(528g) Encapsulation Mechanism of DNA Nucleobases Onto Carbon Nanotubes: Free Energy Landscapes

     Ever since the
unravelling of the double helix structure of DNA, the molecule has been
considered as the building block of biological organisms. Because DNA is the
agent that carries genetic information from one generation to the next, the
knowledge of its physical and chemical properties is of paramount importance.
Much of that information has to do with the way the nucleobases
(NBs)—DNA's own building blocks—interact with their surrounding
environment. Owing to their nanoscale dimensions, carbon nanotubes are able to
selectively interact with DNA almost at the atomic level: homo- and
heteropolymer ssDNA will readily adsorb onto the external walls of small
diameter SWCNTs,¹ according to a p–stacking
mechanism of the NBs on the walls. We have already characterized the SWCNTs
energetic landscape regarding exo- and endoadsorption,
using small probe molecules.^2,3 While
exoadsorption on small diameter nanotubes (< 1 nm) is reasonably understood,
confinement driven adsorption/binding is a phenomenon that remains quite
unexplored. Okada⁴ determined the encapsulation energy of (in
vacuum) ssDNA as a function of nanotube diameter, and suggested an intricate
relationship between both, particularly in the range 1.25 < D (nm) < 1.8. As far as we are aware,
there are no previous studies of intratubularly confined aqueous nucleobases.
We shall address this issue using the well-tempered metadynamics scheme
proposed by Parrinello,⁵ estimating the
free-energy landscapes associated with the encapsulation process of Adenine (purine)
and Thymine (pyrimidine) onto two different diameter (zig-zag) nanotubes, (16,0)
with D = 1.25 nm and (23,0) with D = 1.8 nm. These two topologies
correspond to the narrowest and largest nanopores obtained by electric-arc
discharge.³ Because nanotubes can be electrically charged,⁶ the effect of charge density will be addressed
considering purely hydrophobic solids and nanotubes with electric charges of q = + 0.05 e^–/C. Physiological
conditions (T=310K) are mimicked by explicitly including H₂O
molecules and Cl^– ions in the
calculations.

II.
Results and Discussion

     The
encapsulation mechanism of both purine- and pyrimidine-types onto hydrophobic
SWCNTs is energetically favorable and irreversible during the time windows
spanned in the simulations (0.1 µs); the resulting NB/SWCNT hybrids exhibit
lower free energies compared to the unconfined systems (Fig.1). The highly
symmetrical free-energy landscapes obtained are essentially one dimensionally
isotropic regarding the x₁
order parameter, in direct relationship to the nanotubes quasi 1-D
symmetry. Once confined inside the
nanotube, a nucleobase can explore a region whose boundaries are comprised
between the entrances, 0.15 < x₁
(nm) < L_z (where L_z
is the SWCNT length), and the inner surface of the wall, never returning to the
bulk. It is very interesting to observe that confinement effects are remarkably
stronger at the nanopore center, where free energy differences are largest,  [ x₁,
x₂]^(16,0)
= (1.1, 0.08) nm and [x₁, x₂]^(23,0)
= (1.7, 0.08) nm. Independent umbrella sampling calculations identified a unique
probability distribution maximum corresponding to a domain comprised between
the solid walls and the nanopore center (Fig.2). Once confined, molecules are
in direct contact with the walls at a minimum distance of closest approach of
0.26–0.28 nm. No layer of solvent could be observed between hydrophobic slabs.
This is due to a robust  pi-stacking mechanism of the NB onto the
graphitic mesh, leading to strong dispersive energies of interaction with both the (16,0) and (23,0) topologies (Table I). These results
have been compared with DFT and experimental data for graphene. Contrary to
exoadsorption, intratubular binding affinity increases with nanotube curvature,
as a result of enhanced interaction with the wall immediately opposite to the
adsorptive one. We have estimated a diameter of D = 2.05 nm above which the SWCNT curvature mimics a flat graphene
sheet and confinement effects tend to be negligible. Adenine and thymine
evidence similar interaction energies with the hydrophobic nanopores, however,
adenine confinement results in slightly favored energetics of ca. 0.9 – 1.4 kJ/mol (Table I).
This previously unobserved finding indicate a binding affinity order: purines
> pyrimidines.

     Electrically
charged nanotubes evidence a completely different behavior. The narrowest
nanopore employed here, (16,0), inhibits encapsulation of either the purine or
pyrimidine moiety. Instead, that nanotube favors adsorption at a region very
close to the pore entrance as indicated by the probability distribution maxima at
 W = 0.97 nm and  W = 1.04 nm (Fig.2). When the nucleobases are put in contact with
the (23,0) solids, several encapsulation/exit events can occur (Fig.3); moreover,
the confined systems exhibit lower free energies. The probability distribution
curves indicate two maxima: i) close to the pore
center, overlapping the one obtained for the corresponding hydrophobic solid,
and ii) at the pore entrance. This
second region of confinement is energetically less stable than the pore center
(Table I), particularly in the case of adenine where deltaE=40 kJ/mol. Because the
pyrimidine molecule has a larger dipole moment, its exposure to the bulk
solution carries smaller electrostatic penalties, and therefore its location at
the pore center is stabilized by only ca.
12 kJ/mol compared to the entrance.

References

 ADDIN EN.REFLIST (1) Johnson, R. R.; Johnson, A. T. C.; Klein, M. L.
Nano Letters 2008, 8, 69.

(2) Cruz,
F. J. A. L.; Mota, J. P. B. Phys. Rev. B
2009, 79, 165426.

(3) Cruz,
F. J. A. L.; Esteves, I. A. A. C.; Agnihotri, S.; Mota, J. P. B. J. Phys. Chem. C 2011, 115, 2622.

(4) Kamiya,
K.; Okada, S. Phys. Rev. B 2011, 83, 155444.

(5) Barducci,
A.; Bussi, G.; Parrinello, M. Phys. Rev.
Let. 2008, 100, 020603.

(6) Zhao,
X.; Johnson, J. K. J. Am. Chem. Soc. 2007, 129, 10438.