Biomedical Engineering Reference
In-Depth Information
size already makes computations much more difficult, but the purines moreover con-
tain NH
2
groups that render them slightly nonplanar. Loss of reflection symmetry
both adds to the computational cost and complicates the analysis of the results.
In the pyrimidines, we could carry out separate calculations for the
A
0
A
00
and
representations of the C
s
point group, and resonances could be classified as
or
resonances standing
according to which representation they occurred in, with the
A
00
, where background scattering is weak. In principle, scattering from
the purines should make no use of symmetry, with both the
out clearly in
resonances
to be determined from a single calculation, against a large nonresonant background.
To render our study of the purine bases [
88
] more tractable, we first determined
the effect of enforcing planarity by carrying out static-exchange calculations on
guanine in both planar and nonplanar geometries. The results indicated an upward
shift of about 0.2 eV in the
and
resonance energies of the planar form. With this
information in hand, we then carried out full calculations, including polarization,
for the
A
00
representation of both planar guanine and planar adenine in order to
obtain more accurate
resonance positions. As with the pyrimidines, we then
proceeded to study the effect of incorporating the bases into larger moieties, namely
the nucleosides deoxyguanosine and deoxyadenosine as well as the nucleotide 2
0
-
deoxyadenosine-5
0
-monophosphate or dAMP (Fig.
5.3
), at the static-exchange level
of approximation.
Our calculations including polarization effects placed the three lowest
resonances of adenine at 1.1, 1.8, and 4.1 eV, compared to experimental values
[
65
] of 0.54, 1.36, and 2.17 eV. Agreement is thus quite good for the first two
resonances—though not as good as for the pyrimidines, even after allowing for a
downward shift of 0.2 eV to compensate for imposing planar geometry—but much
poorer for
3
, likely indicating that in this case also it mixes with core-excited
resonances built on triplet excited states. The picture is somewhat different for
guanine: the SMC calculation places the first three
resonances at approximately
1.55, 2.4, and 3.75 eV, whereas the electron transmission measurement places them
at 0.46, 1.37, and 2.36 eV [
65
]. At first sight this comparison indicates much larger
errors in the calculated positions for
1
2
and
than we saw in adenine or any of
3
the pyrimidines, and a smaller error for
. However, as Aflatooni and coworkers
already noted [
65
], the dominant guanine isomer in the gas phase is believed to be
an enol form [
89
], rather than the keto isomer found in DNA, for which we did the
calculation. In contrast, gas-phase adenine is mostly the keto form that we assumed
[
90
]. Here we have, therefore, an example where theory provides information not
directly available from experiment: assuming the same shift of about 0.5 eV needed
to bring our calculated positions for
1
2
of adenine into agreement with
the measurement, we can predict that the same resonances in the keto tautomer of
guanine should lie at about 1.0 and 1.9 eV.
Comparison with the calculations of Tonzani and Greene [
67
] generally follows
the pattern seen in the pyrimidines, with reasonable qualitative agreement but with
their resonance positions shifted to higher energies compared to the SMC results:
thus they place
and
1
at 2.4 eV in both adenine and guanine.
