A

O

H

P-O

O

O

OpGC

N

NH

O O

O

OdGCpO

OpUGC

OdGCpOOH

OpUGCCHO

OHdGCpO

A

O

H

P-O

O

O

OpGC

N

NH

O O

O

OdGCpO

D

D

H

OpBrUGC

OdGCpO

A

D

H

OpBrUGC

OdGCpOA

H

D

0

7.2 (IU)

0.85%

+

40

+.

2' OX

1' OX 2' OX

.2'βD

6.4%

2'αD

2'αD

2'OX

1'OX

0

43%

adenineadenine

kH/kD = 7.5 (BrU)

1' OX38%

2'OX

20 20

2'αD

200

adenine

Retention Time (min)

2'βD 1'OX

1'OX

d(GCABrUGC)2'H

Kinetic Isotope Effect

4040

2'βD

DNA 1'�

DNA 2'α�

Hybrid 1'�

DNA duplex�DNA-RNA Hybrid �

Angew Chem Int Ed 45, 1354 (2006) Nature Protocol 2, 78 (2007)�

Z-DNA �

N

N

N

NH

NH2

O

CH3HO N

HO

N

N

N

O

H2N

OH

HOOO

H3C

anti syn

8-Methylguanine

Nucleic Acids Res., (1996) 24,1272-1278

NaCl (mM)

d(CGCGCG)2 2600

Oligonucleotides

d(CGCm8GCG)2 30d(CGCATGCG)2

d(Cm8GCATm8GCG)2

no Z-DNA observed

45

Midpoint NaCl Concentration in B-Z transition

m8G=8-Methyl deoxyguanosine

H

d(C G C G IU G C G)d(GCm8GCACm8GC) 800

CH3HO N

HO

N

N

N

O

H2N

O

OH

m8rG=8-Methyl guanosined(C G C G IU G C G)d(GCm8rGCACm8rGC) 0

JACS., (2003) 125,13519-13524

-10000

0

10000

210 230 250 270 290 310 330 350

CD spectrum of

Iodouracil-ContainingOligonucleotides

Wavelength (nm)

d(CGCGIUGCG)d(GCGCACGC)

d(C G C G IU G C G)d(GCm8GCACm8GC)

B-form

Z-form

2 °C

θ

NaCl 2 M

OGdCGCpO

O OP

O O-

UO

OpGCG

HO

Ribonuclease T1

+

O G

IU

d(CGCpO

O

OpGCG)

HO G

U

dCGCpO

O

OpGCG

H O2

O G

U

dCGCpO

O

OpGCG

OH.hv 302nm2'βP P P

2'α

3'-d(GC8mGCAC8mGC)-5'5'-d(C G C G IU G C G)-3'Z-DNA

Kawai, K. et al. Tetrahedron Lett. 1999.Kawai, K. et al. J. Am. Chem. Soc. 1999.Oyoshi T. et al. J. Am Chem. Soc. 2003.�

Tyr177

Arg174

Thr191

Tyr177

Arg174

Thr191

2'β

2'α

1'

0

50

100

Z %

0 0.5 1 1.5

UGCG (µM)Z %

UGCG

(µ

M)

5

10

15GXU

rG> U

rGU

Na+ (M)

*

Formation of UGCG Directly

hν

Correlated with Z %

B-form*

Rnase T1

B-form

*

Gel Electrophoresis

Z-form X=Br or I

+

Bending DNA �

The Structures of DNA and Sso7d-DNA Complex

B-DNA Sso7d-DNA Complex

5'-GTAATTAC-3'3'-CATTAATG-5'

Wang, A. H.-J.et al., Nature Structural Biology 1998.

OpUAC

dGTAApO O

NHN

O

O

HOH2C

OpUAC

dGTAApO O

NHN

O

O

OHC

O

dGTAApO O

NHN

O

O

H3C

PO-

OO O

NHN

O

O

OpAC

NHN

O

O

H2C NHN

O

O

HOOH2C

NHN

O

O

H2C

Mechanism of Photoreaction of 5-Iodouracil-Containing DNA with Sso7d

O2 , H

.

.

+H2O

e-

.

Met29

T5-Me

U

5'-GTAATIUAC-3'3'-CAIUTAATG-5'

Intrastrand H abstraction at the methyl group of T5Oyoshi, T.; Wang, A. H.-J.; Sugiyama, H. J. Am. Chem. Soc. 2002.

G-quadruplex

syn

anti

ATT Diagonal loop

Lateral loop

5' A

G

G

G

G

G

G

G

G

G

G

G

G

3'

Na+ Form

Structure 356, 164-168 (1992)

GGG is antiparallel

anti

TA

T

G

G

G

G

G

G

5' A

External loop

3'

G

G

G

G

G

G

K+ Form

Nature 417, 876-880 (2002)

X-ray

GGG is parallel

Bioorg. Med. Chem. In press

K+ Form Solution

Mixed-Chair

HPLC analysis

ODN 1: 5'-AGGGIUTAGGGTTAGGGTTAGGG-3'ODN 2: 5'----------TIUA----------------------------------3'ODN 3: 5'-----------------------IUTA---------------------3'ODN 4: 5'-----------------------TIUA---------------------3'ODN 5: 5'------------------------------------TIUA--------3'ODN 6: 5'------------------------------------IUTA--------3'

K+ Form

Na+ Form Light

Light

Consumption of ODN 1-6

IU is extremely reactive at diagonal loop of antiparallel form! ODN1 ODN2 ODN3 ODN4 ODN5 ODN6 ODN1 ODN2 ODN3 ODN4 ODN5 ODN6

K+ Form

IUAT

0

10

20

30

40

50

60

K+ FormNa+ Form

Na+ Form

5’3’

5’

3’

Structure

Sequence

0 (1%) 0 (2%) 95 (60%)

[d(G4TTTIUG4)]2,[d(TIUG4T)]4d(GTGCTIUACG)GCAGCA-3'A

T CGTCGT-5'IUd(AG3T2AG3TIUAG3T2AG3)/d(TC3A2TC3A2TC3A2TC3)

d(AG3T2AG3TIUAG3T2AG3)

0 (2%) 0 (1%) 90 (50%)2'-deoxyribonolactone in Na+ ions (%)

2'-deoxyribonolactone in K+ ions (%)

0 (1%) 0 (1%) 0 (2%)0 (1%) 0 (2%) 90 (51%)

[d(G4TTIUTG4)]2

89 (35%)

90 (35%)

Diagonal Loop-Specific Ribonolactone Formation

Two Quadruplex Structures are Proposed !

5'-AGGGGAGCTGGGGTAGGTGGGA-3'

IgG switch regions:

ODN 7

2010 300

G

ODN 7

retention time (min)

ribonolactone

A

3'A

5'A

TT

AT

C

GG

G

G

G

G

G

GGG

G

GG

G

3'C

5'CT

C

TTT

G

G

G

G

G

GG

G G

GG

5'-CGGGGGGTTTTGGGCGGC-3'

Rb gene:

ODN 8

T

2010 300

ODN 8

retention time (min)

ribonolactone

Photoreactivities of 5-Bromouracil-containing RNAs

Hironobu Morinaga a, Seiichiro Kizaki a, Tomohiro Takenaka a, Shuhei Kanesato a, Yuta Sannohe a,Ryu Tashiro b, Hiroshi Sugiyama a,c,⇑a Department of Chemistry, Graduate School of Science, Kyoto University, Sakyo, Kyoto 606-8501, Japanb Suzuka University of Medical Science, Faculty of Pharmaceutical Sciences, 3500-3, Minamitamagaki, Suzuka, Mie, Japanc Institute for Integrated Cell-Material Sciences (iCeMS), Kyoto University, Sakyo, Kyoto 606-8502, Japan

a r t i c l e i n f o

Article history:Received 6 October 2012Revised 10 November 2012Accepted 10 November 2012Available online 24 November 2012

Keywords:5-BromouracilElectron transferPhotoreactionRNA

a b s t r a c t

5-Bromouracil (BrU) was incorporated into three types of synthetic RNA and the products of the photoir-radiated BrU-containing RNAs were investigated using HPLC and MS analysis. The photoirradiation ofr(GCABrUGC)2 and r(CGAABrUUGC)/r(GCAAUUCG) in A-form RNA produced the corresponding 20-ketoadenosine (ketoA) product at the 50-neighboring nucleotide, such as r(GCketoAUGC) and r(CGAketoAUUGC),respectively. The photoirradiation of r(CGCGBrUGCG)/r(CmGCACmGCG) in Z-form RNA produced the 20-keto guanosine (ketoG) product r(CGCketoGUGCG), whereas almost no products were observed from thephotoirradiation of r(CGCGBrUGCG)/r(CmGCACmGCG) in A-form RNA. The present results indicate clearlythat hydrogen (H) abstraction by the photochemically generated uracil-5-yl radical selectively occurs atthe C20 position to provide a 20-keto RNA product.

! 2012 Elsevier Ltd. All rights reserved.

1. Introduction

5-Bromouracil (BrU) is a photoreactive base that can be incorpo-rated into DNA or RNA instead of thymine or uracil, respectively.1–4

BrU-substituted DNA and RNA remain functional in vivo.4 The phot-oirradiation of BrU-containing DNA produces the uracil-5-yl radi-cal, which abstracts hydrogen from the 50-neighboring nucleotidein a conformation-specific manner. For example, the photoirradia-tion of BrU-containing DNA in the B-form produces a deoxyribono-lactone-containing product as a C10 oxidation product and anerythrose-containing product as a C20 oxidation product, whereasthe photoirradiation of BrU-containing DNA in the Z-form producesa guanosine as a C20 oxidation product.5 Based on these reactivities,we have proposed that these conformation-specific products ofphotoirradiated BrU-containing DNA can be used for the determi-nation of DNA local structures in living cells.6 The determinationof DNA local structures might be useful for the elucidation of theirroles in living cells.7–15 Similar to DNA local structures, RNA localstructures are also believed to play important biological roles inliving cells. Although some methods, including hydroxyl radicalfootprinting and bioinformatics prediction, have provided signifi-cant information on in vivo RNA local structures,16–20 their natureremains elusive. A new method that allows 1 base-pair resolutionis required for a more detailed understanding of these structures.Hence, the investigation of the photoreactivity of BrU-containing

RNA is important for the BrU-based determination of RNA localstructures.

Moreover, studies on the photoreactivity of BrU-containing DNAor RNA will also provide important information regarding themechanisms underlying radical-induced DNA or RNA damage.Determining the reactive intermediates generated during thedamage process is difficult because the reactive radicals formedinitially are generated randomly within the biopolymers and havea very short half-life. The independent generation of radical inter-mediates within nucleic acids may help solve this problem and elu-cidate the chemistry of radical-induced nucleic acids damage.Information regarding the defined RNA sequence and structure ofthe products of the uracil-5-yl radical provides insight into themechanism underlying radical-induced RNA damage. From thisviewpoint, Greenberg and co-workers generated uracil-5-yl and -6-yl radicals induced using a Norrish type I photocleavage reactionand elucidated the mechanism of RNA strand cleavage initiated bya C20 radical.21,22 As we showed previously that the uracil-5-yl rad-ical can be generated from BrU in DNA under irradiation conditions,here we carried out the photoirradiation of BrU-containing RNA andinvestigated its degradation products.

2. Materials and methods

2.1. Preparation of oligonucleotides

Phosphoramidites were purchased from Glen Research orProligo. Oligonucleotide strands were synthesized on an ABI DNAsynthesizer (Applied Biosystem, Foster City, CA). After purification

0968-0896/$ - see front matter ! 2012 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.bmc.2012.11.010

⇑ Corresponding authorE-mail address: hs@kuchem.kyoto-u.ac.jp (H. Sugiyama).

Bioorganic & Medicinal Chemistry 21 (2013) 466–469

Contents lists available at SciVerse ScienceDirect

Bioorganic & Medicinal Chemistry

journal homepage: www.elsevier .com/locate /bmc

Photoreactivities of 5-Bromouracil-containing RNAs

Hironobu Morinaga a, Seiichiro Kizaki a, Tomohiro Takenaka a, Shuhei Kanesato a, Yuta Sannohe a,Ryu Tashiro b, Hiroshi Sugiyama a,c,⇑a Department of Chemistry, Graduate School of Science, Kyoto University, Sakyo, Kyoto 606-8501, Japanb Suzuka University of Medical Science, Faculty of Pharmaceutical Sciences, 3500-3, Minamitamagaki, Suzuka, Mie, Japanc Institute for Integrated Cell-Material Sciences (iCeMS), Kyoto University, Sakyo, Kyoto 606-8502, Japan

a r t i c l e i n f o

Article history:Received 6 October 2012Revised 10 November 2012Accepted 10 November 2012Available online 24 November 2012

Keywords:5-BromouracilElectron transferPhotoreactionRNA

a b s t r a c t

5-Bromouracil (BrU) was incorporated into three types of synthetic RNA and the products of the photoir-radiated BrU-containing RNAs were investigated using HPLC and MS analysis. The photoirradiation ofr(GCABrUGC)2 and r(CGAABrUUGC)/r(GCAAUUCG) in A-form RNA produced the corresponding 20-ketoadenosine (ketoA) product at the 50-neighboring nucleotide, such as r(GCketoAUGC) and r(CGAketoAUUGC),respectively. The photoirradiation of r(CGCGBrUGCG)/r(CmGCACmGCG) in Z-form RNA produced the 20-keto guanosine (ketoG) product r(CGCketoGUGCG), whereas almost no products were observed from thephotoirradiation of r(CGCGBrUGCG)/r(CmGCACmGCG) in A-form RNA. The present results indicate clearlythat hydrogen (H) abstraction by the photochemically generated uracil-5-yl radical selectively occurs atthe C20 position to provide a 20-keto RNA product.

! 2012 Elsevier Ltd. All rights reserved.

1. Introduction

5-Bromouracil (BrU) is a photoreactive base that can be incorpo-rated into DNA or RNA instead of thymine or uracil, respectively.1–4

BrU-substituted DNA and RNA remain functional in vivo.4 The phot-oirradiation of BrU-containing DNA produces the uracil-5-yl radi-cal, which abstracts hydrogen from the 50-neighboring nucleotidein a conformation-specific manner. For example, the photoirradia-tion of BrU-containing DNA in the B-form produces a deoxyribono-lactone-containing product as a C10 oxidation product and anerythrose-containing product as a C20 oxidation product, whereasthe photoirradiation of BrU-containing DNA in the Z-form producesa guanosine as a C20 oxidation product.5 Based on these reactivities,we have proposed that these conformation-specific products ofphotoirradiated BrU-containing DNA can be used for the determi-nation of DNA local structures in living cells.6 The determinationof DNA local structures might be useful for the elucidation of theirroles in living cells.7–15 Similar to DNA local structures, RNA localstructures are also believed to play important biological roles inliving cells. Although some methods, including hydroxyl radicalfootprinting and bioinformatics prediction, have provided signifi-cant information on in vivo RNA local structures,16–20 their natureremains elusive. A new method that allows 1 base-pair resolutionis required for a more detailed understanding of these structures.Hence, the investigation of the photoreactivity of BrU-containing

RNA is important for the BrU-based determination of RNA localstructures.

Moreover, studies on the photoreactivity of BrU-containing DNAor RNA will also provide important information regarding themechanisms underlying radical-induced DNA or RNA damage.Determining the reactive intermediates generated during thedamage process is difficult because the reactive radicals formedinitially are generated randomly within the biopolymers and havea very short half-life. The independent generation of radical inter-mediates within nucleic acids may help solve this problem and elu-cidate the chemistry of radical-induced nucleic acids damage.Information regarding the defined RNA sequence and structure ofthe products of the uracil-5-yl radical provides insight into themechanism underlying radical-induced RNA damage. From thisviewpoint, Greenberg and co-workers generated uracil-5-yl and -6-yl radicals induced using a Norrish type I photocleavage reactionand elucidated the mechanism of RNA strand cleavage initiated bya C20 radical.21,22 As we showed previously that the uracil-5-yl rad-ical can be generated from BrU in DNA under irradiation conditions,here we carried out the photoirradiation of BrU-containing RNA andinvestigated its degradation products.

2. Materials and methods

2.1. Preparation of oligonucleotides

Phosphoramidites were purchased from Glen Research orProligo. Oligonucleotide strands were synthesized on an ABI DNAsynthesizer (Applied Biosystem, Foster City, CA). After purification

0968-0896/$ - see front matter ! 2012 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.bmc.2012.11.010

⇑ Corresponding authorE-mail address: hs@kuchem.kyoto-u.ac.jp (H. Sugiyama).

Bioorganic & Medicinal Chemistry 21 (2013) 466–469

Contents lists available at SciVerse ScienceDirect

Bioorganic & Medicinal Chemistry

journal homepage: www.elsevier .com/locate /bmc

by HPLC, products were confirmed by ESI-TOFMS (Table S1). DNAconcentrations were determined by using the Nano drop ND-1000 (Nano-drop Technologies, Wilmington, DE).

2.2. Photoreaction and HPLC analysis

The reaction mixture contains RNAs in 20 mM sodium cacodyl-ate buffer (pH 7.0) with or without NaCl (0.1–3 M). After irradia-tion by a transilluminator at 302 nm with 6.0 mJ/cm2 on ice, thereaction mixtures were analyzed by HPLC. HPLC analysis was car-ried out by the PU-980 HPLC system (Jasco) with a Chemcobond 5-ODS-H column. Detection was carried out at 254 nm. Elution waswith 0.05 M ammonium formate containing 0–5% acetonitrile, lin-ear gradient (50 min) (Fig. 1, 3 and 4, Figs. S2 and S4), 0–4% aceto-nitrile, linear gradient (40 min) (Figs. S1 and S3) or 0–2%acetonitrile, linear gradient (20 min) (Fig. S6).

2.3. Enzymatic digestion and identification of keto adenosineand keto guanosine

Using 5 units of antarctic phosphatase (New England Biolabs)and 0.25 units of nuclease P1 (WAKO Pure chemical), the product1, 3 and 4 were digested to mono nucleosides at 37 !C for 4 h. Theythen were analyzed by HPLC comparing the retention time withauthentic material of A, C, G, U, ketoA, and ketoG.

2.4. Synthesis of keto adenosine and keto guanosine

In order to confirm the generation of keto adenosine and ketoguanosine after photoirradiation, they were synthesized accordingto the previous method with a slight modification.27–30

Pfitzner–Moffatt oxidation procedure was used instead of usingDess–Martin periodinane reagent. 1H NMR spectra were recordedon a JEOL JNM ECA-600 spectrometer (600 MHz for 1H), withchemical shifts reported in parts per million relative to residual

solvent and coupling constants in hertz. The following abbrevia-tions were applied to spin multiplicity: s (singlet), d (doublet), t(triplet), q (quartet), and m (multiplet).

2.5. 20-Keto adenosine

1H NMR (600 MHz, DMSO-d6): d 8.18 (s, 1H, H2), 8.13 (s, 1H,H8), 7.19 (s, 2H, NH2), 6.30 (s, 1H, 20-OH), 6.05 (s, 1H, 20-OH),5.91 (s, 1H, H10), 5.45 (d, J = 6.12 Hz, 1H, 30-OH), 5.05 (t,J = 5.44 Hz, 1H, 50-OH), 4.08 (t, J = 6.46 Hz, 1H, H40), 3.70–3.73 (m,1H, H30), 3.68–3.69 (m, 1H, H50), 3.59–3.62 (m, 1H, H50). 1H NMRspectrum is shown in Figure S7.

Figure 1. Photoproducts of r(GCABrUGC). (a) HPLC analysis carried out afterphotoirradiation of the r(GCABrUGC)2 sequence for 2 h showed two major photo-products. (b) The photoproducts were a 20-keto adenosine-containing product (1)and a reduced form (2). The yield was 9.7 ± 0.8% in (1) and 10.2 ± 4.9% in (2).

Figure 2. Comparison of the photoreactivity of BrU in RNA and DNA. The products ofthe photoirradiation of r(GCABrUGC)2 and d(GCABrUGC)2 were analyzed using HPLC,and the consumption of starting hexamers was measured via HPLC. The values arethe means of four independent experiments and SD was shown.

Figure 3. Photoproducts of r(CGAABrUUGC)/r(GCAAUUCG). (a) HPLC analysiscarried out after photoirradiation of r(CGAABrUUGC)/r(GCAAUUCG) for 10 minshowed the presence of three photoproducts. (b) The photoproducts were uracildimer-(3) and 20-keto adenosine-(4) containing products and a reduced form (5).The yield was 69.5 ± 2.4% in (3), 14.2 ± 3.3 in (4) and 16.3 ± 5.7% in (5).

H. Morinaga et al. / Bioorg. Med. Chem. 21 (2013) 466–469 467

2.6. 20-Keto guanosine

1H NMR (600 MHz, DMSO-d6): d 10.55 (s, 1H, H1), 7.76 (s, 1H,H8), 6.43 (s, 2H, NH2), 6.28 (s, 1H, 20-OH), 5.94 (s, 1H, 20-OH),5.69 (s, 1H, H10), 5. 43 (d, J = 5.44 Hz, 1H, H30), 4.95 (t, J = 5.44 Hz,1H, 50-OH), 3.96 (t, J = 5.78 Hz, 1H, H30), 3.64–3.65 (m, 2H, H40

and H50), 3.54–3.56 (m, 1H, H50). 1H NMR spectrum is shown inFigure S8.

3. Results

3.1. Products of photoirradiated r(GCABrUGC)2

First, we investigated the photoirradiation (302 nm) of the self-complementary r(GCABrUGC) sequence. HPLC analysis of the irradi-ated r(GCABrUGC)2 indicated the formation of two major products,1 and 2 as shown in Figure 1. Photoproduct 2 was a reduced prod-uct r(GCAUGC) compared with the authentic material. The product1 was identified as a 20-keto adenosine (ketoA)-containing product,r(GCketoAUGC), as a C20 oxidation product using ESI-TOF MS spec-troscopy and enzymatic digestion to monomers, which were com-pared to the authentic materials. The authentic material of ketoAwas synthesized via the oxidation of a protected adenosine(Fig. S1). Almost no C10-oxidation products were observed afterthe photoirradiation of r(GCABrUGC)2. This was in clear contrastto the observation that competitive C10 and C20 hydrogenabstraction by the uracil-5-yl radical occurs in B-form DNA.1 Based

on these results, we propose the following mechanism for the for-mation of 1 (Scheme 1).

Under irradiation conditions, an initial electron transfer occursfrom photoexcited G to BrU, to produce an anion radical of BrU. Re-lease of the bromide ion generates a uracil-5-yl radical that ab-stracts the C20 hydrogen of the 50-neighboring adenosine, toproduce the C20 radical. The oxidation of the C20 radical by the Gcation radical leads to the production of a C20 cation, providing ke-

toA, which is mainly present as a hydrated form in aqueous solu-tion. In RNA, C20 radical was oxidized by G cation radical andreacted with water but not attacked by molecular oxygen. If molec-ular oxygen attacked C20 radical, oxidation products of comple-mentary strand would have occurred. It was also supported bythe result that there was no difference in the photoreactivity be-tween aerobic and anaerobic condition.

This oxidation mechanism of the C20 radical in photoirradiatedBrU-containing RNA is different from that observed for photoirradi-ated BrU-containing DNA, in which erythrose is generated as a C20

oxidation product via the reaction between the C20 radical andmolecular oxygen. The difference between RNA and DNA in the reac-tion after the generation of the C20 radical can be explained by theexistence of the 20-OH group in RNA, which reduces the oxidationpotential of the C20 radical compared with that observed for DNA.

To compare the reactivity of BrU in photoirradiated RNA andDNA, the consumption of r(GCABrUGC)2 and d(GCABrUGC)2 underirradiation conditions was measured. The results showed that thephotoreactivity of BrU in this sequence in RNA was approximatelytwofold slower than that measured in DNA (Fig. 2). The differencein photoreactivity between RNA and DNA is presumably due to thehalf-life of the anion radical of BrU that produces the uracil-5-ylradical. Obtaining additional general information regarding the dif-ference in photoreactivity between RNA and DNA will require theinvestigation of photoreactivity in a wide variety of sequences.

3.2. Products of photoirradiated r(CGAABrUUGC)/r(GCAAUUCG)

Next, we investigated the products of photoirradiated r(CGAAB-

rUUGC)/r(GCAAUUCG). In DNA, the analogous sequence 50-(G/C)AABrUT-30 has been identified as a hotspot sequence in whichthe efficient photoreaction of BrU occurs to provide alkaline labilesites.23,24

It has been proposed that the efficient photoreaction in this se-quence is caused by the good alignment of three factors: G as theelectron donor, the stacking BrUT as the electron acceptor, and A/Tas the bridge between the electron donor and the acceptor. The HPLCanalysis of photoirradiated r(CGAABrUUGC)/r(GCAAUUCG) showed

Scheme 1.

Figure 4. Photoproducts of r(CGCGBrUGCG)/r(CmGCACmGCG). Photoreaction wasperformed for 2 h in the presence of 0.1 M NaCl (A-form RNA) (a) or 3 M NaCl (Z-form RNA) (b) and the products were analyzed using HPLC. Photoirradiation of thissequence produced one major photoproduct only in the Z-form RNA condition, a 20-keto guanosine-containing product (c). The yield was 54.0 ± 0.1% in (6).

468 H. Morinaga et al. / Bioorg. Med. Chem. 21 (2013) 466–4692.6. 20-Keto guanosine

1H NMR (600 MHz, DMSO-d6): d 10.55 (s, 1H, H1), 7.76 (s, 1H,H8), 6.43 (s, 2H, NH2), 6.28 (s, 1H, 20-OH), 5.94 (s, 1H, 20-OH),5.69 (s, 1H, H10), 5. 43 (d, J = 5.44 Hz, 1H, H30), 4.95 (t, J = 5.44 Hz,1H, 50-OH), 3.96 (t, J = 5.78 Hz, 1H, H30), 3.64–3.65 (m, 2H, H40

and H50), 3.54–3.56 (m, 1H, H50). 1H NMR spectrum is shown inFigure S8.

3. Results

3.1. Products of photoirradiated r(GCABrUGC)2

First, we investigated the photoirradiation (302 nm) of the self-complementary r(GCABrUGC) sequence. HPLC analysis of the irradi-ated r(GCABrUGC)2 indicated the formation of two major products,1 and 2 as shown in Figure 1. Photoproduct 2 was a reduced prod-uct r(GCAUGC) compared with the authentic material. The product1 was identified as a 20-keto adenosine (ketoA)-containing product,r(GCketoAUGC), as a C20 oxidation product using ESI-TOF MS spec-troscopy and enzymatic digestion to monomers, which were com-pared to the authentic materials. The authentic material of ketoAwas synthesized via the oxidation of a protected adenosine(Fig. S1). Almost no C10-oxidation products were observed afterthe photoirradiation of r(GCABrUGC)2. This was in clear contrastto the observation that competitive C10 and C20 hydrogenabstraction by the uracil-5-yl radical occurs in B-form DNA.1 Based

on these results, we propose the following mechanism for the for-mation of 1 (Scheme 1).

Under irradiation conditions, an initial electron transfer occursfrom photoexcited G to BrU, to produce an anion radical of BrU. Re-lease of the bromide ion generates a uracil-5-yl radical that ab-stracts the C20 hydrogen of the 50-neighboring adenosine, toproduce the C20 radical. The oxidation of the C20 radical by the Gcation radical leads to the production of a C20 cation, providing ke-

toA, which is mainly present as a hydrated form in aqueous solu-tion. In RNA, C20 radical was oxidized by G cation radical andreacted with water but not attacked by molecular oxygen. If molec-ular oxygen attacked C20 radical, oxidation products of comple-mentary strand would have occurred. It was also supported bythe result that there was no difference in the photoreactivity be-tween aerobic and anaerobic condition.

This oxidation mechanism of the C20 radical in photoirradiatedBrU-containing RNA is different from that observed for photoirradi-ated BrU-containing DNA, in which erythrose is generated as a C20

oxidation product via the reaction between the C20 radical andmolecular oxygen. The difference between RNA and DNA in the reac-tion after the generation of the C20 radical can be explained by theexistence of the 20-OH group in RNA, which reduces the oxidationpotential of the C20 radical compared with that observed for DNA.

To compare the reactivity of BrU in photoirradiated RNA andDNA, the consumption of r(GCABrUGC)2 and d(GCABrUGC)2 underirradiation conditions was measured. The results showed that thephotoreactivity of BrU in this sequence in RNA was approximatelytwofold slower than that measured in DNA (Fig. 2). The differencein photoreactivity between RNA and DNA is presumably due to thehalf-life of the anion radical of BrU that produces the uracil-5-ylradical. Obtaining additional general information regarding the dif-ference in photoreactivity between RNA and DNA will require theinvestigation of photoreactivity in a wide variety of sequences.

3.2. Products of photoirradiated r(CGAABrUUGC)/r(GCAAUUCG)

Next, we investigated the products of photoirradiated r(CGAAB-

rUUGC)/r(GCAAUUCG). In DNA, the analogous sequence 50-(G/C)AABrUT-30 has been identified as a hotspot sequence in whichthe efficient photoreaction of BrU occurs to provide alkaline labilesites.23,24

It has been proposed that the efficient photoreaction in this se-quence is caused by the good alignment of three factors: G as theelectron donor, the stacking BrUT as the electron acceptor, and A/Tas the bridge between the electron donor and the acceptor. The HPLCanalysis of photoirradiated r(CGAABrUUGC)/r(GCAAUUCG) showed

Scheme 1.

Figure 4. Photoproducts of r(CGCGBrUGCG)/r(CmGCACmGCG). Photoreaction wasperformed for 2 h in the presence of 0.1 M NaCl (A-form RNA) (a) or 3 M NaCl (Z-form RNA) (b) and the products were analyzed using HPLC. Photoirradiation of thissequence produced one major photoproduct only in the Z-form RNA condition, a 20-keto guanosine-containing product (c). The yield was 54.0 ± 0.1% in (6).

468 H. Morinaga et al. / Bioorg. Med. Chem. 21 (2013) 466–469