An ab initio molecular orbital study on the characteristics of 8-hydroxyguanine

Mutation Research, 192 (1987) 83-89

83

Elsevier

MTRL 048

An ab initio molecular orbital study on the characteristics of 8-hydroxyguanine Misako Aida and Susumu Nishimura Biophysics Division, National Cancer Center Research Institute, Tsukiji, Chuo-ku, Tokyo 104 (Japan) (Accepted 15 June 1987)

Keywords: 8-Hydroxyguanine; Ab initio MO; Electrostatic potential.

Summary To investigate the mechanism by which the 8-hydroxyguanine residue in DNA affects the fidelity of DNA replication, the intrinsic properties of this modified base were investigated using an ab initio molecular orbital method. The most stable 8-hydroxyguanine form was revealed to be 6,8-diketo. The addition of an oxygen atom to the 8 position of a guanine base was shown to change the electrostatic potential of the molecule entirely and to give it a negative character. This effect may influence the local structure of 8-hydroxyguanine-containing DNA and the interaction with DNA polymerase, thereby resulting in infidelity of DNA replication.

It has been shown that the hydroxylation of guanine in DNA to produce 8-hydroxyguanine may be an important factor in mutation and carcinogenesis caused by oxygen radicals (Kasai and Nishimura, 1984a,b,c, 1986; Kasai et al., 1984, 1986). Recently, synthetic oligonucleotides containing 8-hydroxyguanine in a specific position were used as a template for DNA synthesis and shown to be misread both at the modified base and at adjacent pyrimidine bases (Kuchino et al., 1987). In their experiment, provided 8-hydroxyguanine was positioned between thymine and cytosine in the oligonucleotide, the 8-hydroxyguanine directed the insertion of adenine, thymine, Correspondence: Dr. S. Nishimura, Biophysics Division, National Cancer Center Research Institute, Tsukiji, Chuo-ku, Tokyo 104 (Japan).

cytosine and guanine with almost equal frequencies, indicating 8-hydroxyguanine completely lacks specific base-pairing. In addition, at the adjacent pyrimidine bases, the insertion of incorrect bases was observed. The molecular mechanism of the misreading of an 8-hydroxyguanine residue is still unclear. To provide a clue toward the solution of this problem, the intrinsic properties of 8-hydroxyguanine were analyzed using an ab initio molecular orbital method, and compared with those of normal guanine. Guanine can take two tautomeric forms, by virtue of the keto-enol tautomerization (Fig. 1). The rare tautomer (Fig. lb) can be paired not ,with cytosine, but with thymine, and this might lead to mutation. It is well known, however, that guanine is usually in the keto form in DNA. Four kinds of

0165-7992/87/$ 03.50 (~ 1987 Elsevier Science Publishers B.V. (Biomedical Division)

84

0/H

0

H__\~i/~\/H /N H

H

~ N "/H

//

/

/

H

H

H

{b} ENOL FORM

KETO FORM

(-,I

Fig. 1. Two tautomeric forms of guanine.

07H

0

H

H

\

\o

0 N/C~N ~C

H

H

1

\N / / H

H

1=1 6-KETO. 8-ENOL FORM

tbl

6.8-DIENOL FORM

0/H

0

0

/N

~ . ~ ~N.I

[,l

.

/

H

H 6,8-DIKETO FORM

/ H

Idl

6-ENOL. 8-KETO FORM

Fig. 2. Four tautomeric forms of 8-hydroxyguanine.

85 0

0

1.212 ,..1

/EN~c

~0!

_C J.'~

H

N

'-/"' ~ N 'K°°°

/o.5

",..11 " " I H ,ze..// --~111.32 " H 12,.7' 8"~ ° ` - ~ x'°s" u H'z"-~,.2~'zz"°/'"'3°U "3¢C ,z2,3

/"

. ,/.~.o ~,<,,,

N '3" H

o.""V H

~;:~"

N~

/o.,.s H H ( . ) KETO FORN

0 ~'H6

I"°8~C Z

H

,~.... /

//,o,.?--c 3, ,~,3,N,~o-

1.371 ,.38, ] H __C//l"z"° 1.395 ,.o52 \X" N~%.. C. 4 ~ ` ~ C ,.3~ H

~

~C

/ H

0 110.53 l ,e.z3 ..N-..~,.-/7;,~o.,~.~

1.388[ .Cj.3, 5 ,/.. ~ ' i. 3~o

N

N ~"'"

/,19.2;, H

//,05.sl

125..//

\N~.,

ells.z3 II

/ ,~5.~o

..,N °'°

N

H

o.3as/ H

(b)

[

"-~120.08 N./ H 121.04 H

ENOL

FORM

Fig. 3.3-21G optimized geometries of (a) the keto form and (b) the enol form of guanine. Bond lengths, in ~mgstroms,are given on the left and bond angles, in degrees, on the right.

tautomeric form can be adopted by 8-hydroxyguanine, by virtue of the keto-enol tautomerization, as shown in Fig. 2. Here, using a geometric optimization procedure, we established the geometry of each and their relative stabilities. Further, the effects of adding an oxygen atom to a guanine residue on the structure of the D N A molecule and its interaction with a protein are discussed in the light of electrostatic potential.

Methods Ab initio SCF calculations were carried out with the I M S P A C K program system ( M o r o k u m a et al., 1980). The geometries of the purine bases were optimized on a 3-21G basis set (Binkley et al., 1980), using gradient optimization techniques (Pulay, 1977). The electrostatic potential m a p was drawn using the graphic program, MGP98 (Hirono et al.,

1984). The classic formula for electrostatic potential at point x, V(x), for a system of charges q~ at points ri in a vacuum is: V(x) = E qi/Iri-x]. i

Net atomic charges, qi, were calculated according to Mulliken's population analysis (Mulliken, 1955), with an STO-3G basis set (Hehre et al., 1969). This was because the charges obtained with the 3-21G basis set were overestimated. The electrostatic potentials were calculated at the molecular surface, i.e., the surface 1.5 ,~ distant f r o m the van der Waals sphere (Hirono et al., 1984).

Results and discussion The optimized 3-21G structures of the two tautomeric forms of guanine and the four

86 0

0 1,212 N

H 0.855 X

,2,, • 1.4.0, ~ ""// C//

0

i 4.28

119.77

C-

H

~ > " , -~

C

1.39e

.N

./o.., "

\ \110.11 0

,. 394.

II

I . 280

1 408

"

H

/H

i

l 10.72

1.343 I 1335

1"32°'~N ' ~ / "

0

C

1343

"':'~ NZ;°"8H' /o.,s H (hi 6 , 8 - D I E N O L

C 113 91

H7.BZXlO9.3o "

IIC 126"22 124."4.2CI 1,7.51

/-12,.15

N--

•

/

H

FORH 0

L371/

0 ~

II 121,

C/

r~ ,..:~...~ "/

,,2~

,.,.o3 C

•

I

II 354. II 1 369 II

1.2H \ \

H~

II

I ~9..-C'~.

\

/0.885

H

/

N 1.oo8

OLO3.

.~c4~.

N-/

,.3.

H

/o.987 H 6,8-DIKETO

(c)

II

117.57/

FORH /H

H 0/0.375 0 995

N'~

1.39y 13,0 C "1.373

0 --C "2°~. ,,sX~

~..-C ~

N I.001

0 111.19

1.331

1.~

6~ 125.3 N

/

1.31~N 1.331

0 - -

,~C.,3

~

1.3,,'-- NT.~" ~ N / ,

H

(d)

I

126.79\ 129.14. 122.30 [ 117.85 \HO.el C " 114.85 ~ C \ . ,~ / ,~'g'-11" \~, 2 3 . 1 2 / 12o.23 ,, ~?/"" / 125.9;' I"~ N / H H

\N./~o..4.

H

115.71

123 92\ .C. H " . N -~ ,31.28/~og ."-.. / ~ " " N /118.2, ~ 121.35 125-4.1 i 120.99

1 I .352 I I 3 4. l

,.5\N:~>.cL~

" ~ ;"

119.4.0 N

.

0 H o 88,\

119.38

N --.~ ,,/,2o 2o~

!

1" 33~. 382X N ~.37~1,~.~C~

/0.8,5 H

""~'N ~t/H

FORH

t

•

0

N~

H

H

H O.9 6 k

123 16 I 117 93 "":. C '

118.6~/

o~, I. 31.C

t25e2N 2oa I I

127 83 "

\1o5.94. ~ C . ; N2,93''~125' .2 '~

/0..3 H 6-KETO,8-ENOL

/o)

C1185, II " II

119.29\

1"3~1.377X ~ ~.....C~ 1.29~C\,.35, H ~ N"~°'°" 890 /.N880 134.3~ N / ~ H

//o5o9 // C ,,3.,3

H

118.75

~r, 110.60

-c 10,.81 127.75\

"~NI ~' 115.97"'~120.75 II

II I II 125.06 123.16 I 119.19

X~ ~,.c,,'-.Li, 3 ~ c \ 121, , ,

0o0

/o.932 H 6-ENOL,8-KETO

L

135-57 ,,,~11tI

-

H

125.o8

N

~

/

,m.26 N

/

H

FORH

Fig. 4. 3-21G optimized geometries of (a) the 6-keto,8-enol form, (b) the 6,8-dienol form, (c) the 6,8-diketo form and (d) the 6-enol,8-keto form of 8-hydroxyguanine. Bond lengths, in ~ngstr6ms, are given on the left and bond angles, in degrees, on the right•

87 tautomeric forms of 8-hydroxyguanine are illustrated in Figs. 3 and 4, respectively. As shown in these figures, the geometry o f the six-membered ring of the purine skeleton varied greatly, as did the shift of the hydrogen atom, according to the keto-enol tautomerization between the 1 and 6 positions. The geometrical variation following the addition of an oxygen a t o m to the 8 position of guanine was large in the 8-keto form. The keto-enol tautomerization between the 7 and 8 positions mainly influenced the geometry of the five-membered ring. Thus, the precise structure of each tautomeric f o r m could be obtained only after the optimization procedure with the ab initio molecular orbital method, when the precise relative energies could be drawn. The basis set used here (3-21G) is satisfactory for these purposes. The energy difference between the two forms of guanine is shown in Table 1. The normal keto form of guanine was found to be 4.95 kcal/mole more stable than the enol form. This result agreed well with the experimental observation that guanine normally exists in keto form. TABLE 1 RELATIVE ENERGIES OF TWO GUANINE TAUTOMERS Form

Relative energy (kcal/mole)

keto enol

0.0 + 4.95

The relative energies of the four forms of 8-hydroxyguanine are summarized in Table 2. The most stable was the 6,8-diketo form (Fig. 2c). Since the 6-keto,8-enol form (Fig. 2a) was 20.93 k c a l / m o l e less stable than the 6,8-diketo form, the

TABLE 2 RELATIVE ENERGIES OF FOUR 8-HYDROXYGUANINE TAUTOMERS Form

Relative energy (kcal/mole)

6-keto,8-euol 6,8-dienol 6,8-diketo 6-enol,8-keto

+ 20.93 + 24.60 0.0 + 2.29

/

H

H

,,~N. J? .... H-N\c--C/~ H"C ~ \ C ' / C \ ff \C--H /

\?

(a

'\

0UANINE : CYTOSINE H .171

(hi

n

H -~,

/H

8-HYOROXYOUANINE : CYTOSINE

Fig. 5. Watson-Crick type base pairs of (a) guanine:cytosine and (b) 8-hydroxyguanine:cytosine. 6-keto,8-enol form should be found to exist only rarely. This is in accord with the X-ray analysis of 9-ethyl-8-hydroxyguanine, which showed a 6,8diketo form (Kasai et al., 1987). O f the 8-keto forms, the 6-keto form (Fig. 2c) was 2.29 kcal/ mole more stable than the 6-enol form (Fig. 2d). Although this energy difference was smaller than it was with guanine, it can be concluded that the 8-keto form exists intrinsically only in 6-keto form. Because of the small energy difference, however, it might be possible for 8-hydroxyguanine to exist in the 6-enol,8-keto form, under the influence of some other components, close to 8-hydroxyguanine, which could stabilize its 6-enol form. These results indicate the base pairing property of 8-hydroxyguanine to be intrinsically the same as that of guanine. When 8-hydroxyguanine is in the anti conformation, with respect to the sugar-base torsion angle, it is paired only with cytosine, as is the case with guanine (Fig. 5). This is in contrast to O6-methylguanine, which is known to induce point mutations: this modified base was revealed, by an ab initio molecular orbital method, to be specifically paired with thymine (Nagata et al., 1982). Thus, the observation of 8-hydroxyguanine affecting the fidelity of D N A replication (Kuchino

88

\ ./

•

~ ;•

q.~.

r

"(3 i

Fig. 6. Maps of electrostatic potential (ESP) at the molecular surfaces, 1.5 .~ distant from van der Waals spheres, for (a) guanine and (b) 8-hydroxyguanine. Dot: red, 9 < ESP; pink, 3 < ESP < 9; yellow, - 3 < ESP < 3; light blue, - 9 < ESP = -3; blue, - 15 < ESP < -9; green, ESP _<_ - 15 (kcal/mole). et al., 1987) should not be ascribed to its potential ability to make an abnormal base pair, although the possibility of a sys-anti conformational change was not taken into consideration in the present study. 8-Hydroxyguanine induced misreading, during D N A synthesis, immediately opposite its own position and, in addition, at the adjacent pyrimidines (Kuchino et al., 1987). The extent of misreading caused by the presence of the 8-hydroxyguanine varied depending on the base sequences surrounding it. It appeared that the 8-hydroxyguanine had a great influence upon its surroundings. One possible origin of such an influence is electrostatic energy; it was, therefore, worth investigating the electrostatic potential of 8-hydroxyguanine. Maps of the electrostatic potentials at the molecular surfaces, 1.5 ,~ distant f r o m van der Waals spheres, for guanine (keto form, Fig. la) and 8-hydroxyguanine (6,8-diketo form, Fig. 2c) are shown in Fig. 6a and b, respectively. A redcolored dot in these figures represents a point where the electrostatic potential (ESP) was greater than or equal to + 9 . 0 kcal/mole (ESP __> 9.0). Pink, yellow, light blue, blue and green dots indicate 3.0 __< ESP < 9.0, - 3 . 0 < ESP < 3.0, - 9 . 0 < ESP _<_ - 3 . 0 , - 15.0 < ESP _<_ - 9 . 0 a n d ESP __< - 1 5 . 0 , respectively. The addition of an oxygen a t o m to the 8 position of guanine remarkably influenced the entire electrostatic

potential. The added oxygen atom gave a negative character to the whole molecule. The change in electrostatic potential is obviously relevant to interaction with metal ions. Besides this simple one, two other kinds of effect could be produced, as described below. • First, the electrostatic force plays a key role in base-base stacking interactions in the D N A double helix, the sequence dependency of stacking interaction energy being due to the electrostatic energy (Aida and Nagata, 1986). In addition, sequencedependent conformational preference for A- or Btype on a D N A fragment results from stacking interactions (Aida a n d Nagata, 1986; Aida, unpublished data). The large change in electrostatic potential of one base may, therefore, result in a variation in the D N A ' s local structure. Second, the electrostatic potential is known to be useful in understanding the qualitative features of protein-substrate interactions (Hayes and Kollman, 1976). Electrostatic force has an important role in interactions between substrates and enzymes. Consequently, it is probable that a marked variation in the electrostatic potential of one base may influence the D N A ' s local structure and the way in which it interacts with a protein such as D N A polymerase. This means that the way in which D N A 1oolymerase acts may be affected, resulting in the induction of misreadings of 8-hydroxyguanine and o f its adjacent bases.

89

It is concluded that 6,8-diketo is the most stable f o r m of 8-hydroxyguanine. The addition of an oxygen a t o m to the 8 position of a guanine entirely changes the electrostatic potential o f the molecule, and this could affect the action of D N A polymerase.

Acknowledgements Numerical calculations were made on a H I T A C M-240H computer at the National Cancer Center, T o k y o , and a H I T A C M-680H computer at the Institute for Molecular Science, Okazaki. We thank the members of these institutes who facilitated the use o f these computers.

References Aida, M., and C. Nagata (1986) An ab initio molecular orbital study on the stacking interaction between nucleic acid bases: dependence on the sequence and relation to the conformation, Int. J. Quantum Chem., 29, 1253-1261. Binkley, J.S., J.A. Pople and W.J. Hehre (1980) Self-consistent molecular orbital methods, 21. Small split-valence basis sets for first-row elements, J. Am. Chem. Soc., 102, 939-947. Hayes, D.M., and P.A. Kollman (1976) Electrostatic potentials of proteins. 1. Carboxypeptidase A, J. Am. Chem. Soc., 98, 3335-3345. Hehre, W.J., R.F. Stewart and J.A. Pople (1969) Selfconsistent molecular-orbital methods. I. Use of gaussian expansions of slater-type atomic orbitals, J. Chem. Phys., 51, 2657-2664. Hirono, S., H. Umeyama and 1. Moriguchi (1984) Electrostatic potential images of drugs targetting dopamine receptors, Chem. Pharm. Bull., 32, 3061-3065. Kasai, H., and S. Nishimura (1984a) Hydroxylation of deoxyguanosine at the C-8 position by ascorbic acid and other reducing agents, Nucleic Acids Res., 12, 2137-2145.

Kasai, H., and S. Nishimura (1984b) Hydroxylation of deoxyguanosine at the C-8 position by polyphenols and aminophenols in the presence of hydrogen peroxide and ferric ion, Gann, 75, 565-566. Kasai, H., and S. Nishimura (1984c) DNA damage induced by asbestos in the presence of hydrogen peroxide, Gann, 75, 841-844. Kasai, H. and S. Nishimura (1986) Hydroxylation of guanine in nucleosides and DNA at the C-8 position by heated glucose and oxygen radical-forming agents, Environ. Health Perspect., 67, 111-116. Kasai, H., H. Tanooka and S. Nishimura (1984) Formation of 8-hydroxiguanine residues in DNA by X-irradiation, Gann, 75, 1037-1039. Kasai, H., P.F. Crain, Y. Kuchino, S. Nishimura, A. Ootsuyama and H. Tanooka (1986) Formation of 8-hydroxyguanine moiety in cellular DNA by agents producing oxygen radicals and evidence for its repair, Carcinogenesis, 7, 1849-1851. Kasai, H., S. Nishimura, Y. Toriumi, A. ltai and Y. litaka (1987) The crystal structure of 9-ethyl-8-hydroxyguanine, Bull. Chem. Soc. Jpn., in press. Kuchino, Y., F. Mori, H. Kasai, H. lnoue, S. Iwai, K. Miura, E. Ohtsuka and S. Nishimura (1987) DNA templates containing 8-hydroxydeoxyguanosine are misread both at the modified base and at adjacent residues, Nature (London), 327, 77-79. Morokuma, K., S. Kato, K. Kitaura, 1. Ohmine, S. Sakai and S. Obara (1980) Institute for Molecular Science Computer Center Library Program, No. 0372. Mullikan, R.S. (1955) Electronic population analysis on LCAO-MO molecular wave functions. 1, J. Chem. Phys., 23, 1833-1840. Nagata, C., E. Takeda and M. Aida (1982) Why O6-alkyl guanine is specifically promutagenic? Ab initio molecular orbital consideration, Mutation Res., 105, 1-8. Pulay, P. (1977) Direct use of the gradient for investigating molecular surfaces, in: H.F. Schaefer Ill (Ed.), Applications of Electronic Structure Theory, Plenum, New York, Ch. 4. Communicated by R.J. Preston